Datasets:

ACLMeeting

/

ACL2024-papers

like 2

Follow

Annual Meeting of the Association for Computational Linguistics 1

https://aclanthology.org/2024.acl-long.101.bib

@inproceedings{li-li-2024-aoe, title = "{A}o{E}: Angle-optimized Embeddings for Semantic Textual Similarity", author = "Li, Xianming and Li, Jing", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.101", pages = "1825--1839", abstract = "Text embedding is pivotal in semantic textual similarity (STS) tasks, which are crucial components in Large Language Model (LLM) applications. STS learning largely relies on the cosine function as the optimization objective to reflect semantic similarity. However, the cosine has saturation zones rendering vanishing gradients and hindering learning subtle semantic differences in text embeddings. To address this issue, we propose a novel Angle-optimized Embedding model, AoE. It optimizes angle differences in complex space to explore similarity in saturation zones better. To set up a comprehensive evaluation, we experimented with existing short-text STS, our newly collected long-text STS, and downstream task datasets. Extensive experimental results on STS and MTEB benchmarks show that AoE significantly outperforms popular text embedding models neglecting cosine saturation zones. It highlights that AoE can produce high-quality text embeddings and broadly benefit downstream tasks.", }

Text embedding is pivotal in semantic textual similarity (STS) tasks, which are crucial components in Large Language Model (LLM) applications. STS learning largely relies on the cosine function as the optimization objective to reflect semantic similarity. However, the cosine has saturation zones rendering vanishing gradients and hindering learning subtle semantic differences in text embeddings. To address this issue, we propose a novel Angle-optimized Embedding model, AoE. It optimizes angle differences in complex space to explore similarity in saturation zones better. To set up a comprehensive evaluation, we experimented with existing short-text STS, our newly collected long-text STS, and downstream task datasets. Extensive experimental results on STS and MTEB benchmarks show that AoE significantly outperforms popular text embedding models neglecting cosine saturation zones. It highlights that AoE can produce high-quality text embeddings and broadly benefit downstream tasks.

AoE: Angle-optimized Embeddings for Semantic Textual Similarity

acl-long.101

Poster

https://aclanthology.org/2024.acl-long.101/

https://aclanthology.org/2024.acl-long.102.bib

@inproceedings{wang-etal-2024-incharacter, title = "{I}n{C}haracter: Evaluating Personality Fidelity in Role-Playing Agents through Psychological Interviews", author = "Wang, Xintao and Xiao, Yunze and Huang, Jen-tse and Yuan, Siyu and Xu, Rui and Guo, Haoran and Tu, Quan and Fei, Yaying and Leng, Ziang and Wang, Wei and Chen, Jiangjie and Li, Cheng and Xiao, Yanghua", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.102", pages = "1840--1873", abstract = "Role-playing agents (RPAs), powered by large language models, have emerged as a flourishing field of applications. However, a key challenge lies in assessing whether RPAs accurately reproduce the personas of target characters, namely their character fidelity. Existing methods mainly focus on the knowledge and linguistic patterns of characters. This paper, instead, introduces a novel perspective to evaluate the personality fidelity of RPAs with psychological scales. Overcoming drawbacks of previous self-report assessments on RPAs, we propose InCharacter, namely **In**terviewing **Character** agents for personality tests. Experiments include various types of RPAs and LLMs, covering 32 distinct characters on 14 widely used psychological scales. The results validate the effectiveness of InCharacter in measuring RPA personalities. Then, with InCharacter, we show that state-of-the-art RPAs exhibit personalities highly aligned with the human-perceived personalities of the characters, achieving an accuracy up to 80.7{\%}.", }

Role-playing agents (RPAs), powered by large language models, have emerged as a flourishing field of applications. However, a key challenge lies in assessing whether RPAs accurately reproduce the personas of target characters, namely their character fidelity. Existing methods mainly focus on the knowledge and linguistic patterns of characters. This paper, instead, introduces a novel perspective to evaluate the personality fidelity of RPAs with psychological scales. Overcoming drawbacks of previous self-report assessments on RPAs, we propose InCharacter, namely **In**terviewing **Character** agents for personality tests. Experiments include various types of RPAs and LLMs, covering 32 distinct characters on 14 widely used psychological scales. The results validate the effectiveness of InCharacter in measuring RPA personalities. Then, with InCharacter, we show that state-of-the-art RPAs exhibit personalities highly aligned with the human-perceived personalities of the characters, achieving an accuracy up to 80.7{\%}.

InCharacter: Evaluating Personality Fidelity in Role-Playing Agents through Psychological Interviews

acl-long.102

Poster

2310.17976

https://huggingface.co/papers/2310.17976

https://aclanthology.org/2024.acl-long.102/

https://aclanthology.org/2024.acl-long.103.bib

@inproceedings{liu-etal-2024-detectgpt, title = "Does {D}etect{GPT} Fully Utilize Perturbation? Bridging Selective Perturbation to Fine-tuned Contrastive Learning Detector would be Better", author = "Liu, Shengchao and Liu, Xiaoming and Wang, Yichen and Cheng, Zehua and Li, Chengzhengxu and Zhang, Zhaohan and Lan, Yu and Shen, Chao", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.103", pages = "1874--1889", abstract = "The burgeoning generative capabilities of large language models (LLMs) have raised growing concerns about abuse, demanding automatic machine-generated text detectors. DetectGPT, a zero-shot metric-based detector, first introduces perturbation and shows great performance improvement. However, in DetectGPT, the random perturbation strategy could introduce noise, and logit regression depends on the threshold, harming the generalizability and applicability of individual or small-batch inputs. Hence, we propose a novel fine-tuned detector, PECOLA, bridging metric-based and fine-tuned methods by contrastive learning on selective perturbation. Selective strategy retains important tokens during perturbation and weights for multi-pair contrastive learning. The experiments show that PECOLA outperforms the state-of-the-art (SOTA) by 1.20{\%} in accuracy on average on four public datasets. And we further analyze the effectiveness, robustness, and generalization of the method.", }

The burgeoning generative capabilities of large language models (LLMs) have raised growing concerns about abuse, demanding automatic machine-generated text detectors. DetectGPT, a zero-shot metric-based detector, first introduces perturbation and shows great performance improvement. However, in DetectGPT, the random perturbation strategy could introduce noise, and logit regression depends on the threshold, harming the generalizability and applicability of individual or small-batch inputs. Hence, we propose a novel fine-tuned detector, PECOLA, bridging metric-based and fine-tuned methods by contrastive learning on selective perturbation. Selective strategy retains important tokens during perturbation and weights for multi-pair contrastive learning. The experiments show that PECOLA outperforms the state-of-the-art (SOTA) by 1.20{\%} in accuracy on average on four public datasets. And we further analyze the effectiveness, robustness, and generalization of the method.

Does DetectGPT Fully Utilize Perturbation? Bridging Selective Perturbation to Fine-tuned Contrastive Learning Detector would be Better

acl-long.103

Poster

2402.00263

https://aclanthology.org/2024.acl-long.103/

https://aclanthology.org/2024.acl-long.104.bib

@inproceedings{ni-etal-2024-afacta, title = "{AF}a{CTA}: Assisting the Annotation of Factual Claim Detection with Reliable {LLM} Annotators", author = "Ni, Jingwei and Shi, Minjing and Stammbach, Dominik and Sachan, Mrinmaya and Ash, Elliott and Leippold, Markus", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.104", pages = "1890--1912", abstract = "With the rise of generative AI, automated fact-checking methods to combat misinformation are becoming more and more important. However, factual claim detection, the first step in a fact-checking pipeline, suffers from two key issues that limit its scalability and generalizability: (1) inconsistency in definitions of the task and what a claim is, and (2) the high cost of manual annotation. To address (1), we review the definitions in related work and propose a unifying definition of factual claims that focuses on verifiability. To address (2), we introduce AFaCTA (Automatic Factual Claim deTection Annotator), a novel framework that assists in the annotation of factual claims with the help of large language models (LLMs). AFaCTA calibrates its annotation confidence with consistency along three predefined reasoning paths. Extensive evaluation and experiments in the domain of political speech reveal that AFaCTA can efficiently assist experts in annotating factual claims and training high-quality classifiers, and can work with or without expert supervision. Our analyses also result in PoliClaim, a comprehensive claim detection dataset spanning diverse political topics.", }

With the rise of generative AI, automated fact-checking methods to combat misinformation are becoming more and more important. However, factual claim detection, the first step in a fact-checking pipeline, suffers from two key issues that limit its scalability and generalizability: (1) inconsistency in definitions of the task and what a claim is, and (2) the high cost of manual annotation. To address (1), we review the definitions in related work and propose a unifying definition of factual claims that focuses on verifiability. To address (2), we introduce AFaCTA (Automatic Factual Claim deTection Annotator), a novel framework that assists in the annotation of factual claims with the help of large language models (LLMs). AFaCTA calibrates its annotation confidence with consistency along three predefined reasoning paths. Extensive evaluation and experiments in the domain of political speech reveal that AFaCTA can efficiently assist experts in annotating factual claims and training high-quality classifiers, and can work with or without expert supervision. Our analyses also result in PoliClaim, a comprehensive claim detection dataset spanning diverse political topics.

AFaCTA: Assisting the Annotation of Factual Claim Detection with Reliable LLM Annotators

acl-long.104

Poster

2402.11073

https://huggingface.co/papers/2402.11073

https://aclanthology.org/2024.acl-long.104/

https://aclanthology.org/2024.acl-long.105.bib

@inproceedings{schimanski-etal-2024-towards, title = "Towards Faithful and Robust {LLM} Specialists for Evidence-Based Question-Answering", author = "Schimanski, Tobias and Ni, Jingwei and Kraus, Mathias and Ash, Elliott and Leippold, Markus", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.105", pages = "1913--1931", abstract = "Advances towards more faithful and traceable answers of Large Language Models (LLMs) are crucial for various research and practical endeavors. One avenue in reaching this goal is basing the answers on reliable sources. However, this Evidence-Based QA has proven to work insufficiently with LLMs in terms of citing the correct sources (source quality) and truthfully representing the information within sources (answer attributability). In this work, we systematically investigate how to robustly fine-tune LLMs for better source quality and answer attributability. Specifically, we introduce a data generation pipeline with automated data quality filters, which can synthesize diversified high-quality training and testing data at scale. We further introduce four test sets to benchmark the robustness of fine-tuned specialist models. Extensive evaluation shows that fine-tuning on synthetic data improves performance on both in- and out-of-distribution. Furthermore, we show that data quality, which can be drastically improved by proposed quality filters, matters more than quantity in improving Evidence-Based QA.", }

Advances towards more faithful and traceable answers of Large Language Models (LLMs) are crucial for various research and practical endeavors. One avenue in reaching this goal is basing the answers on reliable sources. However, this Evidence-Based QA has proven to work insufficiently with LLMs in terms of citing the correct sources (source quality) and truthfully representing the information within sources (answer attributability). In this work, we systematically investigate how to robustly fine-tune LLMs for better source quality and answer attributability. Specifically, we introduce a data generation pipeline with automated data quality filters, which can synthesize diversified high-quality training and testing data at scale. We further introduce four test sets to benchmark the robustness of fine-tuned specialist models. Extensive evaluation shows that fine-tuning on synthetic data improves performance on both in- and out-of-distribution. Furthermore, we show that data quality, which can be drastically improved by proposed quality filters, matters more than quantity in improving Evidence-Based QA.

Towards Faithful and Robust LLM Specialists for Evidence-Based Question-Answering

acl-long.105

Poster

2402.08277

https://aclanthology.org/2024.acl-long.105/

https://aclanthology.org/2024.acl-long.106.bib

@inproceedings{dou-etal-2024-loramoe, title = "{L}o{RAM}o{E}: Alleviating World Knowledge Forgetting in Large Language Models via {M}o{E}-Style Plugin", author = "Dou, Shihan and Zhou, Enyu and Liu, Yan and Gao, Songyang and Shen, Wei and Xiong, Limao and Zhou, Yuhao and Wang, Xiao and Xi, Zhiheng and Fan, Xiaoran and Pu, Shiliang and Zhu, Jiang and Zheng, Rui and Gui, Tao and Zhang, Qi and Huang, Xuanjing", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.106", pages = "1932--1945", abstract = "Supervised fine-tuning (SFT) is a crucial step for large language models (LLMs), enabling them to align with human instructions and enhance their capabilities in downstream tasks. Substantially increasing instruction data is a direct solution to align the model with a broader range of downstream tasks or notably improve its performance on a specific task. However, we find that large-scale increases in instruction data can damage the world knowledge previously stored in LLMs. To address this challenge, we propose LoRAMoE, a novelty framework that introduces several low-rank adapters (LoRA) and integrates them by using a router network, like a plugin version of Mixture of Experts (MoE). It freezes the backbone model and forces a portion of LoRAs to focus on leveraging world knowledge to solve downstream tasks, to alleviate world knowledge forgetting. Experimental results show that, as the instruction data increases, LoRAMoE can significantly improve the ability to process downstream tasks, while maintaining the world knowledge stored in the LLM. Our code is available at https://github.com/Ablustrund/LoRAMoE.", }

Supervised fine-tuning (SFT) is a crucial step for large language models (LLMs), enabling them to align with human instructions and enhance their capabilities in downstream tasks. Substantially increasing instruction data is a direct solution to align the model with a broader range of downstream tasks or notably improve its performance on a specific task. However, we find that large-scale increases in instruction data can damage the world knowledge previously stored in LLMs. To address this challenge, we propose LoRAMoE, a novelty framework that introduces several low-rank adapters (LoRA) and integrates them by using a router network, like a plugin version of Mixture of Experts (MoE). It freezes the backbone model and forces a portion of LoRAs to focus on leveraging world knowledge to solve downstream tasks, to alleviate world knowledge forgetting. Experimental results show that, as the instruction data increases, LoRAMoE can significantly improve the ability to process downstream tasks, while maintaining the world knowledge stored in the LLM. Our code is available at https://github.com/Ablustrund/LoRAMoE.

LoRAMoE: Alleviating World Knowledge Forgetting in Large Language Models via MoE-Style Plugin

acl-long.106

Poster

https://huggingface.co/papers/2312.09979

https://aclanthology.org/2024.acl-long.106/

https://aclanthology.org/2024.acl-long.107.bib

@inproceedings{zhang-etal-2024-self, title = "Self-Alignment for Factuality: Mitigating Hallucinations in {LLM}s via Self-Evaluation", author = "Zhang, Xiaoying and Peng, Baolin and Tian, Ye and Zhou, Jingyan and Jin, Lifeng and Song, Linfeng and Mi, Haitao and Meng, Helen", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.107", pages = "1946--1965", abstract = "Despite showing impressive abilities, large language models (LLMs) often struggle with factual inaccuracies, i.e., {''}hallucinations{''}, even when they hold relevant knowledge. To mitigate these hallucinations, current approaches typically necessitate high-quality human factuality annotations. In this work, we explore Self-Alignment for Factuality, where we leverage the self-evaluation capability of an LLM to provide training signals that steer the model towards factuality. Specifically, we incorporate Self-Eval, a self-evaluation component, to prompt an LLM to validate the factuality of its own generated responses solely based on its internal knowledge. Additionally, we design Self-Knowledge Tuning (SK-Tuning) to augment the LLM{'}s self-evaluation ability by improving the model{'}s confidence estimation and calibration. We then utilize these self-annotated responses to fine-tune the model via Direct Preference Optimization algorithm. We show that the proposed self-alignment approach substantially enhances factual accuracy over Llama family models across three key knowledge-intensive tasks on TruthfulQA and BioGEN.", }

Despite showing impressive abilities, large language models (LLMs) often struggle with factual inaccuracies, i.e., {''}hallucinations{''}, even when they hold relevant knowledge. To mitigate these hallucinations, current approaches typically necessitate high-quality human factuality annotations. In this work, we explore Self-Alignment for Factuality, where we leverage the self-evaluation capability of an LLM to provide training signals that steer the model towards factuality. Specifically, we incorporate Self-Eval, a self-evaluation component, to prompt an LLM to validate the factuality of its own generated responses solely based on its internal knowledge. Additionally, we design Self-Knowledge Tuning (SK-Tuning) to augment the LLM{'}s self-evaluation ability by improving the model{'}s confidence estimation and calibration. We then utilize these self-annotated responses to fine-tune the model via Direct Preference Optimization algorithm. We show that the proposed self-alignment approach substantially enhances factual accuracy over Llama family models across three key knowledge-intensive tasks on TruthfulQA and BioGEN.

Self-Alignment for Factuality: Mitigating Hallucinations in LLMs via Self-Evaluation

acl-long.107

Poster

2402.09267

https://aclanthology.org/2024.acl-long.107/

https://aclanthology.org/2024.acl-long.108.bib

@inproceedings{wang-etal-2024-rag, title = "{M}-{RAG}: Reinforcing Large Language Model Performance through Retrieval-Augmented Generation with Multiple Partitions", author = "Wang, Zheng and Teo, Shu and Ouyang, Jieer and Xu, Yongjun and Shi, Wei", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.108", pages = "1966--1978", abstract = "Retrieval-Augmented Generation (RAG) enhances Large Language Models (LLMs) by retrieving relevant memories from an external database. However, existing RAG methods typically organize all memories in a whole database, potentially limiting focus on crucial memories and introducing noise. In this paper, we introduce a multiple partition paradigm for RAG (called M-RAG), where each database partition serves as a basic unit for RAG execution. Based on this paradigm, we propose a novel framework that leverages LLMs with Multi-Agent Reinforcement Learning to optimize different language generation tasks explicitly. Through comprehensive experiments conducted on seven datasets, spanning three language generation tasks and involving three distinct language model architectures, we confirm that M-RAG consistently outperforms various baseline methods, achieving improvements of 11{\%}, 8{\%}, and 12{\%} for text summarization, machine translation, and dialogue generation, respectively.", }

Retrieval-Augmented Generation (RAG) enhances Large Language Models (LLMs) by retrieving relevant memories from an external database. However, existing RAG methods typically organize all memories in a whole database, potentially limiting focus on crucial memories and introducing noise. In this paper, we introduce a multiple partition paradigm for RAG (called M-RAG), where each database partition serves as a basic unit for RAG execution. Based on this paradigm, we propose a novel framework that leverages LLMs with Multi-Agent Reinforcement Learning to optimize different language generation tasks explicitly. Through comprehensive experiments conducted on seven datasets, spanning three language generation tasks and involving three distinct language model architectures, we confirm that M-RAG consistently outperforms various baseline methods, achieving improvements of 11{\%}, 8{\%}, and 12{\%} for text summarization, machine translation, and dialogue generation, respectively.

M-RAG: Reinforcing Large Language Model Performance through Retrieval-Augmented Generation with Multiple Partitions

acl-long.108

Poster

2405.16420

https://aclanthology.org/2024.acl-long.108/

https://aclanthology.org/2024.acl-long.109.bib

@inproceedings{yang-etal-2024-air, title = "{AIR}-Bench: Benchmarking Large Audio-Language Models via Generative Comprehension", author = "Yang, Qian and Xu, Jin and Liu, Wenrui and Chu, Yunfei and Jiang, Ziyue and Zhou, Xiaohuan and Leng, Yichong and Lv, Yuanjun and Zhao, Zhou and Zhou, Chang and Zhou, Jingren", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.109", pages = "1979--1998", abstract = "Recently, instruction-following audio-language models have received broad attention for human-audio interaction. However, the absence of benchmarks capable of evaluating audio-centric interaction capabilities has impeded advancements in this field. Previous models primarily focus on assessing different fundamental tasks, such as automatic speech recognition, and lack an assessment of the open-ended generative capabilities centered around audio. Thus, it is challenging to track the progression in the Large Audio-Language Models (LALMs) domain and to provide guidance for future improvement.In this paper, we introduce AIR-Bench (Audio InstRuction Benchmark), the first benchmark designed to evaluate the ability of LALMs to understand various types of audio signals (including human speech, natural sounds, and music), and furthermore, to interact with humans in the textual format. AIR-Bench encompasses two dimensions: foundation and chat benchmarks. The former consists of 19 tasks with approximately 19k single-choice questions, intending to inspect the basic single-task ability of LALMs. The latter one contains 2k instances of open-ended question-and-answer data, directly assessing the comprehension of the model on complex audio and its capacity to follow instructions. Both benchmarks require the model to generate hypotheses directly. We design a unified framework that leverages advanced language models, such as GPT-4, to evaluate the scores of generated hypotheses given the meta-information of the audio. Experimental results demonstrate a high level of consistency between GPT-4-based evaluation and human evaluation. By revealing the limitations of existing LALMs through evaluation results, AIR-Bench can provide insights into the direction of future research. Dataset and evaluation code are available at https://github.com/OFA-Sys/AIR-Bench.", }

Recently, instruction-following audio-language models have received broad attention for human-audio interaction. However, the absence of benchmarks capable of evaluating audio-centric interaction capabilities has impeded advancements in this field. Previous models primarily focus on assessing different fundamental tasks, such as automatic speech recognition, and lack an assessment of the open-ended generative capabilities centered around audio. Thus, it is challenging to track the progression in the Large Audio-Language Models (LALMs) domain and to provide guidance for future improvement.In this paper, we introduce AIR-Bench (Audio InstRuction Benchmark), the first benchmark designed to evaluate the ability of LALMs to understand various types of audio signals (including human speech, natural sounds, and music), and furthermore, to interact with humans in the textual format. AIR-Bench encompasses two dimensions: foundation and chat benchmarks. The former consists of 19 tasks with approximately 19k single-choice questions, intending to inspect the basic single-task ability of LALMs. The latter one contains 2k instances of open-ended question-and-answer data, directly assessing the comprehension of the model on complex audio and its capacity to follow instructions. Both benchmarks require the model to generate hypotheses directly. We design a unified framework that leverages advanced language models, such as GPT-4, to evaluate the scores of generated hypotheses given the meta-information of the audio. Experimental results demonstrate a high level of consistency between GPT-4-based evaluation and human evaluation. By revealing the limitations of existing LALMs through evaluation results, AIR-Bench can provide insights into the direction of future research. Dataset and evaluation code are available at https://github.com/OFA-Sys/AIR-Bench.

AIR-Bench: Benchmarking Large Audio-Language Models via Generative Comprehension

acl-long.109

Poster

2402.07729

https://huggingface.co/papers/2402.07729

https://aclanthology.org/2024.acl-long.109/

https://aclanthology.org/2024.acl-long.110.bib

@inproceedings{kocmi-etal-2024-navigating, title = "Navigating the Metrics Maze: Reconciling Score Magnitudes and Accuracies", author = "Kocmi, Tom and Zouhar, Vil{\'e}m and Federmann, Christian and Post, Matt", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.110", pages = "1999--2014", abstract = "Ten years ago a single metric, BLEU, governed progress in machine translation research. For better or worse, there is no such consensus today, and consequently it is difficult for researchers to develop and retain intuitions about metric deltas that drove earlier research and deployment decisions. This paper investigates the {``}dynamic range{''} of a number of modern metrics in an effort to provide a collective understanding of the meaning of differences in scores both within and among metrics; in other words, we ask {``}what point difference x in metric y is required between two systems for humans to notice?{''}. We conduct our evaluation on a new large dataset, ToShip23, using it to discover deltas at which metrics achieve system-level differences that are meaningful to humans, which we measure by pairwise system accuracy. We additionally show that this method of establishing delta-accuracy is more stable than the standard use of statistical p-values in regards to testset size. Where data size permits, we also explore the effect of metric deltas and accuracy across finer-grained features such as translation direction, domain, and system closeness.", }

Ten years ago a single metric, BLEU, governed progress in machine translation research. For better or worse, there is no such consensus today, and consequently it is difficult for researchers to develop and retain intuitions about metric deltas that drove earlier research and deployment decisions. This paper investigates the {``}dynamic range{''} of a number of modern metrics in an effort to provide a collective understanding of the meaning of differences in scores both within and among metrics; in other words, we ask {``}what point difference x in metric y is required between two systems for humans to notice?{''}. We conduct our evaluation on a new large dataset, ToShip23, using it to discover deltas at which metrics achieve system-level differences that are meaningful to humans, which we measure by pairwise system accuracy. We additionally show that this method of establishing delta-accuracy is more stable than the standard use of statistical p-values in regards to testset size. Where data size permits, we also explore the effect of metric deltas and accuracy across finer-grained features such as translation direction, domain, and system closeness.

Navigating the Metrics Maze: Reconciling Score Magnitudes and Accuracies

acl-long.110

Poster

2401.06760

https://aclanthology.org/2024.acl-long.110/

https://aclanthology.org/2024.acl-long.111.bib

@inproceedings{ren-etal-2024-valuebench, title = "{V}alue{B}ench: Towards Comprehensively Evaluating Value Orientations and Understanding of Large Language Models", author = "Ren, Yuanyi and Ye, Haoran and Fang, Hanjun and Zhang, Xin and Song, Guojie", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.111", pages = "2015--2040", abstract = "Large Language Models (LLMs) are transforming diverse fields and gaining increasing influence as human proxies. This development underscores the urgent need for evaluating value orientations and understanding of LLMs to ensure their responsible integration into public-facing applications. This work introduces ValueBench, the first comprehensive psychometric benchmark for evaluating value orientations and understanding in LLMs. ValueBench collects data from 44 established psychometric inventories, encompassing 453 multifaceted value dimensions. We propose an evaluation pipeline grounded in realistic human-AI interactions to probe value orientations, along with novel tasks for evaluating value understanding in an open-ended value space. With extensive experiments conducted on six representative LLMs, we unveil their shared and distinctive value orientations and exhibit their ability to approximate expert conclusions in value-related extraction and generation tasks.", }

Large Language Models (LLMs) are transforming diverse fields and gaining increasing influence as human proxies. This development underscores the urgent need for evaluating value orientations and understanding of LLMs to ensure their responsible integration into public-facing applications. This work introduces ValueBench, the first comprehensive psychometric benchmark for evaluating value orientations and understanding in LLMs. ValueBench collects data from 44 established psychometric inventories, encompassing 453 multifaceted value dimensions. We propose an evaluation pipeline grounded in realistic human-AI interactions to probe value orientations, along with novel tasks for evaluating value understanding in an open-ended value space. With extensive experiments conducted on six representative LLMs, we unveil their shared and distinctive value orientations and exhibit their ability to approximate expert conclusions in value-related extraction and generation tasks.

ValueBench: Towards Comprehensively Evaluating Value Orientations and Understanding of Large Language Models

acl-long.111

Poster

2406.04214

https://aclanthology.org/2024.acl-long.111/

https://aclanthology.org/2024.acl-long.112.bib

@inproceedings{hu-xu-2024-dm, title = "{DM}-{BLI}: Dynamic Multiple Subspaces Alignment for Unsupervised Bilingual Lexicon Induction", author = "Hu, Ling and Xu, Yuemei", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.112", pages = "2041--2052", abstract = "Unsupervised bilingual lexicon induction (BLI) task aims to find word translations between languages and has achieved great success in similar language pairs. However, related works mostly rely on a single linear mapping for language alignment and fail on distant or low-resource language pairs, achieving less than half the performance observed in rich-resource language pairs. In this paper, we introduce DM-BLI, a Dynamic Multiple subspaces alignment framework for unsupervised BLI. DM-BLI improves language alignment by utilizing multiple subspace alignments instead of a single mapping. We begin via unsupervised clustering to discover these subspaces in source embedding space. Then we identify and align corresponding subspaces in the target space using a rough global alignment. DM-BLI further employs intra-cluster and inter-cluster contrastive learning to refine precise alignment for each subspace pair. Experiments conducted on standard BLI datasets for 12 language pairs (6 rich-resource and 6 low-resource) demonstrate substantial gains achieved by our framework. We release our code at https://github.com/huling-2/DM-BLI.git.", }

Unsupervised bilingual lexicon induction (BLI) task aims to find word translations between languages and has achieved great success in similar language pairs. However, related works mostly rely on a single linear mapping for language alignment and fail on distant or low-resource language pairs, achieving less than half the performance observed in rich-resource language pairs. In this paper, we introduce DM-BLI, a Dynamic Multiple subspaces alignment framework for unsupervised BLI. DM-BLI improves language alignment by utilizing multiple subspace alignments instead of a single mapping. We begin via unsupervised clustering to discover these subspaces in source embedding space. Then we identify and align corresponding subspaces in the target space using a rough global alignment. DM-BLI further employs intra-cluster and inter-cluster contrastive learning to refine precise alignment for each subspace pair. Experiments conducted on standard BLI datasets for 12 language pairs (6 rich-resource and 6 low-resource) demonstrate substantial gains achieved by our framework. We release our code at https://github.com/huling-2/DM-BLI.git.

DM-BLI: Dynamic Multiple Subspaces Alignment for Unsupervised Bilingual Lexicon Induction

acl-long.112

Poster

https://aclanthology.org/2024.acl-long.112/

https://aclanthology.org/2024.acl-long.113.bib

@inproceedings{solano-etal-2024-sparsefit, title = "{S}parse{F}it: Few-shot Prompting with Sparse Fine-tuning for Jointly Generating Predictions and Natural Language Explanations", author = "Solano, Jesus and Sanni, Mardhiyah and Camburu, Oana-Maria and Minervini, Pasquale", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.113", pages = "2053--2077", abstract = "Models that generate natural language explanations (NLEs) for their predictions have recently gained increasing interest. However, this approach usually demands large datasets of human-written NLEs for the ground-truth answers at training time, which can be expensive and potentially infeasible for some applications. When only a few NLEs are available (a few-shot setup), fine-tuning pre-trained language models (PLMs) in conjunction with prompt-based learning has recently shown promising results. However, PLMs typically have billions of parameters, making full fine-tuning expensive. We propose SparseFit, a sparse few-shot fine-tuning strategy that leverages discrete prompts to jointly generate predictions and NLEs. We experiment with SparseFit on three sizes of the T5 language model and four datasets and compare it against existing state-of-the-art Parameter-Efficient Fine-Tuning (PEFT) techniques. We find that fine-tuning only 6.8{\%} of the model parameters leads to competitive results for both the task performance and the quality of the generated NLEs compared to full fine-tuning of the model and produces better results on average than other PEFT methods in terms of predictive accuracy and NLE quality.", }

Models that generate natural language explanations (NLEs) for their predictions have recently gained increasing interest. However, this approach usually demands large datasets of human-written NLEs for the ground-truth answers at training time, which can be expensive and potentially infeasible for some applications. When only a few NLEs are available (a few-shot setup), fine-tuning pre-trained language models (PLMs) in conjunction with prompt-based learning has recently shown promising results. However, PLMs typically have billions of parameters, making full fine-tuning expensive. We propose SparseFit, a sparse few-shot fine-tuning strategy that leverages discrete prompts to jointly generate predictions and NLEs. We experiment with SparseFit on three sizes of the T5 language model and four datasets and compare it against existing state-of-the-art Parameter-Efficient Fine-Tuning (PEFT) techniques. We find that fine-tuning only 6.8{\%} of the model parameters leads to competitive results for both the task performance and the quality of the generated NLEs compared to full fine-tuning of the model and produces better results on average than other PEFT methods in terms of predictive accuracy and NLE quality.

SparseFit: Few-shot Prompting with Sparse Fine-tuning for Jointly Generating Predictions and Natural Language Explanations

acl-long.113

Poster

2305.13235

https://huggingface.co/papers/2305.13235

https://aclanthology.org/2024.acl-long.113/

https://aclanthology.org/2024.acl-long.114.bib

@inproceedings{wu-etal-2024-handling, title = "Handling Ambiguity in Emotion: From Out-of-Domain Detection to Distribution Estimation", author = "Wu, Wen and Li, Bo and Zhang, Chao and Chiu, Chung-Cheng and Li, Qiujia and Bai, Junwen and Sainath, Tara and Woodland, Phil", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.114", pages = "2078--2093", abstract = "The subjective perception of emotion leads to inconsistent labels from human annotators. Typically, utterances lacking majority-agreed labels are excluded when training an emotion classifier, which cause problems when encountering ambiguous emotional expressions during testing. This paper investigates three methods to handle ambiguous emotion. First, we show that incorporating utterances without majority-agreed labels as an additional class in the classifier reduces the classification performance of the other emotion classes. Then, we propose detecting utterances with ambiguous emotions as out-of-domain samples by quantifying the uncertainty in emotion classification using evidential deep learning. This approach retains the classification accuracy while effectively detects ambiguous emotion expressions. Furthermore, to obtain fine-grained distinctions among ambiguous emotions, we propose representing emotion as a distribution instead of a single class label. The task is thus re-framed from classification to distribution estimation where every individual annotation is taken into account, not just the majority opinion. The evidential uncertainty measure is extended to quantify the uncertainty in emotion distribution estimation. Experimental results on the IEMOCAP and CREMA-D datasets demonstrate the superior capability of the proposed method in terms of majority class prediction, emotion distribution estimation, and uncertainty estimation.", }

The subjective perception of emotion leads to inconsistent labels from human annotators. Typically, utterances lacking majority-agreed labels are excluded when training an emotion classifier, which cause problems when encountering ambiguous emotional expressions during testing. This paper investigates three methods to handle ambiguous emotion. First, we show that incorporating utterances without majority-agreed labels as an additional class in the classifier reduces the classification performance of the other emotion classes. Then, we propose detecting utterances with ambiguous emotions as out-of-domain samples by quantifying the uncertainty in emotion classification using evidential deep learning. This approach retains the classification accuracy while effectively detects ambiguous emotion expressions. Furthermore, to obtain fine-grained distinctions among ambiguous emotions, we propose representing emotion as a distribution instead of a single class label. The task is thus re-framed from classification to distribution estimation where every individual annotation is taken into account, not just the majority opinion. The evidential uncertainty measure is extended to quantify the uncertainty in emotion distribution estimation. Experimental results on the IEMOCAP and CREMA-D datasets demonstrate the superior capability of the proposed method in terms of majority class prediction, emotion distribution estimation, and uncertainty estimation.

Handling Ambiguity in Emotion: From Out-of-Domain Detection to Distribution Estimation

acl-long.114

Poster

2402.12862

https://aclanthology.org/2024.acl-long.114/

https://aclanthology.org/2024.acl-long.115.bib

@inproceedings{fang-etal-2024-reano, title = "{REANO}: Optimising Retrieval-Augmented Reader Models through Knowledge Graph Generation", author = "Fang, Jinyuan and Meng, Zaiqiao and MacDonald, Craig", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.115", pages = "2094--2112", abstract = "Open domain question answering (ODQA) aims to answer questions with knowledge from an external corpus. Fusion-in-Decoder (FiD) is an effective retrieval-augmented reader model to address this task. Given that FiD independently encodes passages, which overlooks the semantic relationships between passages, some studies use knowledge graphs (KGs) to establish dependencies among passages. However, they only leverage knowledge triples from existing KGs, which suffer from incompleteness and may lack certain information critical for answering given questions. To this end, in order to capture the dependencies between passages while tacking the issue of incompleteness in existing KGs, we propose to enhance the retrieval-augmented reader model with a knowledge graph generation module (REANO). Specifically, REANO consists of a KG generator and an answer predictor. The KG generator aims to generate KGs from the passages and the answer predictor then generates answers based on the passages and the generated KGs. Experimental results on five ODQA datasets indicate that compared with baselines, REANO can improve the exact match score by up to 2.7{\%} on the EntityQuestion dataset, with an average improvement of 1.8{\%} across all the datasets.", }

Open domain question answering (ODQA) aims to answer questions with knowledge from an external corpus. Fusion-in-Decoder (FiD) is an effective retrieval-augmented reader model to address this task. Given that FiD independently encodes passages, which overlooks the semantic relationships between passages, some studies use knowledge graphs (KGs) to establish dependencies among passages. However, they only leverage knowledge triples from existing KGs, which suffer from incompleteness and may lack certain information critical for answering given questions. To this end, in order to capture the dependencies between passages while tacking the issue of incompleteness in existing KGs, we propose to enhance the retrieval-augmented reader model with a knowledge graph generation module (REANO). Specifically, REANO consists of a KG generator and an answer predictor. The KG generator aims to generate KGs from the passages and the answer predictor then generates answers based on the passages and the generated KGs. Experimental results on five ODQA datasets indicate that compared with baselines, REANO can improve the exact match score by up to 2.7{\%} on the EntityQuestion dataset, with an average improvement of 1.8{\%} across all the datasets.

REANO: Optimising Retrieval-Augmented Reader Models through Knowledge Graph Generation

acl-long.115

Poster

https://aclanthology.org/2024.acl-long.115/

https://aclanthology.org/2024.acl-long.116.bib

@inproceedings{zhang-etal-2024-learning, title = "Learning Disentangled Semantic Spaces of Explanations via Invertible Neural Networks", author = "Zhang, Yingji and Carvalho, Danilo and Freitas, Andre", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.116", pages = "2113--2134", abstract = "Disentangled latent spaces usually have better semantic separability and geometrical properties, which leads to better interpretability and more controllable data generation. While this has been well investigated in Computer Vision, in tasks such as image disentanglement, in the NLP domain, sentence disentanglement is still comparatively under-investigated. Most previous work have concentrated on disentangling task-specific generative factors, such as sentiment, within the context of style transfer. In this work, we focus on a more general form of sentence disentanglement, targeting the localised modification and control of more general sentence semantic features. To achieve this, we contribute to a novel notion of sentence semantic disentanglement and introduce a flow-based invertible neural network (INN) mechanism integrated with a transformer-based language Autoencoder (AE) in order to deliver latent spaces with better separability properties. Experimental results demonstrate that the model can conform the distributed latent space into a better semantically disentangled sentence space, leading to improved language interpretability and controlled generation when compared to the recent state-of-the-art language VAE models.", }

Disentangled latent spaces usually have better semantic separability and geometrical properties, which leads to better interpretability and more controllable data generation. While this has been well investigated in Computer Vision, in tasks such as image disentanglement, in the NLP domain, sentence disentanglement is still comparatively under-investigated. Most previous work have concentrated on disentangling task-specific generative factors, such as sentiment, within the context of style transfer. In this work, we focus on a more general form of sentence disentanglement, targeting the localised modification and control of more general sentence semantic features. To achieve this, we contribute to a novel notion of sentence semantic disentanglement and introduce a flow-based invertible neural network (INN) mechanism integrated with a transformer-based language Autoencoder (AE) in order to deliver latent spaces with better separability properties. Experimental results demonstrate that the model can conform the distributed latent space into a better semantically disentangled sentence space, leading to improved language interpretability and controlled generation when compared to the recent state-of-the-art language VAE models.

Learning Disentangled Semantic Spaces of Explanations via Invertible Neural Networks

acl-long.116

Poster

2305.01713

https://aclanthology.org/2024.acl-long.116/

https://aclanthology.org/2024.acl-long.117.bib

@inproceedings{ma-etal-2024-mops, title = "{M}o{PS}: Modular Story Premise Synthesis for Open-Ended Automatic Story Generation", author = "Ma, Yan and Qiao, Yu and Liu, Pengfei", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.117", pages = "2135--2169", abstract = "A story premise succinctly defines a story{'}s main idea, foundation, and trajectory. It serves as the initial trigger in automatic story generation. Existing sources of story premises are limited by a lack of diversity, uneven quality, and high costs that make them difficult to scale. In response, we introduce Modular Story Premise Synthesis (MoPS) which breaks down story premises into modules like background and persona for automated design and generation. MoPS consists of three phases: (1) Pre-collect a consistent set of candidates for each module to form a nested dictionary. (2) Extract a key path from the nested dictionary as the premise design. (3) Instruct an LLM to integrate the design into a coherent premise sentence. Thorough evaluations demonstrate that our synthesized premises excel in diversity, fascination, completeness, and originality compared to those induced from large language models and captured from public story datasets. Similarly, the extended novels and scripts generated from our premises also exhibit higher quality. In supplementary materials, we provide the MoPS code suite, along with 7.5k generated premises and 1k extended stories.", }

A story premise succinctly defines a story{'}s main idea, foundation, and trajectory. It serves as the initial trigger in automatic story generation. Existing sources of story premises are limited by a lack of diversity, uneven quality, and high costs that make them difficult to scale. In response, we introduce Modular Story Premise Synthesis (MoPS) which breaks down story premises into modules like background and persona for automated design and generation. MoPS consists of three phases: (1) Pre-collect a consistent set of candidates for each module to form a nested dictionary. (2) Extract a key path from the nested dictionary as the premise design. (3) Instruct an LLM to integrate the design into a coherent premise sentence. Thorough evaluations demonstrate that our synthesized premises excel in diversity, fascination, completeness, and originality compared to those induced from large language models and captured from public story datasets. Similarly, the extended novels and scripts generated from our premises also exhibit higher quality. In supplementary materials, we provide the MoPS code suite, along with 7.5k generated premises and 1k extended stories.

MoPS: Modular Story Premise Synthesis for Open-Ended Automatic Story Generation

acl-long.117

Poster

2406.05690

https://huggingface.co/papers/2406.05690

https://aclanthology.org/2024.acl-long.117/

https://aclanthology.org/2024.acl-long.118.bib

@inproceedings{chen-etal-2024-open, title = "Open-Set Semi-Supervised Text Classification via Adversarial Disagreement Maximization", author = "Chen, Junfan and Zhang, Richong and Chen, Junchi and Hu, Chunming", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.118", pages = "2170--2180", abstract = "Open-Set Semi-Supervised Text Classification (OSTC) aims to train a classification model on a limited set of labeled texts, alongside plenty of unlabeled texts that include both in-distribution and out-of-distribution examples. In this paper, we revisit the main challenge in OSTC, i.e., outlier detection, from a measurement disagreement perspective and innovatively propose to improve OSTC performance by directly maximizing the measurement disagreements. Based on the properties of in-measurement and cross-measurements, we design an Adversarial Disagreement Maximization (ADM) model that synergeticly optimizes the measurement disagreements. In addition, we develop an abnormal example detection and measurement calibration approach to guarantee the effectiveness of ADM training. Experiment results and comprehensive analysis of three benchmarks demonstrate the effectiveness of our model.", }

Open-Set Semi-Supervised Text Classification (OSTC) aims to train a classification model on a limited set of labeled texts, alongside plenty of unlabeled texts that include both in-distribution and out-of-distribution examples. In this paper, we revisit the main challenge in OSTC, i.e., outlier detection, from a measurement disagreement perspective and innovatively propose to improve OSTC performance by directly maximizing the measurement disagreements. Based on the properties of in-measurement and cross-measurements, we design an Adversarial Disagreement Maximization (ADM) model that synergeticly optimizes the measurement disagreements. In addition, we develop an abnormal example detection and measurement calibration approach to guarantee the effectiveness of ADM training. Experiment results and comprehensive analysis of three benchmarks demonstrate the effectiveness of our model.

Open-Set Semi-Supervised Text Classification via Adversarial Disagreement Maximization

acl-long.118

Poster

https://aclanthology.org/2024.acl-long.118/

https://aclanthology.org/2024.acl-long.119.bib

@inproceedings{ye-etal-2024-toolsword, title = "{T}ool{S}word: Unveiling Safety Issues of Large Language Models in Tool Learning Across Three Stages", author = "Ye, Junjie and Li, Sixian and Li, Guanyu and Huang, Caishuang and Gao, Songyang and Wu, Yilong and Zhang, Qi and Gui, Tao and Huang, Xuanjing", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.119", pages = "2181--2211", abstract = "Tool learning is widely acknowledged as a foundational approach or deploying large language models (LLMs) in real-world scenarios. While current research primarily emphasizes leveraging tools to augment LLMs, it frequently neglects emerging safety considerations tied to their application. To fill this gap, we present $ToolSword$, a comprehensive framework dedicated to meticulously investigating safety issues linked to LLMs in tool learning. Specifically, ToolSword delineates six safety scenarios for LLMs in tool learning, encompassing $malicious$ $queries$ and $jailbreak$ $attacks$ in the input stage, $noisy$ $misdirection$ and $risky$ $cues$ in the execution stage, and $harmful$ $feedback$ and $error$ $conflicts$ in the output stage. Experiments conducted on 11 open-source and closed-source LLMs reveal enduring safety challenges in tool learning, such as handling harmful queries, employing risky tools, and delivering detrimental feedback, which even GPT-4 is susceptible to. Moreover, we conduct further studies with the aim of fostering research on tool learning safety. The data will be released upon acceptance of the paper.", }

Tool learning is widely acknowledged as a foundational approach or deploying large language models (LLMs) in real-world scenarios. While current research primarily emphasizes leveraging tools to augment LLMs, it frequently neglects emerging safety considerations tied to their application. To fill this gap, we present $ToolSword$, a comprehensive framework dedicated to meticulously investigating safety issues linked to LLMs in tool learning. Specifically, ToolSword delineates six safety scenarios for LLMs in tool learning, encompassing $malicious$ $queries$ and $jailbreak$ $attacks$ in the input stage, $noisy$ $misdirection$ and $risky$ $cues$ in the execution stage, and $harmful$ $feedback$ and $error$ $conflicts$ in the output stage. Experiments conducted on 11 open-source and closed-source LLMs reveal enduring safety challenges in tool learning, such as handling harmful queries, employing risky tools, and delivering detrimental feedback, which even GPT-4 is susceptible to. Moreover, we conduct further studies with the aim of fostering research on tool learning safety. The data will be released upon acceptance of the paper.

ToolSword: Unveiling Safety Issues of Large Language Models in Tool Learning Across Three Stages

acl-long.119

Poster

2402.10753

https://aclanthology.org/2024.acl-long.119/

https://aclanthology.org/2024.acl-long.120.bib

@inproceedings{hosseini-etal-2024-synthetic, title = "A synthetic data approach for domain generalization of {NLI} models", author = "Hosseini, Mohammad Javad and Petrov, Andrey and Fabrikant, Alex and Louis, Annie", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.120", pages = "2212--2226", abstract = "Natural Language Inference (NLI) remains an important benchmark task for LLMs. NLI datasets are a springboard for transfer learning to other semantic tasks, and NLI models are standard tools for identifying the faithfulness of model-generated text. There are several large scale NLI datasets today, and models have improved greatly by hill-climbing on these collections. Yet their realistic performance on out-of-distribution/domain data is less well-understood. We explore the opportunity for synthetic high-quality datasets to adapt NLI models for zero-shot use in downstream applications across new and unseen text domains. We demonstrate a new approach for generating NLI data in diverse domains and lengths, so far not covered by existing training sets. The resulting examples have meaningful premises, the hypotheses are formed in creative ways rather than simple edits to a few premise tokens, and the labels have high accuracy. We show that models trained on this data (685K synthetic examples) have the best generalization to completely new downstream test settings. On the TRUE benchmark, a T5-small model trained with our data improves around 7{\%} on average compared to training on the best alternative dataset. The improvements are more pronounced for smaller models, while still meaningful on a T5 XXL model. We also demonstrate gains on test sets when in-domain training data is augmented with our domain-general synthetic data.", }

Natural Language Inference (NLI) remains an important benchmark task for LLMs. NLI datasets are a springboard for transfer learning to other semantic tasks, and NLI models are standard tools for identifying the faithfulness of model-generated text. There are several large scale NLI datasets today, and models have improved greatly by hill-climbing on these collections. Yet their realistic performance on out-of-distribution/domain data is less well-understood. We explore the opportunity for synthetic high-quality datasets to adapt NLI models for zero-shot use in downstream applications across new and unseen text domains. We demonstrate a new approach for generating NLI data in diverse domains and lengths, so far not covered by existing training sets. The resulting examples have meaningful premises, the hypotheses are formed in creative ways rather than simple edits to a few premise tokens, and the labels have high accuracy. We show that models trained on this data (685K synthetic examples) have the best generalization to completely new downstream test settings. On the TRUE benchmark, a T5-small model trained with our data improves around 7{\%} on average compared to training on the best alternative dataset. The improvements are more pronounced for smaller models, while still meaningful on a T5 XXL model. We also demonstrate gains on test sets when in-domain training data is augmented with our domain-general synthetic data.

A synthetic data approach for domain generalization of NLI models

acl-long.120

Poster

2402.12368

https://aclanthology.org/2024.acl-long.120/

https://aclanthology.org/2024.acl-long.121.bib

@inproceedings{wu-etal-2024-enhancing, title = "Enhancing Contrastive Learning with Noise-Guided Attack: Towards Continual Relation Extraction in the Wild", author = "Wu, Ting and Liu, Jingyi and Zheng, Rui and Gui, Tao and Zhang, Qi and Huang, Xuanjing", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.121", pages = "2227--2239", abstract = "The principle of continual relation extraction (CRE) involves adapting to emerging novel relations while preserving old knowledge. Existing CRE approaches excel in preserving old knowledge but falter when confronted with contaminated data streams, likely due to an artificial assumption of no annotation errors. Recognizing the prevalence of noisy labels in real-world datasets, we introduce a more practical learning scenario, termed as \textit{noisy-CRE}. In response to this challenge, we propose a noise-resistant contrastive framework called Noise-guided Attack in Contrastive Learning (NaCL), aimed at learning incremental corrupted relations. Diverging from conventional approaches like sample discarding or relabeling in the presence of noisy labels, NaCL takes a transformative route by modifying the feature space through targeted attack. This attack aims to align the feature space with the provided, albeit inaccurate, labels, thereby enhancing contrastive representations. Extensive empirical validations demonstrate the consistent performance improvement of NaCL with increasing noise rates, surpassing state-of-the-art methods.", }

The principle of continual relation extraction (CRE) involves adapting to emerging novel relations while preserving old knowledge. Existing CRE approaches excel in preserving old knowledge but falter when confronted with contaminated data streams, likely due to an artificial assumption of no annotation errors. Recognizing the prevalence of noisy labels in real-world datasets, we introduce a more practical learning scenario, termed as \textit{noisy-CRE}. In response to this challenge, we propose a noise-resistant contrastive framework called Noise-guided Attack in Contrastive Learning (NaCL), aimed at learning incremental corrupted relations. Diverging from conventional approaches like sample discarding or relabeling in the presence of noisy labels, NaCL takes a transformative route by modifying the feature space through targeted attack. This attack aims to align the feature space with the provided, albeit inaccurate, labels, thereby enhancing contrastive representations. Extensive empirical validations demonstrate the consistent performance improvement of NaCL with increasing noise rates, surpassing state-of-the-art methods.

Enhancing Contrastive Learning with Noise-Guided Attack: Towards Continual Relation Extraction in the Wild

acl-long.121

Poster

2305.07085

https://aclanthology.org/2024.acl-long.121/

https://aclanthology.org/2024.acl-long.122.bib

@inproceedings{zhao-etal-2024-lrquant, title = "{LRQ}uant: Learnable and Robust Post-Training Quantization for Large Language Models", author = "Zhao, Jiaqi and Zhang, Miao and Zeng, Chao and Wang, Ming and Liu, Xuebo and Nie, Liqiang", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.122", pages = "2240--2255", abstract = "Post-training quantization (PTQ) for large language models (LLMs) significantly accelerates model inference and relieves memory constraints, without incurring model training. A {``}smoothing paradigm{''} is commonly used in LLM quantization, which transfers the quantization difficulty of activation to weight quantization using mathematically equivalent transformations. However, existing methods face two issues: 1) Most smoothing parameters are hand-crafted defined which leads to suboptimal results; 2) There are significant performance degradations when tested on unseen datasets. To address these challenges, this paper introduces a robust learnable smooth-based PTQ framework, called LRQuant. Firstly, we consider a learnable paradigm to find optimal smoothing parameters which are initialized by logarithmic activation equivalent. In addition, we empirically found that only relying on MSE loss could hardly lead to optimal quantization results, and we then propose a novel loss function based on the negative logarithm of cosine similarity (NLC loss) between outputs of full-precision and quantized block. At last, we pioneeringly introduce Test-time adaptation (TTA) into LLM quantization, which allows for rapid model adaptation during testing to improve generalization performance. More surprisingly, we find that by using our TTA method, we can achieve better results on test sets than directly using test sets for calibration in some cases while avoiding catastrophic forgetting. Codes are available at https://github.com/zjq0455/RLQ.", }

Post-training quantization (PTQ) for large language models (LLMs) significantly accelerates model inference and relieves memory constraints, without incurring model training. A {``}smoothing paradigm{''} is commonly used in LLM quantization, which transfers the quantization difficulty of activation to weight quantization using mathematically equivalent transformations. However, existing methods face two issues: 1) Most smoothing parameters are hand-crafted defined which leads to suboptimal results; 2) There are significant performance degradations when tested on unseen datasets. To address these challenges, this paper introduces a robust learnable smooth-based PTQ framework, called LRQuant. Firstly, we consider a learnable paradigm to find optimal smoothing parameters which are initialized by logarithmic activation equivalent. In addition, we empirically found that only relying on MSE loss could hardly lead to optimal quantization results, and we then propose a novel loss function based on the negative logarithm of cosine similarity (NLC loss) between outputs of full-precision and quantized block. At last, we pioneeringly introduce Test-time adaptation (TTA) into LLM quantization, which allows for rapid model adaptation during testing to improve generalization performance. More surprisingly, we find that by using our TTA method, we can achieve better results on test sets than directly using test sets for calibration in some cases while avoiding catastrophic forgetting. Codes are available at https://github.com/zjq0455/RLQ.

LRQuant: Learnable and Robust Post-Training Quantization for Large Language Models

acl-long.122

Oral

https://aclanthology.org/2024.acl-long.122/

https://aclanthology.org/2024.acl-long.123.bib

@inproceedings{weber-genzel-etal-2024-varierr, title = "{V}ari{E}rr {NLI}: Separating Annotation Error from Human Label Variation", author = "Weber-Genzel, Leon and Peng, Siyao and De Marneffe, Marie-Catherine and Plank, Barbara", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.123", pages = "2256--2269", abstract = "Human label variation arises when annotators assign different labels to the same item for valid reasons, while annotation errors occur when labels are assigned for invalid reasons. These two issues are prevalent in NLP benchmarks, yet existing research has studied them in isolation. To the best of our knowledge, there exists no prior work that focuses on teasing apart error from signal, especially in cases where signal is beyond black-and-white.To fill this gap, we introduce a systematic methodology and a new dataset, VariErr (variation versus error), focusing on the NLI task in English. We propose a 2-round annotation procedure with annotators explaining each label and subsequently judging the validity of label-explanation pairs.VariErr contains 7,732 validity judgments on 1,933 explanations for 500 re-annotated MNLI items. We assess the effectiveness of various automatic error detection (AED) methods and GPTs in uncovering errors versus human label variation. We find that state-of-the-art AED methods significantly underperform GPTs and humans. While GPT-4 is the best system, it still falls short of human performance. Our methodology is applicable beyond NLI, offering fertile ground for future research on error versus plausible variation, which in turn can yield better and more trustworthy NLP systems.", }

Human label variation arises when annotators assign different labels to the same item for valid reasons, while annotation errors occur when labels are assigned for invalid reasons. These two issues are prevalent in NLP benchmarks, yet existing research has studied them in isolation. To the best of our knowledge, there exists no prior work that focuses on teasing apart error from signal, especially in cases where signal is beyond black-and-white.To fill this gap, we introduce a systematic methodology and a new dataset, VariErr (variation versus error), focusing on the NLI task in English. We propose a 2-round annotation procedure with annotators explaining each label and subsequently judging the validity of label-explanation pairs.VariErr contains 7,732 validity judgments on 1,933 explanations for 500 re-annotated MNLI items. We assess the effectiveness of various automatic error detection (AED) methods and GPTs in uncovering errors versus human label variation. We find that state-of-the-art AED methods significantly underperform GPTs and humans. While GPT-4 is the best system, it still falls short of human performance. Our methodology is applicable beyond NLI, offering fertile ground for future research on error versus plausible variation, which in turn can yield better and more trustworthy NLP systems.

VariErr NLI: Separating Annotation Error from Human Label Variation

acl-long.123

Oral

2403.01931

https://aclanthology.org/2024.acl-long.123/

https://aclanthology.org/2024.acl-long.124.bib

@inproceedings{yin-etal-2024-benchmarking, title = "Benchmarking Knowledge Boundary for Large Language Models: A Different Perspective on Model Evaluation", author = "Yin, Xunjian and Zhang, Xu and Ruan, Jie and Wan, Xiaojun", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.124", pages = "2270--2286", abstract = "In recent years, substantial advancements have been made in the development of large language models, achieving remarkable performance across diverse tasks.To evaluate the knowledge ability of language models, previous studies have proposed lots of benchmarks based on question-answering pairs.We argue that it is not reliable and comprehensive to evaluate language models with a fixed question or limited paraphrases as the query, since language models are sensitive to prompt.Therefore, we introduce a novel concept named knowledge boundary to encompass both prompt-agnostic and prompt-sensitive knowledge within language models.Knowledge boundary avoids prompt sensitivity in language model evaluations, rendering them more dependable and robust.To explore the knowledge boundary for a given model, we propose projected gradient descent method with semantic constraints, a new algorithm designed to identify the optimal prompt for each piece of knowledge.Experiments demonstrate a superior performance of our algorithm in computing the knowledge boundary compared to existing methods.Furthermore, we evaluate the ability of multiple language models in several domains with knowledge boundary.", }

In recent years, substantial advancements have been made in the development of large language models, achieving remarkable performance across diverse tasks.To evaluate the knowledge ability of language models, previous studies have proposed lots of benchmarks based on question-answering pairs.We argue that it is not reliable and comprehensive to evaluate language models with a fixed question or limited paraphrases as the query, since language models are sensitive to prompt.Therefore, we introduce a novel concept named knowledge boundary to encompass both prompt-agnostic and prompt-sensitive knowledge within language models.Knowledge boundary avoids prompt sensitivity in language model evaluations, rendering them more dependable and robust.To explore the knowledge boundary for a given model, we propose projected gradient descent method with semantic constraints, a new algorithm designed to identify the optimal prompt for each piece of knowledge.Experiments demonstrate a superior performance of our algorithm in computing the knowledge boundary compared to existing methods.Furthermore, we evaluate the ability of multiple language models in several domains with knowledge boundary.

Benchmarking Knowledge Boundary for Large Language Models: A Different Perspective on Model Evaluation

acl-long.124

Poster

2402.11493

https://aclanthology.org/2024.acl-long.124/

https://aclanthology.org/2024.acl-long.125.bib

@inproceedings{yoon-etal-2024-listt5, title = "{L}ist{T}5: Listwise Reranking with Fusion-in-Decoder Improves Zero-shot Retrieval", author = "Yoon, Soyoung and Choi, Eunbi and Kim, Jiyeon and Yun, Hyeongu and Kim, Yireun and Hwang, Seung-won", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.125", pages = "2287--2308", abstract = "We propose ListT5, a novel reranking approach based on Fusion-in-Decoder (FiD) that handles multiple candidate passages at both train and inference time. We also introduce an efficient inference framework for listwise ranking based on m-ary tournament sort with output caching. We evaluate and compare our model on the BEIR benchmark for zero-shot retrieval task, demonstrating that ListT5 (1) outperforms the state-of-the-art RankT5 baseline with a notable +1.3 gain in the average NDCG@10 score, (2) has an efficiency comparable to pointwise ranking models and surpasses the efficiency of previous listwise ranking models, and (3) overcomes the lost-in-the-middle problem of previous listwise rerankers. Our code, model checkpoints, and the evaluation framework will be fully open-sourced.", }

We propose ListT5, a novel reranking approach based on Fusion-in-Decoder (FiD) that handles multiple candidate passages at both train and inference time. We also introduce an efficient inference framework for listwise ranking based on m-ary tournament sort with output caching. We evaluate and compare our model on the BEIR benchmark for zero-shot retrieval task, demonstrating that ListT5 (1) outperforms the state-of-the-art RankT5 baseline with a notable +1.3 gain in the average NDCG@10 score, (2) has an efficiency comparable to pointwise ranking models and surpasses the efficiency of previous listwise ranking models, and (3) overcomes the lost-in-the-middle problem of previous listwise rerankers. Our code, model checkpoints, and the evaluation framework will be fully open-sourced.

ListT5: Listwise Reranking with Fusion-in-Decoder Improves Zero-shot Retrieval

acl-long.125

Oral

2402.15838

https://huggingface.co/papers/2402.15838

https://aclanthology.org/2024.acl-long.125/

https://aclanthology.org/2024.acl-long.126.bib

@inproceedings{chen-etal-2024-exploring-potential, title = "Exploring the Potential of Large Language Models in Computational Argumentation", author = "Chen, Guizhen and Cheng, Liying and Luu, Anh Tuan and Bing, Lidong", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.126", pages = "2309--2330", abstract = "Computational argumentation has become an essential tool in various domains, including law, public policy, and artificial intelligence. It is an emerging research field in natural language processing that attracts increasing attention. Research on computational argumentation mainly involves two types of tasks: argument mining and argument generation. As large language models (LLMs) have demonstrated impressive capabilities in understanding context and generating natural language, it is worthwhile to evaluate the performance of LLMs on diverse computational argumentation tasks. This work aims to embark on an assessment of LLMs, such as ChatGPT, Flan models, and LLaMA2 models, in both zero-shot and few-shot settings. We organize existing tasks into six main categories and standardize the format of fourteen openly available datasets. In addition, we present a new benchmark dataset on counter speech generation that aims to holistically evaluate the end-to-end performance of LLMs on argument mining and argument generation. Extensive experiments show that LLMs exhibit commendable performance across most of the datasets, demonstrating their capabilities in the field of argumentation. Our analysis offers valuable suggestions for evaluating computational argumentation and its integration with LLMs in future research endeavors.", }

Computational argumentation has become an essential tool in various domains, including law, public policy, and artificial intelligence. It is an emerging research field in natural language processing that attracts increasing attention. Research on computational argumentation mainly involves two types of tasks: argument mining and argument generation. As large language models (LLMs) have demonstrated impressive capabilities in understanding context and generating natural language, it is worthwhile to evaluate the performance of LLMs on diverse computational argumentation tasks. This work aims to embark on an assessment of LLMs, such as ChatGPT, Flan models, and LLaMA2 models, in both zero-shot and few-shot settings. We organize existing tasks into six main categories and standardize the format of fourteen openly available datasets. In addition, we present a new benchmark dataset on counter speech generation that aims to holistically evaluate the end-to-end performance of LLMs on argument mining and argument generation. Extensive experiments show that LLMs exhibit commendable performance across most of the datasets, demonstrating their capabilities in the field of argumentation. Our analysis offers valuable suggestions for evaluating computational argumentation and its integration with LLMs in future research endeavors.

Exploring the Potential of Large Language Models in Computational Argumentation

acl-long.126

Poster

2311.09022

https://aclanthology.org/2024.acl-long.126/

https://aclanthology.org/2024.acl-long.127.bib

@inproceedings{moskvoretskii-etal-2024-taxollama, title = "{T}axo{LL}a{MA}: {W}ord{N}et-based Model for Solving Multiple Lexical Semantic Tasks", author = "Moskvoretskii, Viktor and Neminova, Ekaterina and Lobanova, Alina and Panchenko, Alexander and Nikishina, Irina", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.127", pages = "2331--2350", abstract = "In this paper, we explore the capabilities of LLMs in capturing lexical-semantic knowledge from WordNet on the example of the LLaMA-2-7b model and test it on multiple lexical semantic tasks. As the outcome of our experiments, we present TaxoLLaMA, the {``}all-in-one{''} model for taxonomy-related tasks, lightweight due to 4-bit quantization and LoRA. TaxoLLaMA achieves 11 SOTA results, and 4 top-2 results out of 16 tasks on the Taxonomy Enrichment, Hypernym Discovery, Taxonomy Construction, and Lexical Entailment tasks. Moreover, it demonstrates a very strong zero-shot performance on Lexical Entailment and Taxonomy Construction with no fine-tuning. We also explore its hidden multilingual and domain adaptation capabilities with a little tuning or few-shot learning. All datasets, code, and pre-trained models are available online (code: https://github.com/VityaVitalich/TaxoLLaMA)", }

In this paper, we explore the capabilities of LLMs in capturing lexical-semantic knowledge from WordNet on the example of the LLaMA-2-7b model and test it on multiple lexical semantic tasks. As the outcome of our experiments, we present TaxoLLaMA, the {``}all-in-one{''} model for taxonomy-related tasks, lightweight due to 4-bit quantization and LoRA. TaxoLLaMA achieves 11 SOTA results, and 4 top-2 results out of 16 tasks on the Taxonomy Enrichment, Hypernym Discovery, Taxonomy Construction, and Lexical Entailment tasks. Moreover, it demonstrates a very strong zero-shot performance on Lexical Entailment and Taxonomy Construction with no fine-tuning. We also explore its hidden multilingual and domain adaptation capabilities with a little tuning or few-shot learning. All datasets, code, and pre-trained models are available online (code: https://github.com/VityaVitalich/TaxoLLaMA)

TaxoLLaMA: WordNet-based Model for Solving Multiple Lexical Semantic Tasks

acl-long.127

Poster

2403.09207

https://huggingface.co/papers/2403.09207

https://aclanthology.org/2024.acl-long.127/

https://aclanthology.org/2024.acl-long.128.bib

@inproceedings{wang-etal-2024-candle, title = "{CANDLE}: Iterative Conceptualization and Instantiation Distillation from Large Language Models for Commonsense Reasoning", author = "Wang, Weiqi and Fang, Tianqing and Li, Chunyang and Shi, Haochen and Ding, Wenxuan and Xu, Baixuan and Wang, Zhaowei and Bai, Jiaxin and Liu, Xin and Jiayang, Cheng and Chan, Chunkit and Song, Yangqiu", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.128", pages = "2351--2374", abstract = "The sequential process of conceptualization and instantiation is essential to generalizable commonsense reasoning as it allows the application of existing knowledge to unfamiliar scenarios. However, existing works tend to undervalue the step of instantiation and heavilyrely on pre-built concept taxonomies and human annotations to collect both types of knowledge, resulting in a lack of instantiated knowledge to complete reasoning, high cost, and limited scalability. To tackle these challenges, we introduce CANDLE (ConceptuAlizationand INstantiation Distillation from Large Language ModEls), a distillation framework that iteratively performs contextualized conceptualization and instantiation over commonsense knowledge bases by instructing large language models to generate both types of knowledge with critic filtering. By applying CANDLE to ATOMIC (Sap et al., 2019a), we construct a comprehensive knowledge base comprising six million conceptualizations and instantiated commonsense knowledge triples. Both types of knowledge are firmly rooted in the original ATOMIC dataset, and intrinsic evaluations demonstrate their exceptional quality and diversity. Empirical results indicate that distilling CANDLE on student models provides benefits across three downstream tasks. Our data and models are publicly available at https://github.com/HKUST-KnowComp/CANDLE.", }

The sequential process of conceptualization and instantiation is essential to generalizable commonsense reasoning as it allows the application of existing knowledge to unfamiliar scenarios. However, existing works tend to undervalue the step of instantiation and heavilyrely on pre-built concept taxonomies and human annotations to collect both types of knowledge, resulting in a lack of instantiated knowledge to complete reasoning, high cost, and limited scalability. To tackle these challenges, we introduce CANDLE (ConceptuAlizationand INstantiation Distillation from Large Language ModEls), a distillation framework that iteratively performs contextualized conceptualization and instantiation over commonsense knowledge bases by instructing large language models to generate both types of knowledge with critic filtering. By applying CANDLE to ATOMIC (Sap et al., 2019a), we construct a comprehensive knowledge base comprising six million conceptualizations and instantiated commonsense knowledge triples. Both types of knowledge are firmly rooted in the original ATOMIC dataset, and intrinsic evaluations demonstrate their exceptional quality and diversity. Empirical results indicate that distilling CANDLE on student models provides benefits across three downstream tasks. Our data and models are publicly available at https://github.com/HKUST-KnowComp/CANDLE.

CANDLE: Iterative Conceptualization and Instantiation Distillation from Large Language Models for Commonsense Reasoning

acl-long.128

Poster

2401.07286

https://huggingface.co/papers/2401.07286

https://aclanthology.org/2024.acl-long.128/

https://aclanthology.org/2024.acl-long.129.bib

@inproceedings{hao-etal-2024-meft, title = "{MEFT}: Memory-Efficient Fine-Tuning through Sparse Adapter", author = "Hao, Jitai and Sun, Weiwei and Xin, Xin and Meng, Qi and Chen, Zhumin and Ren, Pengjie and Ren, Zhaochun", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.129", pages = "2375--2388", abstract = "Parameter-Efficient Fine-tuning (PEFT) facilitates the fine-tuning of Large Language Models (LLMs) under limited resources. However, the fine-tuning performance with PEFT on complex, knowledge-intensive tasks is limited due to the constrained model capacity, which originates from the limited number of additional trainable parameters. To overcome this limitation, we introduce a novel mechanism that fine-tunes LLMs with adapters of larger size yet memory-efficient. This is achieved by leveraging the inherent activation sparsity in the Feed-Forward Networks (FFNs) of LLMs and utilizing the larger capacity of Central Processing Unit (CPU) memory compared to Graphics Processing Unit (GPU). We store and update the parameters of larger adapters on the CPU. Moreover, we employ a Mixture of Experts (MoE)-like architecture to mitigate unnecessary CPU computations and reduce the communication volume between the GPU and CPU. This is particularly beneficial over the limited bandwidth of PCI Express (PCIe). Our method can achieve fine-tuning results comparable to those obtained with larger memory capacities, even when operating under more limited resources such as a 24GB memory single GPU setup, with acceptable loss in training efficiency. Our codes are available at https://github.com/CURRENTF/MEFT.", }

Parameter-Efficient Fine-tuning (PEFT) facilitates the fine-tuning of Large Language Models (LLMs) under limited resources. However, the fine-tuning performance with PEFT on complex, knowledge-intensive tasks is limited due to the constrained model capacity, which originates from the limited number of additional trainable parameters. To overcome this limitation, we introduce a novel mechanism that fine-tunes LLMs with adapters of larger size yet memory-efficient. This is achieved by leveraging the inherent activation sparsity in the Feed-Forward Networks (FFNs) of LLMs and utilizing the larger capacity of Central Processing Unit (CPU) memory compared to Graphics Processing Unit (GPU). We store and update the parameters of larger adapters on the CPU. Moreover, we employ a Mixture of Experts (MoE)-like architecture to mitigate unnecessary CPU computations and reduce the communication volume between the GPU and CPU. This is particularly beneficial over the limited bandwidth of PCI Express (PCIe). Our method can achieve fine-tuning results comparable to those obtained with larger memory capacities, even when operating under more limited resources such as a 24GB memory single GPU setup, with acceptable loss in training efficiency. Our codes are available at https://github.com/CURRENTF/MEFT.

MEFT: Memory-Efficient Fine-Tuning through Sparse Adapter

acl-long.129

Poster

2406.04984

https://aclanthology.org/2024.acl-long.129/

https://aclanthology.org/2024.acl-long.130.bib

@inproceedings{chavan-etal-2024-surgical, title = "Surgical Feature-Space Decomposition of {LLM}s: Why, When and How?", author = "Chavan, Arnav and Lele, Nahush and Gupta, Deepak", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.130", pages = "2389--2400", abstract = "Low-rank approximations, of the weight and feature space can enhance the performance of deep learning models, whether in terms of improving generalization or reducing the latency of inference. However, there is no clear consensus yet on how, when and why these approximations are helpful for large language models (LLMs). In this work, we empirically study the efficacy of weight and feature space decomposition in transformer-based LLMs. We demonstrate that surgical decomposition not only provides critical insights into the trade-off between compression and language modelling performance, but also sometimes enhances commonsense reasoning performance of LLMs. Our empirical analysis identifies specific network segments that intrinsically exhibit a low-rank structure. Furthermore, we extend our investigation to the implications of low-rank approximations on model bias. Overall, our findings offer a novel perspective on optimizing LLMs, presenting the low-rank approximation not only as a tool for performance enhancements, but also as a means to potentially rectify biases within these models.", }

Low-rank approximations, of the weight and feature space can enhance the performance of deep learning models, whether in terms of improving generalization or reducing the latency of inference. However, there is no clear consensus yet on how, when and why these approximations are helpful for large language models (LLMs). In this work, we empirically study the efficacy of weight and feature space decomposition in transformer-based LLMs. We demonstrate that surgical decomposition not only provides critical insights into the trade-off between compression and language modelling performance, but also sometimes enhances commonsense reasoning performance of LLMs. Our empirical analysis identifies specific network segments that intrinsically exhibit a low-rank structure. Furthermore, we extend our investigation to the implications of low-rank approximations on model bias. Overall, our findings offer a novel perspective on optimizing LLMs, presenting the low-rank approximation not only as a tool for performance enhancements, but also as a means to potentially rectify biases within these models.

Surgical Feature-Space Decomposition of LLMs: Why, When and How?

acl-long.130

Poster

2405.13039

https://aclanthology.org/2024.acl-long.130/

https://aclanthology.org/2024.acl-long.131.bib

@inproceedings{yin-etal-2024-reasoning, title = "Reasoning in Flux: Enhancing Large Language Models Reasoning through Uncertainty-aware Adaptive Guidance", author = "Yin, Zhangyue and Sun, Qiushi and Guo, Qipeng and Zeng, Zhiyuan and Li, Xiaonan and Dai, Junqi and Cheng, Qinyuan and Huang, Xuanjing and Qiu, Xipeng", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.131", pages = "2401--2416", abstract = "Machine reasoning, which involves solving complex problems through step-by-step deduction and analysis, is a crucial indicator of the capabilities of Large Language Models (LLMs). However, as the complexity of tasks escalates, LLMs often encounter increasing errors in their multi-step reasoning process. This study delves into the underlying factors contributing to these reasoning errors and seeks to leverage uncertainty to refine them. Specifically, we introduce Uncertainty-aware Adaptive Guidance (UAG), a novel approach for guiding LLM reasoning onto an accurate and reliable trajectory. UAG first identifies and evaluates uncertainty signals within each step of the reasoning chain. Upon detecting a significant increase in uncertainty, UAG intervenes by retracting to a previously reliable state and then introduces certified reasoning clues for refinement. By dynamically adjusting the reasoning process, UAG offers a plug-and-play solution for improving LLMs{'} performance in complex reasoning. Extensive experiments across various reasoning tasks demonstrate that UAG not only enhances the reasoning abilities of LLMs but also consistently outperforms several strong baselines with minimal computational overhead. Further analysis reveals that UAG is notably effective in identifying and diminishing reasoning errors.", }

Machine reasoning, which involves solving complex problems through step-by-step deduction and analysis, is a crucial indicator of the capabilities of Large Language Models (LLMs). However, as the complexity of tasks escalates, LLMs often encounter increasing errors in their multi-step reasoning process. This study delves into the underlying factors contributing to these reasoning errors and seeks to leverage uncertainty to refine them. Specifically, we introduce Uncertainty-aware Adaptive Guidance (UAG), a novel approach for guiding LLM reasoning onto an accurate and reliable trajectory. UAG first identifies and evaluates uncertainty signals within each step of the reasoning chain. Upon detecting a significant increase in uncertainty, UAG intervenes by retracting to a previously reliable state and then introduces certified reasoning clues for refinement. By dynamically adjusting the reasoning process, UAG offers a plug-and-play solution for improving LLMs{'} performance in complex reasoning. Extensive experiments across various reasoning tasks demonstrate that UAG not only enhances the reasoning abilities of LLMs but also consistently outperforms several strong baselines with minimal computational overhead. Further analysis reveals that UAG is notably effective in identifying and diminishing reasoning errors.

Reasoning in Flux: Enhancing Large Language Models Reasoning through Uncertainty-aware Adaptive Guidance

acl-long.131

Poster

https://aclanthology.org/2024.acl-long.131/

https://aclanthology.org/2024.acl-long.132.bib

@inproceedings{dong-etal-2024-modality, title = "Modality-Aware Integration with Large Language Models for Knowledge-Based Visual Question Answering", author = "Dong, Junnan and Zhang, Qinggang and Zhou, Huachi and Zha, Daochen and Zheng, Pai and Huang, Xiao", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.132", pages = "2417--2429", abstract = "Knowledge-based visual question answering (KVQA) has been extensively studied to answer visual questions with external knowledge, e.g., knowledge graphs (KGs). While several attempts have been proposed to leverage large language models (LLMs) as an implicit knowledge source, it remains challenging since LLMs may generate hallucinations. Moreover, multiple knowledge sources, e.g., images, KGs and LLMs, cannot be readily aligned for complex scenarios. To tackle these, we present a novel modality-aware integration with LLMs for KVQA ($\texttt{MAIL}$). It carefully leverages multimodal knowledge for both image understanding and knowledge reasoning. Specifically, $(i)$ we propose a two-stage prompting strategy with LLMs to densely embody the image into a *scene graph* with detailed visual features; $(ii)$ We construct a coupled *concept graph* by linking the mentioned entities with external facts. $(iii)$ A tailored pseudo-siamese graph medium fusion is designed for sufficient multimodal fusion. We utilize the shared mentioned entities in two graphs as mediums to bridge a tight inter-modal exchange, while maximally preserving insightful intra-modal learning by constraining the fusion within mediums. Extensive experiments show the superiority of $\texttt{MAIL}$.", }

Knowledge-based visual question answering (KVQA) has been extensively studied to answer visual questions with external knowledge, e.g., knowledge graphs (KGs). While several attempts have been proposed to leverage large language models (LLMs) as an implicit knowledge source, it remains challenging since LLMs may generate hallucinations. Moreover, multiple knowledge sources, e.g., images, KGs and LLMs, cannot be readily aligned for complex scenarios. To tackle these, we present a novel modality-aware integration with LLMs for KVQA ($\texttt{MAIL}$). It carefully leverages multimodal knowledge for both image understanding and knowledge reasoning. Specifically, $(i)$ we propose a two-stage prompting strategy with LLMs to densely embody the image into a *scene graph* with detailed visual features; $(ii)$ We construct a coupled *concept graph* by linking the mentioned entities with external facts. $(iii)$ A tailored pseudo-siamese graph medium fusion is designed for sufficient multimodal fusion. We utilize the shared mentioned entities in two graphs as mediums to bridge a tight inter-modal exchange, while maximally preserving insightful intra-modal learning by constraining the fusion within mediums. Extensive experiments show the superiority of $\texttt{MAIL}$.

Modality-Aware Integration with Large Language Models for Knowledge-Based Visual Question Answering

acl-long.132

Poster

2402.12728

https://aclanthology.org/2024.acl-long.132/

https://aclanthology.org/2024.acl-long.133.bib

@inproceedings{liu-etal-2024-unlocking-data, title = "Unlocking Data-free Low-bit Quantization with Matrix Decomposition for {KV} Cache Compression", author = "Liu, Peiyu and Gao, Ze-Feng and Zhao, Xin and Ma, Yipeng and Wang, Tao and Wen, Ji-Rong", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.133", pages = "2430--2440", abstract = "Key-value (KV) caching is an important technique to accelerate the inference of large language models (LLMs), but incurs significant memory overhead. To compress the size of KV cache, existing methods often compromise precision or require extra data for calibration, limiting their practicality in LLM deployment. In this paper, we introduce \textbf{DecoQuant}, a novel data-free low-bit quantization technique based on tensor decomposition methods, to effectively compress KV cache. Our core idea is to adjust the outlier distribution of the original matrix by performing tensor decomposition, so that the quantization difficulties are migrated from the matrix to decomposed local tensors. Specially, we find that outliers mainly concentrate on small local tensors, while large tensors tend to have a narrower value range. Based on this finding, we propose to apply low-bit quantization to the large tensor, while maintaining high-precision representation for the small tensor. Furthermore, we utilize the proposed quantization method to compress the KV cache of LLMs to accelerate the inference, and develop an efficient dequantization kernel tailored specifically for DecoQuant. Through extensive experiments, DecoQuant demonstrates remarkable efficiency gains, showcasing up to a 75{\%} reduction in memory footprint while maintaining comparable generation quality.", }

Key-value (KV) caching is an important technique to accelerate the inference of large language models (LLMs), but incurs significant memory overhead. To compress the size of KV cache, existing methods often compromise precision or require extra data for calibration, limiting their practicality in LLM deployment. In this paper, we introduce \textbf{DecoQuant}, a novel data-free low-bit quantization technique based on tensor decomposition methods, to effectively compress KV cache. Our core idea is to adjust the outlier distribution of the original matrix by performing tensor decomposition, so that the quantization difficulties are migrated from the matrix to decomposed local tensors. Specially, we find that outliers mainly concentrate on small local tensors, while large tensors tend to have a narrower value range. Based on this finding, we propose to apply low-bit quantization to the large tensor, while maintaining high-precision representation for the small tensor. Furthermore, we utilize the proposed quantization method to compress the KV cache of LLMs to accelerate the inference, and develop an efficient dequantization kernel tailored specifically for DecoQuant. Through extensive experiments, DecoQuant demonstrates remarkable efficiency gains, showcasing up to a 75{\%} reduction in memory footprint while maintaining comparable generation quality.

Unlocking Data-free Low-bit Quantization with Matrix Decomposition for KV Cache Compression

acl-long.133

Poster

2405.12591

https://aclanthology.org/2024.acl-long.133/

https://aclanthology.org/2024.acl-long.134.bib

@inproceedings{kim-etal-2024-verifiner, title = "{V}erifi{NER}: Verification-augmented {NER} via Knowledge-grounded Reasoning with Large Language Models", author = "Kim, Seoyeon and Seo, Kwangwook and Chae, Hyungjoo and Yeo, Jinyoung and Lee, Dongha", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.134", pages = "2441--2461", abstract = "Recent approaches in domain-specific named entity recognition (NER), such as biomedical NER, have shown remarkable advances. However, they still lack of faithfulness, producing erroneous predictions. We assume that knowledge of entities can be useful in verifying the correctness of the predictions. Despite the usefulness of knowledge, resolving such errors with knowledge is nontrivial, since the knowledge itself does not directly indicate the ground-truth label. To this end, we propose VerifiNER, a post-hoc verification framework that identifies errors from existing NER methods using knowledge and revises them into more faithful predictions. Our framework leverages the reasoning abilities of large language models to adequately ground on knowledge and the contextual information in the verification process. We validate effectiveness of VerifiNER through extensive experiments on biomedical datasets. The results suggest that VerifiNER can successfully verify errors from existing models as a model-agnostic approach. Further analyses on out-of-domain and low-resource settings show the usefulness of VerifiNER on real-world applications.", }

Recent approaches in domain-specific named entity recognition (NER), such as biomedical NER, have shown remarkable advances. However, they still lack of faithfulness, producing erroneous predictions. We assume that knowledge of entities can be useful in verifying the correctness of the predictions. Despite the usefulness of knowledge, resolving such errors with knowledge is nontrivial, since the knowledge itself does not directly indicate the ground-truth label. To this end, we propose VerifiNER, a post-hoc verification framework that identifies errors from existing NER methods using knowledge and revises them into more faithful predictions. Our framework leverages the reasoning abilities of large language models to adequately ground on knowledge and the contextual information in the verification process. We validate effectiveness of VerifiNER through extensive experiments on biomedical datasets. The results suggest that VerifiNER can successfully verify errors from existing models as a model-agnostic approach. Further analyses on out-of-domain and low-resource settings show the usefulness of VerifiNER on real-world applications.

VerifiNER: Verification-augmented NER via Knowledge-grounded Reasoning with Large Language Models

acl-long.134

Poster

2402.18374

https://huggingface.co/papers/2402.18374

https://aclanthology.org/2024.acl-long.134/

https://aclanthology.org/2024.acl-long.135.bib

@inproceedings{li-etal-2024-making, title = "Making Long-Context Language Models Better Multi-Hop Reasoners", author = "Li, Yanyang and Liang, Shuo and Lyu, Michael and Wang, Liwei", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.135", pages = "2462--2475", abstract = "Recent advancements in long-context modeling have enhanced language models (LMs) for complex tasks across multiple NLP applications. Despite this progress, we find that these models struggle with multi-hop reasoning and exhibit decreased performance in the presence of noisy contexts. In this paper, we introduce Reasoning with Attributions, a novel approach that prompts LMs to supply attributions for each assertion during their reasoning. We validate our approach through experiments on three multi-hop datasets, employing both proprietary and open-source models, and demonstrate its efficacy and resilience. Furthermore, we explore methods to augment reasoning capabilities via fine-tuning and offer an attribution-annotated dataset and a specialized training strategy. Our fine-tuned model achieves competitive performance on multi-hop reasoning benchmarks, closely paralleling proprietary LMs such as ChatGPT and Claude-instant.", }

Recent advancements in long-context modeling have enhanced language models (LMs) for complex tasks across multiple NLP applications. Despite this progress, we find that these models struggle with multi-hop reasoning and exhibit decreased performance in the presence of noisy contexts. In this paper, we introduce Reasoning with Attributions, a novel approach that prompts LMs to supply attributions for each assertion during their reasoning. We validate our approach through experiments on three multi-hop datasets, employing both proprietary and open-source models, and demonstrate its efficacy and resilience. Furthermore, we explore methods to augment reasoning capabilities via fine-tuning and offer an attribution-annotated dataset and a specialized training strategy. Our fine-tuned model achieves competitive performance on multi-hop reasoning benchmarks, closely paralleling proprietary LMs such as ChatGPT and Claude-instant.

Making Long-Context Language Models Better Multi-Hop Reasoners

acl-long.135

Poster

2408.03246

https://aclanthology.org/2024.acl-long.135/

https://aclanthology.org/2024.acl-long.136.bib

@inproceedings{liu-etal-2024-translico, title = "{T}ransli{C}o: A Contrastive Learning Framework to Address the Script Barrier in Multilingual Pretrained Language Models", author = "Liu, Yihong and Ma, Chunlan and Ye, Haotian and Schuetze, Hinrich", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.136", pages = "2476--2499", abstract = "The world{'}s more than 7000 languages are written in at least 293 scripts. Due to various reasons, many closely related languages use different scripts, which poses a difficulty for multilingual pretrained language models (mPLMs) in learning crosslingual knowledge through lexical overlap. As a consequence, mPLMs are faced with a script barrier: representations from different scripts are located in different subspaces, which can result in crosslingual transfer involving languages of different scripts performing suboptimally. To address this problem, we propose TransliCo, a framework that optimizes the Transliteration Contrastive Modeling (TCM) objective to fine-tune an mPLM by contrasting sentences in its training data and their transliterations in a unified script (in our case Latin), which enhances uniformity in the representation space for different scripts. Using Glot500-m, an mPLM pretrained on over 500 languages, as our source model, we fine-tune it on a small portion (5{\%}) of its training data, and refer to the resulting model as Furina. We show that Furina not only better aligns representations from distinct scripts but also outperforms the original Glot500-m on various zero-shot crosslingual transfer tasks. Additionally, we achieve consistent improvement in a case study on the Indic group where the languages exhibit areal features but use different scripts. We make our code and models publicly available.", }

The world{'}s more than 7000 languages are written in at least 293 scripts. Due to various reasons, many closely related languages use different scripts, which poses a difficulty for multilingual pretrained language models (mPLMs) in learning crosslingual knowledge through lexical overlap. As a consequence, mPLMs are faced with a script barrier: representations from different scripts are located in different subspaces, which can result in crosslingual transfer involving languages of different scripts performing suboptimally. To address this problem, we propose TransliCo, a framework that optimizes the Transliteration Contrastive Modeling (TCM) objective to fine-tune an mPLM by contrasting sentences in its training data and their transliterations in a unified script (in our case Latin), which enhances uniformity in the representation space for different scripts. Using Glot500-m, an mPLM pretrained on over 500 languages, as our source model, we fine-tune it on a small portion (5{\%}) of its training data, and refer to the resulting model as Furina. We show that Furina not only better aligns representations from distinct scripts but also outperforms the original Glot500-m on various zero-shot crosslingual transfer tasks. Additionally, we achieve consistent improvement in a case study on the Indic group where the languages exhibit areal features but use different scripts. We make our code and models publicly available.

TransliCo: A Contrastive Learning Framework to Address the Script Barrier in Multilingual Pretrained Language Models

acl-long.136

Poster

2401.06620

https://aclanthology.org/2024.acl-long.136/

https://aclanthology.org/2024.acl-long.137.bib

@inproceedings{raina-etal-2024-extreme, title = "Extreme Miscalibration and the Illusion of Adversarial Robustness", author = "Raina, Vyas and Tan, Samson and Cevher, Volkan and Rawal, Aditya and Zha, Sheng and Karypis, George", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.137", pages = "2500--2525", abstract = "Deep learning-based Natural Language Processing (NLP) models are vulnerable to adversarial attacks, where small perturbations can cause a model to misclassify. Adversarial Training (AT) is often used to increase model robustness. However, we have discovered an intriguing phenomenon: deliberately or accidentally miscalibrating models masks gradients in a way that interferes with adversarial attack search methods, giving rise to an apparent increase in robustness. We show that this observed gain in robustness is an illusion of robustness (IOR), and demonstrate how an adversary can perform various forms of test-time temperature calibration to nullify the aforementioned interference and allow the adversarial attack to find adversarial examples. Hence, we urge the NLP community to incorporate test-time temperature scaling into their robustness evaluations to ensure that any observed gains are genuine. Finally, we show how the temperature can be scaled during training to improve genuine robustness.", }

Deep learning-based Natural Language Processing (NLP) models are vulnerable to adversarial attacks, where small perturbations can cause a model to misclassify. Adversarial Training (AT) is often used to increase model robustness. However, we have discovered an intriguing phenomenon: deliberately or accidentally miscalibrating models masks gradients in a way that interferes with adversarial attack search methods, giving rise to an apparent increase in robustness. We show that this observed gain in robustness is an illusion of robustness (IOR), and demonstrate how an adversary can perform various forms of test-time temperature calibration to nullify the aforementioned interference and allow the adversarial attack to find adversarial examples. Hence, we urge the NLP community to incorporate test-time temperature scaling into their robustness evaluations to ensure that any observed gains are genuine. Finally, we show how the temperature can be scaled during training to improve genuine robustness.

Extreme Miscalibration and the Illusion of Adversarial Robustness

acl-long.137

Poster

2402.17509

https://aclanthology.org/2024.acl-long.137/

https://aclanthology.org/2024.acl-long.138.bib

@inproceedings{zheng-etal-2024-hycorec, title = "{H}y{C}o{R}ec: Hypergraph-Enhanced Multi-Preference Learning for Alleviating Matthew Effect in Conversational Recommendation", author = "Zheng, Yongsen and Xu, Ruilin and Chen, Ziliang and Wang, Guohua and Qian, Mingjie and Qin, Jinghui and Lin, Liang", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.138", pages = "2526--2537", abstract = "The Matthew effect is a notorious issue in Recommender Systems (RSs), i.e., the rich get richer and the poor get poorer, wherein popular items are overexposed while less popular ones are regularly ignored. Most methods examine Matthew effect in static or nearly-static recommendation scenarios. However, the Matthew effect will be increasingly amplified when the user interacts with the system over time. To address these issues, we propose a novel paradigm, Hypergraph-Enhanced Multi-Preference Learning for Alleviating Matthew Effect in Conversational Recommendation (HyCoRec), which aims to alleviate the Matthew effect in conversational recommendation. Concretely, HyCoRec devotes to alleviate the Matthew effect by learning multi-aspect preferences, i.e., item-, entity-, word-, review-, and knowledge-aspect preferences, to effectively generate responses in the conversational task and accurately predict items in the recommendation task when the user chats with the system over time. Extensive experiments conducted on two benchmarks validate that HyCoRec achieves new state-of-the-art performance and the superior of alleviating Matthew effect.", }

The Matthew effect is a notorious issue in Recommender Systems (RSs), i.e., the rich get richer and the poor get poorer, wherein popular items are overexposed while less popular ones are regularly ignored. Most methods examine Matthew effect in static or nearly-static recommendation scenarios. However, the Matthew effect will be increasingly amplified when the user interacts with the system over time. To address these issues, we propose a novel paradigm, Hypergraph-Enhanced Multi-Preference Learning for Alleviating Matthew Effect in Conversational Recommendation (HyCoRec), which aims to alleviate the Matthew effect in conversational recommendation. Concretely, HyCoRec devotes to alleviate the Matthew effect by learning multi-aspect preferences, i.e., item-, entity-, word-, review-, and knowledge-aspect preferences, to effectively generate responses in the conversational task and accurately predict items in the recommendation task when the user chats with the system over time. Extensive experiments conducted on two benchmarks validate that HyCoRec achieves new state-of-the-art performance and the superior of alleviating Matthew effect.

HyCoRec: Hypergraph-Enhanced Multi-Preference Learning for Alleviating Matthew Effect in Conversational Recommendation

acl-long.138

Poster

https://aclanthology.org/2024.acl-long.138/

https://aclanthology.org/2024.acl-long.139.bib

@inproceedings{sadat-caragea-2024-co, title = "Co-training for Low Resource Scientific Natural Language Inference", author = "Sadat, Mobashir and Caragea, Cornelia", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.139", pages = "2538--2550", abstract = "Scientific Natural Language Inference (NLI) is the task of predicting the semantic relation between a pair of sentences extracted from research articles. The automatic annotation method based on distant supervision for the training set of SciNLI, the first and most popular dataset for this task, results in label noise which inevitably degenerates the performance of classifiers. In this paper, we propose a novel co-training method that assigns weights based on the training dynamics of the classifiers to the distantly supervised labels, reflective of the manner they are used in the subsequent training epochs. That is, unlike the existing semi-supervised learning (SSL) approaches, we consider the historical behavior of the classifiers to evaluate the quality of the automatically annotated labels. Furthermore, by assigning importance weights instead of filtering out examples based on an arbitrary threshold on the predicted confidence, we maximize the usage of automatically labeled data, while ensuring that the noisy labels have a minimal impact on model training. The proposed method obtains an improvement of 1.5{\%} in Macro F1 over the distant supervision baseline, and substantial improvements over several other strong SSL baselines. We make our code and data available on Github.", }

Scientific Natural Language Inference (NLI) is the task of predicting the semantic relation between a pair of sentences extracted from research articles. The automatic annotation method based on distant supervision for the training set of SciNLI, the first and most popular dataset for this task, results in label noise which inevitably degenerates the performance of classifiers. In this paper, we propose a novel co-training method that assigns weights based on the training dynamics of the classifiers to the distantly supervised labels, reflective of the manner they are used in the subsequent training epochs. That is, unlike the existing semi-supervised learning (SSL) approaches, we consider the historical behavior of the classifiers to evaluate the quality of the automatically annotated labels. Furthermore, by assigning importance weights instead of filtering out examples based on an arbitrary threshold on the predicted confidence, we maximize the usage of automatically labeled data, while ensuring that the noisy labels have a minimal impact on model training. The proposed method obtains an improvement of 1.5{\%} in Macro F1 over the distant supervision baseline, and substantial improvements over several other strong SSL baselines. We make our code and data available on Github.

Co-training for Low Resource Scientific Natural Language Inference

acl-long.139

Poster

2406.14666

https://aclanthology.org/2024.acl-long.139/

https://aclanthology.org/2024.acl-long.140.bib

@inproceedings{wang-etal-2024-rlhfpoison, title = "{RLHFP}oison: Reward Poisoning Attack for Reinforcement Learning with Human Feedback in Large Language Models", author = "Wang, Jiongxiao and Wu, Junlin and Chen, Muhao and Vorobeychik, Yevgeniy and Xiao, Chaowei", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.140", pages = "2551--2570", abstract = "Reinforcement Learning with Human Feedback (RLHF) is a methodology designed to align Large Language Models (LLMs) with human preferences, playing an important role in LLMs alignment. Despite its advantages, RLHF relies on human annotators to rank the text, which can introduce potential security vulnerabilities if any adversarial annotator (i.e., attackers) manipulates the ranking score by up-ranking any malicious text to steer the LLM adversarially. To assess the red-teaming of RLHF against human preference data poisoning, we propose RankPoison, a poisoning attack method on candidates{'} selection of preference rank flipping to reach certain malicious behaviors (e.g., generating longer sequences, which can increase the computational cost). With poisoned dataset generated by RankPoison, we can perform poisoning attacks on LLMs to generate longer tokens without hurting the original safety alignment performance. Moreover, applying RankPoison, we also successfully implement a backdoor attack where LLMs can generate longer answers under questions with the trigger word. Our findings highlight critical security challenges in RLHF, underscoring the necessity for more robust alignment methods for LLMs.", }

Reinforcement Learning with Human Feedback (RLHF) is a methodology designed to align Large Language Models (LLMs) with human preferences, playing an important role in LLMs alignment. Despite its advantages, RLHF relies on human annotators to rank the text, which can introduce potential security vulnerabilities if any adversarial annotator (i.e., attackers) manipulates the ranking score by up-ranking any malicious text to steer the LLM adversarially. To assess the red-teaming of RLHF against human preference data poisoning, we propose RankPoison, a poisoning attack method on candidates{'} selection of preference rank flipping to reach certain malicious behaviors (e.g., generating longer sequences, which can increase the computational cost). With poisoned dataset generated by RankPoison, we can perform poisoning attacks on LLMs to generate longer tokens without hurting the original safety alignment performance. Moreover, applying RankPoison, we also successfully implement a backdoor attack where LLMs can generate longer answers under questions with the trigger word. Our findings highlight critical security challenges in RLHF, underscoring the necessity for more robust alignment methods for LLMs.

RLHFPoison: Reward Poisoning Attack for Reinforcement Learning with Human Feedback in Large Language Models

acl-long.140

Poster

2311.09641

https://huggingface.co/papers/2311.09641

https://aclanthology.org/2024.acl-long.140/

https://aclanthology.org/2024.acl-long.141.bib

@inproceedings{nylund-etal-2024-time, title = "Time is Encoded in the Weights of Finetuned Language Models", author = "Nylund, Kai and Gururangan, Suchin and Smith, Noah", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.141", pages = "2571--2587", abstract = "We present time vectors, a simple tool to customize language models to new time periods. Time vectors are created by finetuning a language model on data from a single time (e.g., a year or month), and then subtracting the weights of the original pretrained model. This vector specifies a direction in weight space that, as our experiments show, improves performance on text from that time period. Time vectors specialized to adjacent time periods appear to be positioned closer together in a manifold. Using this structure, we interpolate between time vectors to induce new models that perform better on intervening and future time periods, without any additional training. We demonstrate the consistency of our findings across different tasks, domains, model sizes, and time scales. Our results suggest that time is encoded in the weight space of finetuned models.", }

We present time vectors, a simple tool to customize language models to new time periods. Time vectors are created by finetuning a language model on data from a single time (e.g., a year or month), and then subtracting the weights of the original pretrained model. This vector specifies a direction in weight space that, as our experiments show, improves performance on text from that time period. Time vectors specialized to adjacent time periods appear to be positioned closer together in a manifold. Using this structure, we interpolate between time vectors to induce new models that perform better on intervening and future time periods, without any additional training. We demonstrate the consistency of our findings across different tasks, domains, model sizes, and time scales. Our results suggest that time is encoded in the weight space of finetuned models.

Time is Encoded in the Weights of Finetuned Language Models

acl-long.141

Poster

2312.13401

https://huggingface.co/papers/2312.13401

https://aclanthology.org/2024.acl-long.141/

https://aclanthology.org/2024.acl-long.142.bib

@inproceedings{yen-etal-2024-long, title = "Long-Context Language Modeling with Parallel Context Encoding", author = "Yen, Howard and Gao, Tianyu and Chen, Danqi", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.142", pages = "2588--2610", abstract = "Extending large language models (LLMs) to process longer inputs is crucial for numerous applications. However, the considerable computational cost of transformers, coupled with limited generalization of positional encoding, restricts the size of their context window. We introduce Cross-Attention to Parallel Encodings (CAPE), a framework that can be applied to any existing decoder-only LLMs for context expansion. CAPE leverages a small encoder to process a long input chunk by chunk and enables the frozen decoder to cross-attend to the additional contexts. CAPE is efficient, generalizable, and versatile: trained with 8K-token documents, CAPE extends the context window of LLaMA-2 to 128K tokens, offering $10\times$ of the throughput with only 1/6 of the memory. CAPE yields strong performance on language modeling and in-context learning. CAPE also excels in retrieval-augmented applications, while existing long-context models degenerate with retrieved contexts. We further introduce a CAPE variant that can extend the context window of instruction-tuned models with only unlabeled data, and showcase its effectiveness on LLaMA-2-Chat, leading to a strong instruction-following model that can leverage very long context on downstream tasks.", }

Extending large language models (LLMs) to process longer inputs is crucial for numerous applications. However, the considerable computational cost of transformers, coupled with limited generalization of positional encoding, restricts the size of their context window. We introduce Cross-Attention to Parallel Encodings (CAPE), a framework that can be applied to any existing decoder-only LLMs for context expansion. CAPE leverages a small encoder to process a long input chunk by chunk and enables the frozen decoder to cross-attend to the additional contexts. CAPE is efficient, generalizable, and versatile: trained with 8K-token documents, CAPE extends the context window of LLaMA-2 to 128K tokens, offering $10\times$ of the throughput with only 1/6 of the memory. CAPE yields strong performance on language modeling and in-context learning. CAPE also excels in retrieval-augmented applications, while existing long-context models degenerate with retrieved contexts. We further introduce a CAPE variant that can extend the context window of instruction-tuned models with only unlabeled data, and showcase its effectiveness on LLaMA-2-Chat, leading to a strong instruction-following model that can leverage very long context on downstream tasks.

Long-Context Language Modeling with Parallel Context Encoding

acl-long.142

Poster

2402.16617

https://huggingface.co/papers/2402.16617

https://aclanthology.org/2024.acl-long.142/

https://aclanthology.org/2024.acl-long.143.bib

@inproceedings{yao-etal-2024-sirllm, title = "{S}ir{LLM}: Streaming Infinite Retentive {LLM}", author = "Yao, Yao and Li, Zuchao and Zhao, Hai", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.143", pages = "2611--2624", abstract = "As Large Language Models (LLMs) become increasingly prevalent in various domains, their ability to process inputs of any length and maintain a degree of memory becomes essential. However, the one-off input of overly long texts is limited, as studies have shown that when input lengths exceed the LLMs{'} pre-trained text length, there is a dramatic decline in text generation capabilities. Moreover, simply extending the length of pre-training texts is impractical due to the difficulty in obtaining long text data and the substantial memory consumption costs this would entail for LLMs. Recent efforts have employed streaming inputs to alleviate the pressure of excessively long text inputs, but this approach can significantly impair the model{'}s long-term memory capabilities.Motivated by this challenge, we introduce Streaming Infinite Retentive LLM (SirLLM), which allows LLMs to maintain longer memory during infinite-length dialogues without the need for fine-tuning. SirLLM utilizes the Token Entropy metric and a memory decay mechanism to filter key phrases, endowing LLMs with both long-lasting and flexible memory. We designed three distinct tasks and constructed three datasets to measure the effectiveness of SirLLM from various angles: (1) DailyDialog; (2) Grocery Shopping; (3) Rock-Paper-Scissors. Our experimental results robustly demonstrate that SirLLM can achieve stable and significant improvements across different LLMs and tasks, compellingly proving its effectiveness. When having a coversation, {``}A sir could forget himself,{''} but SirLLM never does! Our code is publicly available at https://github.com/Zoeyyao27/SirLLMhttps://github.com/Zoeyyao27/SirLLM", }

As Large Language Models (LLMs) become increasingly prevalent in various domains, their ability to process inputs of any length and maintain a degree of memory becomes essential. However, the one-off input of overly long texts is limited, as studies have shown that when input lengths exceed the LLMs{'} pre-trained text length, there is a dramatic decline in text generation capabilities. Moreover, simply extending the length of pre-training texts is impractical due to the difficulty in obtaining long text data and the substantial memory consumption costs this would entail for LLMs. Recent efforts have employed streaming inputs to alleviate the pressure of excessively long text inputs, but this approach can significantly impair the model{'}s long-term memory capabilities.Motivated by this challenge, we introduce Streaming Infinite Retentive LLM (SirLLM), which allows LLMs to maintain longer memory during infinite-length dialogues without the need for fine-tuning. SirLLM utilizes the Token Entropy metric and a memory decay mechanism to filter key phrases, endowing LLMs with both long-lasting and flexible memory. We designed three distinct tasks and constructed three datasets to measure the effectiveness of SirLLM from various angles: (1) DailyDialog; (2) Grocery Shopping; (3) Rock-Paper-Scissors. Our experimental results robustly demonstrate that SirLLM can achieve stable and significant improvements across different LLMs and tasks, compellingly proving its effectiveness. When having a coversation, {``}A sir could forget himself,{''} but SirLLM never does! Our code is publicly available at https://github.com/Zoeyyao27/SirLLMhttps://github.com/Zoeyyao27/SirLLM

SirLLM: Streaming Infinite Retentive LLM

acl-long.143

Oral

2405.12528

https://huggingface.co/papers/2405.12528

https://aclanthology.org/2024.acl-long.143/

https://aclanthology.org/2024.acl-long.144.bib

@inproceedings{feng-etal-2024-imo, title = "{IMO}: Greedy Layer-Wise Sparse Representation Learning for Out-of-Distribution Text Classification with Pre-trained Models", author = "Feng, Tao and Qu, Lizhen and Li, Zhuang and Zhan, Haolan and Hua, Yuncheng and Haf, Reza", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.144", pages = "2625--2639", abstract = "Machine learning models have made incredible progress, but they still struggle when applied to examples from unseen domains. This study focuses on a specific problem of domain generalization, where a model is trained on one source domain and tested on multiple target domains that are unseen during training. We propose IMO: Invariant features Masks for Out-of-Distribution text classification, to achieve OOD generalization by learning invariant features. During training, IMO would learn sparse mask layers to remove irrelevant features for prediction, where the remaining features keep invariant. Additionally, IMO has an attention module at the token level to focus on tokens that are useful for prediction. Our comprehensive experiments show that IMO substantially outperforms strong baselines in terms of various evaluation metrics and settings.", }

Machine learning models have made incredible progress, but they still struggle when applied to examples from unseen domains. This study focuses on a specific problem of domain generalization, where a model is trained on one source domain and tested on multiple target domains that are unseen during training. We propose IMO: Invariant features Masks for Out-of-Distribution text classification, to achieve OOD generalization by learning invariant features. During training, IMO would learn sparse mask layers to remove irrelevant features for prediction, where the remaining features keep invariant. Additionally, IMO has an attention module at the token level to focus on tokens that are useful for prediction. Our comprehensive experiments show that IMO substantially outperforms strong baselines in terms of various evaluation metrics and settings.

IMO: Greedy Layer-Wise Sparse Representation Learning for Out-of-Distribution Text Classification with Pre-trained Models

acl-long.144

Poster

2404.13504

https://aclanthology.org/2024.acl-long.144/

https://aclanthology.org/2024.acl-long.145.bib

@inproceedings{hu-etal-2024-generative, title = "Generative Pretrained Structured Transformers: Unsupervised Syntactic Language Models at Scale", author = "Hu, Xiang and Ji, Pengyu and Zhu, Qingyang and Wu, Wei and Tu, Kewei", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.145", pages = "2640--2657", abstract = "A syntactic language model (SLM) incrementally generates a sentence with its syntactic tree in a left-to-right manner.We present Generative Pretrained Structured Transformers (GPST), an unsupervised SLM at scale capable of being pre-trained from scratch on raw texts with high parallelism. GPST circumvents the limitations of previous SLMs such as relying on gold trees and sequential training. It consists of two components, a usual SLM supervised by a uni-directional language modeling loss, and an additional composition model, which induces syntactic parse trees and computes constituent representations, supervised by a bi-directional language modeling loss. We propose a representation surrogate to enable joint parallel training of the two models in a hard-EM fashion.We pre-train GPST on OpenWebText, a corpus with billion tokens, and demonstrate the superiority of GPST over GPT-2 with a comparable size in numerous tasks covering both language understanding and language generation. Meanwhile, GPST also significantly outperforms existing unsupervised SLMs on left-to-right grammar induction, while holding a substantial acceleration on training.", }

A syntactic language model (SLM) incrementally generates a sentence with its syntactic tree in a left-to-right manner.We present Generative Pretrained Structured Transformers (GPST), an unsupervised SLM at scale capable of being pre-trained from scratch on raw texts with high parallelism. GPST circumvents the limitations of previous SLMs such as relying on gold trees and sequential training. It consists of two components, a usual SLM supervised by a uni-directional language modeling loss, and an additional composition model, which induces syntactic parse trees and computes constituent representations, supervised by a bi-directional language modeling loss. We propose a representation surrogate to enable joint parallel training of the two models in a hard-EM fashion.We pre-train GPST on OpenWebText, a corpus with billion tokens, and demonstrate the superiority of GPST over GPT-2 with a comparable size in numerous tasks covering both language understanding and language generation. Meanwhile, GPST also significantly outperforms existing unsupervised SLMs on left-to-right grammar induction, while holding a substantial acceleration on training.

Generative Pretrained Structured Transformers: Unsupervised Syntactic Language Models at Scale

acl-long.145

Poster

2403.08293

https://aclanthology.org/2024.acl-long.145/

https://aclanthology.org/2024.acl-long.146.bib

@inproceedings{zhang-etal-2024-mela, title = "{MELA}: Multilingual Evaluation of Linguistic Acceptability", author = "Zhang, Ziyin and Liu, Yikang and Huang, Weifang and Mao, Junyu and Wang, Rui and Hu, Hai", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.146", pages = "2658--2674", abstract = "In this work, we present the largest benchmark to date on linguistic acceptability: Multilingual Evaluation of Linguistic Acceptability{---}MELA, with 46K samples covering 10 languages from a diverse set of language families. We establish LLM baselines on this benchmark, and investigate cross-lingual transfer in acceptability judgements with XLM-R. In pursuit of multilingual interpretability, we conduct probing experiments with fine-tuned XLM-R to explore the process of syntax capability acquisition. Our results show that GPT-4o exhibits a strong multilingual ability, outperforming fine-tuned XLM-R, while open-source multilingual models lag behind by a noticeable gap. Cross-lingual transfer experiments show that transfer in acceptability judgment is non-trivial: 500 Icelandic fine-tuning examples lead to 23 MCC performance in a completely unrelated language{---}Chinese. Results of our probing experiments indicate that training on MELA improves the performance of XLM-R on syntax-related tasks.", }

In this work, we present the largest benchmark to date on linguistic acceptability: Multilingual Evaluation of Linguistic Acceptability{---}MELA, with 46K samples covering 10 languages from a diverse set of language families. We establish LLM baselines on this benchmark, and investigate cross-lingual transfer in acceptability judgements with XLM-R. In pursuit of multilingual interpretability, we conduct probing experiments with fine-tuned XLM-R to explore the process of syntax capability acquisition. Our results show that GPT-4o exhibits a strong multilingual ability, outperforming fine-tuned XLM-R, while open-source multilingual models lag behind by a noticeable gap. Cross-lingual transfer experiments show that transfer in acceptability judgment is non-trivial: 500 Icelandic fine-tuning examples lead to 23 MCC performance in a completely unrelated language{---}Chinese. Results of our probing experiments indicate that training on MELA improves the performance of XLM-R on syntax-related tasks.

MELA: Multilingual Evaluation of Linguistic Acceptability

acl-long.146

Poster

2311.09033

https://aclanthology.org/2024.acl-long.146/

https://aclanthology.org/2024.acl-long.147.bib

@inproceedings{zhou-etal-2024-copyne, title = "{C}opy{NE}: Better Contextual {ASR} by Copying Named Entities", author = "Zhou, Shilin and Li, Zhenghua and Hong, Yu and Zhang, Min and Wang, Zhefeng and Huai, Baoxing", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.147", pages = "2675--2686", abstract = "End-to-end automatic speech recognition (ASR) systems have made significant progress in general scenarios. However, it remains challenging to transcribe contextual named entities (NEs) in the contextual ASR scenario. Previous approaches have attempted to address this by utilizing the NE dictionary. These approaches treat entities as individual tokens and generate them token-by-token, which may result in incomplete transcriptions of entities. In this paper, we treat entities as indivisible wholes and introduce the idea of copying into ASR. We design a systematic mechanism called CopyNE, which can copy entities from the NE dictionary. By copying all tokens of an entity at once, we can reduce errors during entity transcription, ensuring the completeness of the entity. Experiments demonstrate that CopyNE consistently improves the accuracy of transcribing entities compared to previous approaches. Even when based on the strong Whisper, CopyNE still achieves notable improvements.", }

End-to-end automatic speech recognition (ASR) systems have made significant progress in general scenarios. However, it remains challenging to transcribe contextual named entities (NEs) in the contextual ASR scenario. Previous approaches have attempted to address this by utilizing the NE dictionary. These approaches treat entities as individual tokens and generate them token-by-token, which may result in incomplete transcriptions of entities. In this paper, we treat entities as indivisible wholes and introduce the idea of copying into ASR. We design a systematic mechanism called CopyNE, which can copy entities from the NE dictionary. By copying all tokens of an entity at once, we can reduce errors during entity transcription, ensuring the completeness of the entity. Experiments demonstrate that CopyNE consistently improves the accuracy of transcribing entities compared to previous approaches. Even when based on the strong Whisper, CopyNE still achieves notable improvements.

CopyNE: Better Contextual ASR by Copying Named Entities

acl-long.147

Poster

2305.12839

https://aclanthology.org/2024.acl-long.147/

https://aclanthology.org/2024.acl-long.148.bib

@inproceedings{chen-etal-2024-table, title = "Is Table Retrieval a Solved Problem? Exploring Join-Aware Multi-Table Retrieval", author = "Chen, Peter Baile and Zhang, Yi and Roth, Dan", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.148", pages = "2687--2699", abstract = "Retrieving relevant tables containing the necessary information to accurately answer a given question over tables is critical to open-domain question-answering (QA) systems. Previous methods assume the answer to such a question can be found either in a single table or multiple tables identified through question decomposition or rewriting. However, neither of these approaches is sufficient, as many questions require retrieving multiple tables and joining them through a join plan that cannot be discerned from the user query itself. If the join plan is not considered in the retrieval stage, the subsequent steps of reasoning and answering based on those retrieved tables are likely to be incorrect. To address this problem, we introduce a method that uncovers useful join relations for any query and database during table retrieval. We use a novel re-ranking method formulated as a mixed-integer program that considers not only table-query relevance but also table-table relevance that requires inferring join relationships. Our method outperforms the state-of-the-art approaches for table retrieval by up to 9.3{\%} in F1 score and for end-to-end QA by up to 5.4{\%} in accuracy.", }

Retrieving relevant tables containing the necessary information to accurately answer a given question over tables is critical to open-domain question-answering (QA) systems. Previous methods assume the answer to such a question can be found either in a single table or multiple tables identified through question decomposition or rewriting. However, neither of these approaches is sufficient, as many questions require retrieving multiple tables and joining them through a join plan that cannot be discerned from the user query itself. If the join plan is not considered in the retrieval stage, the subsequent steps of reasoning and answering based on those retrieved tables are likely to be incorrect. To address this problem, we introduce a method that uncovers useful join relations for any query and database during table retrieval. We use a novel re-ranking method formulated as a mixed-integer program that considers not only table-query relevance but also table-table relevance that requires inferring join relationships. Our method outperforms the state-of-the-art approaches for table retrieval by up to 9.3{\%} in F1 score and for end-to-end QA by up to 5.4{\%} in accuracy.

Is Table Retrieval a Solved Problem? Exploring Join-Aware Multi-Table Retrieval

acl-long.148

Poster

2404.09889

https://aclanthology.org/2024.acl-long.148/

https://aclanthology.org/2024.acl-long.149.bib

@inproceedings{chen-etal-2024-generalizing, title = "Generalizing Conversational Dense Retrieval via {LLM}-Cognition Data Augmentation", author = "Chen, Haonan and Dou, Zhicheng and Mao, Kelong and Liu, Jiongnan and Zhao, Ziliang", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.149", pages = "2700--2718", abstract = "Conversational search utilizes muli-turn natural language contexts to retrieve relevant passages. Existing conversational dense retrieval models mostly view a conversation as a fixed sequence of questions and responses, overlooking the severe data sparsity problem {--} that is, users can perform a conversation in various ways, and these alternate conversations are unrecorded. Consequently, they often struggle to generalize to diverse conversations in real-world scenarios. In this work, we propose a framework for generalizing Conversational dense retrieval via LLM-cognition data Augmentation (ConvAug). We first generate multi-level augmented conversations to capture the diverse nature of conversational contexts. Inspired by human cognition, we devise a cognition-aware prompting process to mitigate the generation of false positives, false negatives, and hallucinations. Moreover, we develop a difficulty-adaptive sample filter that selects challenging samples for complex conversations, thereby giving the model a larger learning space. A contrastive learning objective is then employed to train a better conversational context encoder. Extensive experiments conducted on four public datasets, under both normal and zero-shot settings, demonstrate the effectiveness, generalizability, and applicability of ConvAug. The code is released at https://github.com/haon-chen/ConvAug.", }

Conversational search utilizes muli-turn natural language contexts to retrieve relevant passages. Existing conversational dense retrieval models mostly view a conversation as a fixed sequence of questions and responses, overlooking the severe data sparsity problem {--} that is, users can perform a conversation in various ways, and these alternate conversations are unrecorded. Consequently, they often struggle to generalize to diverse conversations in real-world scenarios. In this work, we propose a framework for generalizing Conversational dense retrieval via LLM-cognition data Augmentation (ConvAug). We first generate multi-level augmented conversations to capture the diverse nature of conversational contexts. Inspired by human cognition, we devise a cognition-aware prompting process to mitigate the generation of false positives, false negatives, and hallucinations. Moreover, we develop a difficulty-adaptive sample filter that selects challenging samples for complex conversations, thereby giving the model a larger learning space. A contrastive learning objective is then employed to train a better conversational context encoder. Extensive experiments conducted on four public datasets, under both normal and zero-shot settings, demonstrate the effectiveness, generalizability, and applicability of ConvAug. The code is released at https://github.com/haon-chen/ConvAug.

Generalizing Conversational Dense Retrieval via LLM-Cognition Data Augmentation

acl-long.149

Poster

2402.07092

https://aclanthology.org/2024.acl-long.149/

https://aclanthology.org/2024.acl-long.150.bib

@inproceedings{sun-etal-2024-itd, title = "{I}t{D}: Large Language Models Can Teach Themselves Induction through Deduction", author = "Sun, Wangtao and Xu, Haotian and Yu, Xuanqing and Chen, Pei and He, Shizhu and Zhao, Jun and Liu, Kang", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.150", pages = "2719--2731", abstract = "Although Large Language Models (LLMs) are showing impressive performance on a wide range of Natural Language Processing tasks, researchers have found that they still have limited ability to conduct induction. Recent works mainly adopt {``}post processes{''} paradigms to improve the performance of LLMs on induction (e.g., the hypothesis search {\&} refinement methods), but their performance is still constrained by the inherent inductive capability of the LLMs. In this paper, we propose a novel framework, Induction through Deduction (ItD), to enable the LLMs to teach themselves induction through deduction. The ItD framework is composed of two main components: a Deductive Data Generation module to generate induction data and a Naive Bayesian Induction module to optimize the fine-tuning and decoding of LLMs. Our empirical results showcase the effectiveness of ItD on two induction benchmarks, achieving relative performance improvement of 36{\%} and 10{\%} compared with previous state-of-the-art, respectively. Our ablation study verifies the effectiveness of two key modules of ItD. We also verify the effectiveness of ItD across different LLMs and deductors. The data and code of this paper can be found at https://github.com/forangel2014/ItD.", }

Although Large Language Models (LLMs) are showing impressive performance on a wide range of Natural Language Processing tasks, researchers have found that they still have limited ability to conduct induction. Recent works mainly adopt {``}post processes{''} paradigms to improve the performance of LLMs on induction (e.g., the hypothesis search {\&} refinement methods), but their performance is still constrained by the inherent inductive capability of the LLMs. In this paper, we propose a novel framework, Induction through Deduction (ItD), to enable the LLMs to teach themselves induction through deduction. The ItD framework is composed of two main components: a Deductive Data Generation module to generate induction data and a Naive Bayesian Induction module to optimize the fine-tuning and decoding of LLMs. Our empirical results showcase the effectiveness of ItD on two induction benchmarks, achieving relative performance improvement of 36{\%} and 10{\%} compared with previous state-of-the-art, respectively. Our ablation study verifies the effectiveness of two key modules of ItD. We also verify the effectiveness of ItD across different LLMs and deductors. The data and code of this paper can be found at https://github.com/forangel2014/ItD.

ItD: Large Language Models Can Teach Themselves Induction through Deduction

acl-long.150

Poster

2403.05789

https://aclanthology.org/2024.acl-long.150/

https://aclanthology.org/2024.acl-long.151.bib

@inproceedings{lu-etal-2024-mathgenie, title = "{M}ath{G}enie: Generating Synthetic Data with Question Back-translation for Enhancing Mathematical Reasoning of {LLM}s", author = "Lu, Zimu and Zhou, Aojun and Ren, Houxing and Wang, Ke and Shi, Weikang and Pan, Junting and Zhan, Mingjie and Li, Hongsheng", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.151", pages = "2732--2747", abstract = "Large language models (LLMs) have exhibited great potential in mathematical reasoning. However, there remains a performance gap in this area between existing open-source models and closed-source models such as GPT-4. In this paper, we introduce MathGenie, a novel method for generating diverse and reliable math problems by leveraging the ground-truth solutions of the seed data. We augment these ground-truth solutions and use a specially finetuned model to translate these augmented solutions back into new questions. Subsequently, we generate code-integrated solutions for these questions. To ensure the correctness of the code-integrated solutions, we employ rationale-based verification for filtering. Then, we finetune various pretrained models, ranging from 7B to 70B, on the newly curated data, resulting in a family of models known as MathGenie. These models consistently outperform previous open-source models across five representative mathematical reasoning datasets, achieving state-of-the-art performance. In particular, MathGenie-InternLM2 achieves an accuracy of 87.7{\%} on GSM8K and 55.7{\%} on MATH, securing the best overall score.", }

Large language models (LLMs) have exhibited great potential in mathematical reasoning. However, there remains a performance gap in this area between existing open-source models and closed-source models such as GPT-4. In this paper, we introduce MathGenie, a novel method for generating diverse and reliable math problems by leveraging the ground-truth solutions of the seed data. We augment these ground-truth solutions and use a specially finetuned model to translate these augmented solutions back into new questions. Subsequently, we generate code-integrated solutions for these questions. To ensure the correctness of the code-integrated solutions, we employ rationale-based verification for filtering. Then, we finetune various pretrained models, ranging from 7B to 70B, on the newly curated data, resulting in a family of models known as MathGenie. These models consistently outperform previous open-source models across five representative mathematical reasoning datasets, achieving state-of-the-art performance. In particular, MathGenie-InternLM2 achieves an accuracy of 87.7{\%} on GSM8K and 55.7{\%} on MATH, securing the best overall score.

MathGenie: Generating Synthetic Data with Question Back-translation for Enhancing Mathematical Reasoning of LLMs

acl-long.151

Poster

2402.16352

https://huggingface.co/papers/2402.16352

https://aclanthology.org/2024.acl-long.151/

https://aclanthology.org/2024.acl-long.152.bib

@inproceedings{xu-etal-2024-rethinking, title = "Rethinking Task-Oriented Dialogue Systems: From Complex Modularity to Zero-Shot Autonomous Agent", author = "Xu, Heng-Da and Mao, Xian-Ling and Yang, Puhai and Sun, Fanshu and Huang, Heyan", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.152", pages = "2748--2763", abstract = "Task-oriented dialogue (TOD) systems are predominantly designed to be composed of several functional modules (e.g. dialogue state tracker, dialogue policy, natural language generation) whether they are pipeline or end-to-end architectures. However, this modular design not only heavily relies on massive fully-annotated data, but also suffers from many intrinsic drawbacks, such as serious error accumulation, poor generalization ability, high customization cost, and low fault tolerance rate. In this paper, we rethink the architecture of the task-oriented dialogue systems and propose a novel fully zero-shot autonomous TOD agent, named AutoTOD, where all the delicate modules in traditional TOD systems are deprecated and all it needs is a general-purpose instruction-following language model (e.g. GPT-4). AutoTOD only leverages a simple instruction schema consisting of the description of tasks and external APIs, and can autonomously decide what to do at each dialogue turn, including asking for information, calling APIs, summarizing API results, and correcting previous mistakes. Moreover, we propose a simulation-based evaluation framework to better validate the abilities of TOD models in real-life scenarios. Extensive experiments conducted on the MultiWOZ and SGD datasets show the superior task completion ability and flexible language skills of AutoTOD.", }

Task-oriented dialogue (TOD) systems are predominantly designed to be composed of several functional modules (e.g. dialogue state tracker, dialogue policy, natural language generation) whether they are pipeline or end-to-end architectures. However, this modular design not only heavily relies on massive fully-annotated data, but also suffers from many intrinsic drawbacks, such as serious error accumulation, poor generalization ability, high customization cost, and low fault tolerance rate. In this paper, we rethink the architecture of the task-oriented dialogue systems and propose a novel fully zero-shot autonomous TOD agent, named AutoTOD, where all the delicate modules in traditional TOD systems are deprecated and all it needs is a general-purpose instruction-following language model (e.g. GPT-4). AutoTOD only leverages a simple instruction schema consisting of the description of tasks and external APIs, and can autonomously decide what to do at each dialogue turn, including asking for information, calling APIs, summarizing API results, and correcting previous mistakes. Moreover, we propose a simulation-based evaluation framework to better validate the abilities of TOD models in real-life scenarios. Extensive experiments conducted on the MultiWOZ and SGD datasets show the superior task completion ability and flexible language skills of AutoTOD.

Rethinking Task-Oriented Dialogue Systems: From Complex Modularity to Zero-Shot Autonomous Agent

acl-long.152

Poster

https://aclanthology.org/2024.acl-long.152/

https://aclanthology.org/2024.acl-long.153.bib

@inproceedings{ravaut-etal-2024-context, title = "On Context Utilization in Summarization with Large Language Models", author = "Ravaut, Mathieu and Sun, Aixin and Chen, Nancy and Joty, Shafiq", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.153", pages = "2764--2781", abstract = "Large language models (LLMs) excel in abstractive summarization tasks, delivering fluent and pertinent summaries. Recent advancements have extended their capabilities to handle long-input contexts, exceeding 100k tokens. However, in question answering, language models exhibit uneven utilization of their input context. They tend to favor the initial and final segments, resulting in a U-shaped performance pattern concerning where the answer is located within the input. This bias raises concerns, particularly in summarization where crucial content may be dispersed throughout the source document(s). Besides, in summarization, mapping facts from the source to the summary is not trivial as salient content is usually re-phrased. In this paper, we conduct the first comprehensive study on context utilization and position bias in summarization. Our analysis encompasses 6 LLMs, 10 datasets, and 5 evaluation metrics. We introduce a new evaluation benchmark called MiddleSum on the which we benchmark two alternative inference methods to alleviate position bias: hierarchical summarization and incremental summarization. Our code and data can be found here: https://github.com/ntunlp/MiddleSum.", }

Large language models (LLMs) excel in abstractive summarization tasks, delivering fluent and pertinent summaries. Recent advancements have extended their capabilities to handle long-input contexts, exceeding 100k tokens. However, in question answering, language models exhibit uneven utilization of their input context. They tend to favor the initial and final segments, resulting in a U-shaped performance pattern concerning where the answer is located within the input. This bias raises concerns, particularly in summarization where crucial content may be dispersed throughout the source document(s). Besides, in summarization, mapping facts from the source to the summary is not trivial as salient content is usually re-phrased. In this paper, we conduct the first comprehensive study on context utilization and position bias in summarization. Our analysis encompasses 6 LLMs, 10 datasets, and 5 evaluation metrics. We introduce a new evaluation benchmark called MiddleSum on the which we benchmark two alternative inference methods to alleviate position bias: hierarchical summarization and incremental summarization. Our code and data can be found here: https://github.com/ntunlp/MiddleSum.

On Context Utilization in Summarization with Large Language Models

acl-long.153

Poster

2310.10570

https://aclanthology.org/2024.acl-long.153/

https://aclanthology.org/2024.acl-long.154.bib

@inproceedings{zhu-etal-2024-inters, title = "{INTERS}: Unlocking the Power of Large Language Models in Search with Instruction Tuning", author = "Zhu, Yutao and Zhang, Peitian and Zhang, Chenghao and Chen, Yifei and Xie, Binyu and Liu, Zheng and Wen, Ji-Rong and Dou, Zhicheng", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.154", pages = "2782--2809", abstract = "Large language models (LLMs) have demonstrated impressive capabilities in various natural language processing tasks. Despite this, their application to information retrieval (IR) tasks is still challenging due to the infrequent occurrence of many IR-specific concepts in natural language. While prompt-based methods can provide task descriptions to LLMs, they often fall short in facilitating a comprehensive understanding and execution of IR tasks, thereby limiting LLMs{'} applicability. To address this gap, in this work, we explore the potential of instruction tuning to enhance LLMs{'} proficiency in IR tasks. We introduce a novel instruction tuning dataset, INTERS, encompassing 20 tasks across three fundamental IR categories: query understanding, document understanding, and query-document relationship understanding. The data are derived from 43 distinct datasets with manually written templates. Our empirical results reveal that INTERS significantly boosts the performance of various publicly available LLMs, such as LLaMA, Mistral, and Falcon, in IR tasks. Furthermore, we conduct extensive experiments to analyze the effects of instruction design, template diversity, few-shot demonstrations, and the volume of instructions on performance. We make our dataset and the fine-tuned models publicly accessible at https://github.com/DaoD/INTERS.", }

Large language models (LLMs) have demonstrated impressive capabilities in various natural language processing tasks. Despite this, their application to information retrieval (IR) tasks is still challenging due to the infrequent occurrence of many IR-specific concepts in natural language. While prompt-based methods can provide task descriptions to LLMs, they often fall short in facilitating a comprehensive understanding and execution of IR tasks, thereby limiting LLMs{'} applicability. To address this gap, in this work, we explore the potential of instruction tuning to enhance LLMs{'} proficiency in IR tasks. We introduce a novel instruction tuning dataset, INTERS, encompassing 20 tasks across three fundamental IR categories: query understanding, document understanding, and query-document relationship understanding. The data are derived from 43 distinct datasets with manually written templates. Our empirical results reveal that INTERS significantly boosts the performance of various publicly available LLMs, such as LLaMA, Mistral, and Falcon, in IR tasks. Furthermore, we conduct extensive experiments to analyze the effects of instruction design, template diversity, few-shot demonstrations, and the volume of instructions on performance. We make our dataset and the fine-tuned models publicly accessible at https://github.com/DaoD/INTERS.

INTERS: Unlocking the Power of Large Language Models in Search with Instruction Tuning

acl-long.154

Poster

2401.06532

https://huggingface.co/papers/2401.06532

https://aclanthology.org/2024.acl-long.154/

https://aclanthology.org/2024.acl-long.155.bib

@inproceedings{zhou-etal-2024-enhancing-context, title = "Enhancing In-Context Learning via Implicit Demonstration Augmentation", author = "Zhou, Xiaoling and Ye, Wei and Wang, Yidong and Jiang, Chaoya and Lee, Zhemg and Xie, Rui and Zhang, Shikun", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.155", pages = "2810--2828", abstract = "The emergence of in-context learning (ICL) enables large pre-trained language models (PLMs) to make predictions for unseen inputs without updating parameters. Despite its potential, ICL{'}s effectiveness heavily relies on the quality, quantity, and permutation of demonstrations, commonly leading to suboptimal and unstable performance. In this paper, we tackle this challenge for the first time from the perspective of demonstration augmentation. Specifically, we start with enriching representations of demonstrations by leveraging their deep feature distribution. We then theoretically reveal that when the number of augmented copies approaches infinity, the augmentation is approximately equal to a novel logit calibration mechanism integrated with specific statistical properties. This insight results in a simple yet highly efficient method that significantly improves the average and worst-case accuracy across diverse PLMs and tasks. Moreover, our method effectively reduces performance variance among varying demonstrations, permutations, and templates, and displays the capability to address imbalanced class distributions.", }

The emergence of in-context learning (ICL) enables large pre-trained language models (PLMs) to make predictions for unseen inputs without updating parameters. Despite its potential, ICL{'}s effectiveness heavily relies on the quality, quantity, and permutation of demonstrations, commonly leading to suboptimal and unstable performance. In this paper, we tackle this challenge for the first time from the perspective of demonstration augmentation. Specifically, we start with enriching representations of demonstrations by leveraging their deep feature distribution. We then theoretically reveal that when the number of augmented copies approaches infinity, the augmentation is approximately equal to a novel logit calibration mechanism integrated with specific statistical properties. This insight results in a simple yet highly efficient method that significantly improves the average and worst-case accuracy across diverse PLMs and tasks. Moreover, our method effectively reduces performance variance among varying demonstrations, permutations, and templates, and displays the capability to address imbalanced class distributions.

Enhancing In-Context Learning via Implicit Demonstration Augmentation

acl-long.155

Poster

https://aclanthology.org/2024.acl-long.155/

https://aclanthology.org/2024.acl-long.156.bib

@inproceedings{wang-etal-2024-prolora, title = "{PR}o{L}o{RA}: Partial Rotation Empowers More Parameter-Efficient {L}o{RA}", author = "Wang, Sheng and Xue, Boyang and Ye, Jiacheng and Jiang, Jiyue and Chen, Liheng and Kong, Lingpeng and Wu, Chuan", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.156", pages = "2829--2841", abstract = "With the rapid scaling of large language models (LLMs), serving numerouslow-rank adaptations (LoRAs) concurrently has become increasingly impractical,leading to unaffordable costs and necessitating more parameter-efficientfinetuning methods. In this work, we introduce $\text{\textbf{P}artially \textbf{Ro}tation-enhanced \textbf{Lo}w-\textbf{R}ank \textbf{A}daptation (PRoLoRA)}$, an intra-layer sharing mechanism comprising fouressential components: broadcast reduction, rotation enhancement,partially-sharing refinement, and rectified initialization strategy. As asuperset of LoRA, PRoLoRA retains its advantages, and effectively circumventthe drawbacks of peer parameter-sharing methods with superior model capacity,practical feasibility, and broad applicability. Empirical experimentsdemonstrate the remarkably higher parameter efficiency of PRoLoRA in bothspecific parameter budget and performance target scenarios, and its scalabilityto larger LLMs. Notably, with one time less trainable parameters, PRoLoRA stilloutperforms LoRA on multiple instruction tuning datasets. Subsequently, anablation study is conducted to validate the necessity of individual componentsand highlight the superiority of PRoLoRA over three potential variants.Hopefully, the conspicuously higher parameter efficiency can establish PRoLoRAas a resource-friendly alternative to LoRA.", }

With the rapid scaling of large language models (LLMs), serving numerouslow-rank adaptations (LoRAs) concurrently has become increasingly impractical,leading to unaffordable costs and necessitating more parameter-efficientfinetuning methods. In this work, we introduce $\text{\textbf{P}artially \textbf{Ro}tation-enhanced \textbf{Lo}w-\textbf{R}ank \textbf{A}daptation (PRoLoRA)}$, an intra-layer sharing mechanism comprising fouressential components: broadcast reduction, rotation enhancement,partially-sharing refinement, and rectified initialization strategy. As asuperset of LoRA, PRoLoRA retains its advantages, and effectively circumventthe drawbacks of peer parameter-sharing methods with superior model capacity,practical feasibility, and broad applicability. Empirical experimentsdemonstrate the remarkably higher parameter efficiency of PRoLoRA in bothspecific parameter budget and performance target scenarios, and its scalabilityto larger LLMs. Notably, with one time less trainable parameters, PRoLoRA stilloutperforms LoRA on multiple instruction tuning datasets. Subsequently, anablation study is conducted to validate the necessity of individual componentsand highlight the superiority of PRoLoRA over three potential variants.Hopefully, the conspicuously higher parameter efficiency can establish PRoLoRAas a resource-friendly alternative to LoRA.

PRoLoRA: Partial Rotation Empowers More Parameter-Efficient LoRA

acl-long.156

Poster

2402.16902

https://aclanthology.org/2024.acl-long.156/

https://aclanthology.org/2024.acl-long.157.bib

@inproceedings{cai-etal-2024-improving-event, title = "Improving Event Definition Following For Zero-Shot Event Detection", author = "Cai, Zefan and Kung, Po-Nien and Suvarna, Ashima and Ma, Mingyu and Bansal, Hritik and Chang, Baobao and Brantingham, P. Jeffrey and Wang, Wei and Peng, Nanyun", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.157", pages = "2842--2863", abstract = "Existing approaches on zero-shot event detection usually train models on datasets annotated with known event types, and prompt them with unseen event definitions. These approaches yield sporadic successes, yet generally fall short of expectations.In this work, we aim to improve zero-shot event detection by training models to better follow event definitions. We hypothesize that a diverse set of event types and definitions are the key for models to learn to follow event definitions while existing event extraction datasets focus on annotating many high-quality examples for a few event types. To verify our hypothesis, we construct an automatically generated Diverse Event Definition (DivED) dataset and conduct comparative studies. Our experiments reveal that a large number of event types (200) and diverse event definitions can significantly boost event extraction performance; on the other hand, the performance does not scale with over ten examples per event type.Beyond scaling, we incorporate event ontology information and hard-negative samples during training, further boosting the performance. Based on these findings, we fine-tuned a LLaMA-2-7B model on our DivED dataset, yielding performance that surpasses SOTA large language models like GPT-3.5 across three open benchmarks on zero-shot event detection.", }

Existing approaches on zero-shot event detection usually train models on datasets annotated with known event types, and prompt them with unseen event definitions. These approaches yield sporadic successes, yet generally fall short of expectations.In this work, we aim to improve zero-shot event detection by training models to better follow event definitions. We hypothesize that a diverse set of event types and definitions are the key for models to learn to follow event definitions while existing event extraction datasets focus on annotating many high-quality examples for a few event types. To verify our hypothesis, we construct an automatically generated Diverse Event Definition (DivED) dataset and conduct comparative studies. Our experiments reveal that a large number of event types (200) and diverse event definitions can significantly boost event extraction performance; on the other hand, the performance does not scale with over ten examples per event type.Beyond scaling, we incorporate event ontology information and hard-negative samples during training, further boosting the performance. Based on these findings, we fine-tuned a LLaMA-2-7B model on our DivED dataset, yielding performance that surpasses SOTA large language models like GPT-3.5 across three open benchmarks on zero-shot event detection.

Improving Event Definition Following For Zero-Shot Event Detection

acl-long.157

Poster

2403.02586

https://aclanthology.org/2024.acl-long.157/

https://aclanthology.org/2024.acl-long.158.bib

@inproceedings{wei-etal-2024-mud, title = "Through the {MUD}: A Multi-Defendant Charge Prediction Benchmark with Linked Crime Elements", author = "Wei, Xiao and Xu, Qi and Yu, Hang and Liu, Qian and Cambria, Erik", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.158", pages = "2864--2878", abstract = "The current charge prediction datasets mostly focus on single-defendant criminal cases.However, real-world criminal cases usually involve multiple defendants whose criminal facts are intertwined. In an early attempt to fill this gap, we introduce a new benchmark that encompasses legal cases involving multiple defendants, where each defendant is labeled with a charge and four types of crime elements, \textit{i.e.,} \textit{Object Element}, \textit{Objective Element}, \textit{Subject Element}, and \textit{Subjective Element}. Based on the dataset, we further develop an interpretable model called EJudge that incorporates crime elements and legal rules to infer charges. We observe that predicting crime charges while providing corresponding rationales benefits the interpretable AI system. Extensive experiments show that EJudge significantly surpasses state-of-the-art methods, which verify the importance of crime elements and legal rules in multi-defendant charge prediction. The source code and dataset are available at \url{https://anonymous.4open.science/r/MCP_1-6010}.", }

The current charge prediction datasets mostly focus on single-defendant criminal cases.However, real-world criminal cases usually involve multiple defendants whose criminal facts are intertwined. In an early attempt to fill this gap, we introduce a new benchmark that encompasses legal cases involving multiple defendants, where each defendant is labeled with a charge and four types of crime elements, \textit{i.e.,} \textit{Object Element}, \textit{Objective Element}, \textit{Subject Element}, and \textit{Subjective Element}. Based on the dataset, we further develop an interpretable model called EJudge that incorporates crime elements and legal rules to infer charges. We observe that predicting crime charges while providing corresponding rationales benefits the interpretable AI system. Extensive experiments show that EJudge significantly surpasses state-of-the-art methods, which verify the importance of crime elements and legal rules in multi-defendant charge prediction. The source code and dataset are available at \url{https://anonymous.4open.science/r/MCP_1-6010}.

Through the MUD: A Multi-Defendant Charge Prediction Benchmark with Linked Crime Elements

acl-long.158

Poster

https://aclanthology.org/2024.acl-long.158/

https://aclanthology.org/2024.acl-long.159.bib

@inproceedings{cheng-etal-2024-interpreting, title = "Interpreting Conversational Dense Retrieval by Rewriting-Enhanced Inversion of Session Embedding", author = "Cheng, Yiruo and Mao, Kelong and Dou, Zhicheng", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.159", pages = "2879--2893", abstract = "Conversational dense retrieval has shown to be effective in conversational search. However, a major limitation of conversational dense retrieval is their lack of interpretability, hindering intuitive understanding of model behaviors for targeted improvements. This paper presents CONVINV, a simple yet effective approach to shed light on interpretable conversational dense retrieval models. CONVINV transforms opaque conversational session embeddings into explicitly interpretable text while faithfully maintaining their original retrieval performance as much as possible. Such transformation is achieved by training a recently proposed Vec2Text model based on the ad-hoc query encoder, leveraging the fact that the session and query embeddings share the same space in existing conversational dense retrieval.To further enhance interpretability, we propose to incorporate external interpretable query rewrites into the transformation process. Extensive evaluations on three conversational search benchmarks demonstrate that CONVINV can yield more interpretable text and faithfully preserve original retrieval performance than baselines. Our work connects opaque session embeddings with transparent query rewriting, paving the way toward trustworthy conversational search.", }

Conversational dense retrieval has shown to be effective in conversational search. However, a major limitation of conversational dense retrieval is their lack of interpretability, hindering intuitive understanding of model behaviors for targeted improvements. This paper presents CONVINV, a simple yet effective approach to shed light on interpretable conversational dense retrieval models. CONVINV transforms opaque conversational session embeddings into explicitly interpretable text while faithfully maintaining their original retrieval performance as much as possible. Such transformation is achieved by training a recently proposed Vec2Text model based on the ad-hoc query encoder, leveraging the fact that the session and query embeddings share the same space in existing conversational dense retrieval.To further enhance interpretability, we propose to incorporate external interpretable query rewrites into the transformation process. Extensive evaluations on three conversational search benchmarks demonstrate that CONVINV can yield more interpretable text and faithfully preserve original retrieval performance than baselines. Our work connects opaque session embeddings with transparent query rewriting, paving the way toward trustworthy conversational search.

Interpreting Conversational Dense Retrieval by Rewriting-Enhanced Inversion of Session Embedding

acl-long.159

Poster

2402.12774

https://aclanthology.org/2024.acl-long.159/

https://aclanthology.org/2024.acl-long.160.bib

@inproceedings{wang-etal-2024-stumbling, title = "Stumbling Blocks: Stress Testing the Robustness of Machine-Generated Text Detectors Under Attacks", author = "Wang, Yichen and Feng, Shangbin and Hou, Abe and Pu, Xiao and Shen, Chao and Liu, Xiaoming and Tsvetkov, Yulia and He, Tianxing", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.160", pages = "2894--2925", abstract = "The widespread use of large language models (LLMs) is increasing the demand for methods that detect machine-generated text to prevent misuse. The goal of our study is to stress test the detectors{'} robustness to malicious attacks under realistic scenarios. We comprehensively study the robustness of popular machine-generated text detectors under attacks from diverse categories: editing, paraphrasing, co-generating, and prompting. Our attacks assume limited access to the generator LLMs, and we compare the performance of detectors on different attacks under different budget levels. Our experiments reveal that almost none of the existing detectors remain robust under all the attacks, and all detectors exhibit different loopholes. Averaging all detectors, the performance drops by 35{\%} across all attacks. Further, we investigate the reasons behind these defects and propose initial out-of-the-box patches.", }

The widespread use of large language models (LLMs) is increasing the demand for methods that detect machine-generated text to prevent misuse. The goal of our study is to stress test the detectors{'} robustness to malicious attacks under realistic scenarios. We comprehensively study the robustness of popular machine-generated text detectors under attacks from diverse categories: editing, paraphrasing, co-generating, and prompting. Our attacks assume limited access to the generator LLMs, and we compare the performance of detectors on different attacks under different budget levels. Our experiments reveal that almost none of the existing detectors remain robust under all the attacks, and all detectors exhibit different loopholes. Averaging all detectors, the performance drops by 35{\%} across all attacks. Further, we investigate the reasons behind these defects and propose initial out-of-the-box patches.

Stumbling Blocks: Stress Testing the Robustness of Machine-Generated Text Detectors Under Attacks

acl-long.160

Poster

2402.11638

https://huggingface.co/papers/2402.11638

https://aclanthology.org/2024.acl-long.160/

https://aclanthology.org/2024.acl-long.161.bib

@inproceedings{huang-etal-2024-training, title = "Training Language Models to Generate Text with Citations via Fine-grained Rewards", author = "Huang, Chengyu and Wu, Zeqiu and Hu, Yushi and Wang, Wenya", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.161", pages = "2926--2949", abstract = "While recent Large Language Models (LLMs) have proven useful in answering user queries, they are prone to hallucination, and their responses often lack credibility due to missing references to reliable sources. An intuitive solution to these issues would be to include in-text citations referring to external documents as evidence. While previous works have directly prompted LLMs to generate in-text citations, their performances are far from satisfactory, especially when it comes to smaller LLMs. In this work, we propose an effective training framework using fine-grained rewards to teach LLMs to generate highly supportive and relevant citations, while ensuring the correctness of their responses. We also conduct a systematic analysis of applying these fine-grained rewards to common LLM training strategies, demonstrating its advantage over conventional practices. We conduct extensive experiments on Question Answering (QA) datasets taken from the ALCE benchmark and validate the model{'}s generalizability using EXPERTQA. On LLaMA-2-7B, the incorporation of fine-grained rewards achieves the best performance among the baselines, even surpassing that of GPT-3.5-turbo.", }

While recent Large Language Models (LLMs) have proven useful in answering user queries, they are prone to hallucination, and their responses often lack credibility due to missing references to reliable sources. An intuitive solution to these issues would be to include in-text citations referring to external documents as evidence. While previous works have directly prompted LLMs to generate in-text citations, their performances are far from satisfactory, especially when it comes to smaller LLMs. In this work, we propose an effective training framework using fine-grained rewards to teach LLMs to generate highly supportive and relevant citations, while ensuring the correctness of their responses. We also conduct a systematic analysis of applying these fine-grained rewards to common LLM training strategies, demonstrating its advantage over conventional practices. We conduct extensive experiments on Question Answering (QA) datasets taken from the ALCE benchmark and validate the model{'}s generalizability using EXPERTQA. On LLaMA-2-7B, the incorporation of fine-grained rewards achieves the best performance among the baselines, even surpassing that of GPT-3.5-turbo.

Training Language Models to Generate Text with Citations via Fine-grained Rewards

acl-long.161

Poster

2402.04315

https://huggingface.co/papers/2402.04315

https://aclanthology.org/2024.acl-long.161/

https://aclanthology.org/2024.acl-long.162.bib

@inproceedings{li-etal-2024-hypergraph, title = "Hypergraph based Understanding for Document Semantic Entity Recognition", author = "Li, Qiwei and Li, Zuchao and Wang, Ping and Ai, Haojun and Zhao, Hai", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.162", pages = "2950--2960", abstract = "Semantic entity recognition is an important task in the field of visually-rich document understanding. It distinguishes the semantic types of text by analyzing the position relationship between text nodes and the relation between text content. The existing document understanding models mainly focus on entity categories while ignoring the extraction of entity boundaries. We build a novel hypergraph attention document semantic entity recognition framework, HGA, which uses hypergraph attention to focus on entity boundaries and entity categories at the same time. It can conduct a more detailed analysis of the document text representation analyzed by the upstream model and achieves a better performance of semantic information. We apply this method on the basis of GraphLayoutLM to construct a new semantic entity recognition model HGALayoutLM. Our experiment results on FUNSD, CORD, XFUND and SROIE show that our method can effectively improve the performance of semantic entity recognition tasks based on the original model. The results of HGALayoutLM on FUNSD and XFUND reach the new state-of-the-art results.", }

Semantic entity recognition is an important task in the field of visually-rich document understanding. It distinguishes the semantic types of text by analyzing the position relationship between text nodes and the relation between text content. The existing document understanding models mainly focus on entity categories while ignoring the extraction of entity boundaries. We build a novel hypergraph attention document semantic entity recognition framework, HGA, which uses hypergraph attention to focus on entity boundaries and entity categories at the same time. It can conduct a more detailed analysis of the document text representation analyzed by the upstream model and achieves a better performance of semantic information. We apply this method on the basis of GraphLayoutLM to construct a new semantic entity recognition model HGALayoutLM. Our experiment results on FUNSD, CORD, XFUND and SROIE show that our method can effectively improve the performance of semantic entity recognition tasks based on the original model. The results of HGALayoutLM on FUNSD and XFUND reach the new state-of-the-art results.

Hypergraph based Understanding for Document Semantic Entity Recognition

acl-long.162

Poster

2407.06904

https://aclanthology.org/2024.acl-long.162/

https://aclanthology.org/2024.acl-long.163.bib

@inproceedings{li-etal-2024-gsm, title = "{GSM}-Plus: A Comprehensive Benchmark for Evaluating the Robustness of {LLM}s as Mathematical Problem Solvers", author = "Li, Qintong and Cui, Leyang and Zhao, Xueliang and Kong, Lingpeng and Bi, Wei", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.163", pages = "2961--2984", abstract = "Large language models (LLMs) have achieved impressive performance across various mathematical reasoning benchmarks. However, there are increasing debates regarding whether these models truly understand and apply mathematical knowledge or merely rely on shortcuts for mathematical reasoning. One essential and frequently occurring evidence is that when the math questions are slightly changed, LLMs can behave incorrectly. This motivates us to evaluate the robustness of LLMs{'} math reasoning capability by testing a wide range of question variations. We introduce the adversarial grade school math (GSM-Plus) dataset, an extension of GSM8K augmented with various mathematical perturbations. Our experiments on 25 LLMs and 4 prompting techniques show that while LLMs exhibit different levels of math reasoning abilities, their performances are far from robust. In particular, even for problems that have been solved in GSM8K, LLMs can make mistakes when new statements are added or the question targets are altered. We also explore whether more robust performance can be achieved by composing existing prompting methods, in which we try an iterative method that generates and verifies each intermediate thought based on its reasoning goal and calculation result.", }

Large language models (LLMs) have achieved impressive performance across various mathematical reasoning benchmarks. However, there are increasing debates regarding whether these models truly understand and apply mathematical knowledge or merely rely on shortcuts for mathematical reasoning. One essential and frequently occurring evidence is that when the math questions are slightly changed, LLMs can behave incorrectly. This motivates us to evaluate the robustness of LLMs{'} math reasoning capability by testing a wide range of question variations. We introduce the adversarial grade school math (GSM-Plus) dataset, an extension of GSM8K augmented with various mathematical perturbations. Our experiments on 25 LLMs and 4 prompting techniques show that while LLMs exhibit different levels of math reasoning abilities, their performances are far from robust. In particular, even for problems that have been solved in GSM8K, LLMs can make mistakes when new statements are added or the question targets are altered. We also explore whether more robust performance can be achieved by composing existing prompting methods, in which we try an iterative method that generates and verifies each intermediate thought based on its reasoning goal and calculation result.

GSM-Plus: A Comprehensive Benchmark for Evaluating the Robustness of LLMs as Mathematical Problem Solvers

acl-long.163

Poster

2402.19255

https://huggingface.co/papers/2402.19255

https://aclanthology.org/2024.acl-long.163/

https://aclanthology.org/2024.acl-long.164.bib

@inproceedings{min-etal-2024-synergetic, title = "Synergetic Event Understanding: A Collaborative Approach to Cross-Document Event Coreference Resolution with Large Language Models", author = "Min, Qingkai and Guo, Qipeng and Hu, Xiangkun and Huang, Songfang and Zhang, Zheng and Zhang, Yue", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.164", pages = "2985--3002", abstract = "Cross-document event coreference resolution (CDECR) involves clustering event mentions across multiple documents that refer to the same real-world events. Existing approaches utilize fine-tuning of small language models (SLMs) like BERT to address the compatibility among the contexts of event mentions. However, due to the complexity and diversity of contexts, these models are prone to learning simple co-occurrences. Recently, large language models (LLMs) like ChatGPT have demonstrated impressive contextual understanding, yet they encounter challenges in adapting to specific information extraction (IE) tasks. In this paper, we propose a collaborative approach for CDECR, leveraging the capabilities of both a universally capable LLM and a task-specific SLM. The collaborative strategy begins with the LLM accurately and comprehensively summarizing events through prompting. Then, the SLM refines its learning of event representations based on these insights during fine-tuning. Experimental results demonstrate that our approach surpasses the performance of both the large and small language models individually, forming a complementary advantage. Across various datasets, our approach achieves state-of-the-art performance, underscoring its effectiveness in diverse scenarios.", }

Cross-document event coreference resolution (CDECR) involves clustering event mentions across multiple documents that refer to the same real-world events. Existing approaches utilize fine-tuning of small language models (SLMs) like BERT to address the compatibility among the contexts of event mentions. However, due to the complexity and diversity of contexts, these models are prone to learning simple co-occurrences. Recently, large language models (LLMs) like ChatGPT have demonstrated impressive contextual understanding, yet they encounter challenges in adapting to specific information extraction (IE) tasks. In this paper, we propose a collaborative approach for CDECR, leveraging the capabilities of both a universally capable LLM and a task-specific SLM. The collaborative strategy begins with the LLM accurately and comprehensively summarizing events through prompting. Then, the SLM refines its learning of event representations based on these insights during fine-tuning. Experimental results demonstrate that our approach surpasses the performance of both the large and small language models individually, forming a complementary advantage. Across various datasets, our approach achieves state-of-the-art performance, underscoring its effectiveness in diverse scenarios.

Synergetic Event Understanding: A Collaborative Approach to Cross-Document Event Coreference Resolution with Large Language Models

acl-long.164

Poster

2406.02148

https://aclanthology.org/2024.acl-long.164/

https://aclanthology.org/2024.acl-long.165.bib

@inproceedings{qiao-etal-2024-autoact, title = "{A}uto{A}ct: Automatic Agent Learning from Scratch for {QA} via Self-Planning", author = "Qiao, Shuofei and Zhang, Ningyu and Fang, Runnan and Luo, Yujie and Zhou, Wangchunshu and Jiang, Yuchen and Lv, Chengfei and Chen, Huajun", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.165", pages = "3003--3021", abstract = "Language agents have achieved considerable performance on various complex question-answering tasks by planning with external tools. Despite the incessant exploration in this field, existing language agent systems still struggle with costly, non-reproducible data reliance and face the challenge of compelling a single model for multiple functions. To this end, we introduce AutoAct, an automatic agent learning framework for QA that does not rely on large-scale annotated data and synthetic planning trajectories from closed-source models (e.g., GPT-4). Given limited data with a tool library, AutoAct first automatically synthesizes planning trajectories without any assistance from humans or strong closed-source models. Then, AutoAct leverages a division-of-labor strategy to automatically differentiate based on the target task information and synthesized trajectories, producing a sub-agent group to complete the task. We conduct comprehensive experiments with different LLMs, which demonstrates that AutoAct yields better or parallel performance compared to various strong baselines. Further analysis demonstrates the effectiveness of the division-of-labor strategy, with the trajectory quality generated by AutoAct generally outperforming that of others.", }

Language agents have achieved considerable performance on various complex question-answering tasks by planning with external tools. Despite the incessant exploration in this field, existing language agent systems still struggle with costly, non-reproducible data reliance and face the challenge of compelling a single model for multiple functions. To this end, we introduce AutoAct, an automatic agent learning framework for QA that does not rely on large-scale annotated data and synthetic planning trajectories from closed-source models (e.g., GPT-4). Given limited data with a tool library, AutoAct first automatically synthesizes planning trajectories without any assistance from humans or strong closed-source models. Then, AutoAct leverages a division-of-labor strategy to automatically differentiate based on the target task information and synthesized trajectories, producing a sub-agent group to complete the task. We conduct comprehensive experiments with different LLMs, which demonstrates that AutoAct yields better or parallel performance compared to various strong baselines. Further analysis demonstrates the effectiveness of the division-of-labor strategy, with the trajectory quality generated by AutoAct generally outperforming that of others.

AutoAct: Automatic Agent Learning from Scratch for QA via Self-Planning

acl-long.165

Poster

2401.05268

https://huggingface.co/papers/2401.05268

https://aclanthology.org/2024.acl-long.165/

https://aclanthology.org/2024.acl-long.166.bib

@inproceedings{t-y-s-s-etal-2024-chronoslex, title = "{C}hronos{L}ex: Time-aware Incremental Training for Temporal Generalization of Legal Classification Tasks", author = "T.y.s.s, Santosh and Vuong, Tuan-Quang and Grabmair, Matthias", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.166", pages = "3022--3039", abstract = "This study investigates the challenges posed by the dynamic nature of legal multi-label text classification tasks, where legal concepts evolve over time. Existing models often overlook the temporal dimension in their training process, leading to suboptimal performance of those models over time, as they treat training data as a single homogeneous block. To address this, we introduce ChronosLex, an incremental training paradigm that trains models on chronological splits, preserving the temporal order of the data. However, this incremental approach raises concerns about overfitting to recent data, prompting an assessment of mitigation strategies using continual learning and temporal invariant methods. Our experimental results over six legal multi-label text classification datasets reveal that continual learning methods prove effective in preventing overfitting thereby enhancing temporal generalizability, while temporal invariant methods struggle to capture these dynamics of temporal shifts.", }

This study investigates the challenges posed by the dynamic nature of legal multi-label text classification tasks, where legal concepts evolve over time. Existing models often overlook the temporal dimension in their training process, leading to suboptimal performance of those models over time, as they treat training data as a single homogeneous block. To address this, we introduce ChronosLex, an incremental training paradigm that trains models on chronological splits, preserving the temporal order of the data. However, this incremental approach raises concerns about overfitting to recent data, prompting an assessment of mitigation strategies using continual learning and temporal invariant methods. Our experimental results over six legal multi-label text classification datasets reveal that continual learning methods prove effective in preventing overfitting thereby enhancing temporal generalizability, while temporal invariant methods struggle to capture these dynamics of temporal shifts.

ChronosLex: Time-aware Incremental Training for Temporal Generalization of Legal Classification Tasks

acl-long.166

Poster

2405.14211

https://aclanthology.org/2024.acl-long.166/

https://aclanthology.org/2024.acl-long.167.bib

@inproceedings{gao-etal-2024-virtual, title = "Virtual Compiler Is All You Need For Assembly Code Search", author = "Gao, Zeyu and Wang, Hao and Wang, Yuanda and Zhang, Chao", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.167", pages = "3040--3051", abstract = "Assembly code search is vital for reducing the burden on reverse engineers, allowing them to quickly identify specific functions using natural language within vast binary programs.Despite its significance, this critical task is impeded by the complexities involved in building high-quality datasets. This paper explores training a Large Language Model (LLM) to emulate a general compiler. By leveraging Ubuntu packages to compile a dataset of 20 billion tokens, we further continue pre-train CodeLlama as a Virtual Compiler (ViC), capable of compiling any source code to assembly code. This approach allows for {``}virtual{''} compilation across a wide range of programming languages without the need for a real compiler, preserving semantic equivalency and expanding the possibilities for assembly code dataset construction. Furthermore, we use ViC to construct a sufficiently large dataset for assembly code search. Employing this extensive dataset, we achieve a substantial improvement in assembly code search performance, with our model surpassing the leading baseline by 26{\%}.", }

Assembly code search is vital for reducing the burden on reverse engineers, allowing them to quickly identify specific functions using natural language within vast binary programs.Despite its significance, this critical task is impeded by the complexities involved in building high-quality datasets. This paper explores training a Large Language Model (LLM) to emulate a general compiler. By leveraging Ubuntu packages to compile a dataset of 20 billion tokens, we further continue pre-train CodeLlama as a Virtual Compiler (ViC), capable of compiling any source code to assembly code. This approach allows for {``}virtual{''} compilation across a wide range of programming languages without the need for a real compiler, preserving semantic equivalency and expanding the possibilities for assembly code dataset construction. Furthermore, we use ViC to construct a sufficiently large dataset for assembly code search. Employing this extensive dataset, we achieve a substantial improvement in assembly code search performance, with our model surpassing the leading baseline by 26{\%}.

Virtual Compiler Is All You Need For Assembly Code Search

acl-long.167

Poster

https://aclanthology.org/2024.acl-long.167/

https://aclanthology.org/2024.acl-long.168.bib

@inproceedings{ren-etal-2024-melora, title = "{MEL}o{RA}: Mini-Ensemble Low-Rank Adapters for Parameter-Efficient Fine-Tuning", author = "Ren, Pengjie and Shi, Chengshun and Wu, Shiguang and Zhang, Mengqi and Ren, Zhaochun and Rijke, Maarten and Chen, Zhumin and Pei, Jiahuan", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.168", pages = "3052--3064", abstract = "Parameter-efficient fine-tuning (PEFT) is a popular method for tailoring pre-trained large language models (LLMs), especially as the models{'} scale and the diversity of tasks increase. Low-rank adaptation (LoRA) is based on the idea that the adaptation process is intrinsically low-dimensional, i.e., significant model changes can be represented with relatively few parameters. However, decreasing the rank encounters challenges with generalization errors for specific tasks when compared to full-parameter fine-tuning. We present MELoRA, a mini-ensemble low-rank adapters that uses fewer trainable parameters while maintaining a higher rank, thereby offering improved performance potential.The core idea is to freeze original pretrained weights and train a group of mini LoRAs with only a small number of parameters. This can capture a significant degree of diversity among mini LoRAs, thus promoting better generalization ability. We conduct a theoretical analysis and empirical studies on various NLP tasks. Our experimental results show that, compared to LoRA, MELoRA achieves better performance with 8 times fewer trainable parameters on natural language understanding tasks and 36 times fewer trainable parameters on instruction following tasks, which demonstrates the effectiveness of MELoRA.", }

Parameter-efficient fine-tuning (PEFT) is a popular method for tailoring pre-trained large language models (LLMs), especially as the models{'} scale and the diversity of tasks increase. Low-rank adaptation (LoRA) is based on the idea that the adaptation process is intrinsically low-dimensional, i.e., significant model changes can be represented with relatively few parameters. However, decreasing the rank encounters challenges with generalization errors for specific tasks when compared to full-parameter fine-tuning. We present MELoRA, a mini-ensemble low-rank adapters that uses fewer trainable parameters while maintaining a higher rank, thereby offering improved performance potential.The core idea is to freeze original pretrained weights and train a group of mini LoRAs with only a small number of parameters. This can capture a significant degree of diversity among mini LoRAs, thus promoting better generalization ability. We conduct a theoretical analysis and empirical studies on various NLP tasks. Our experimental results show that, compared to LoRA, MELoRA achieves better performance with 8 times fewer trainable parameters on natural language understanding tasks and 36 times fewer trainable parameters on instruction following tasks, which demonstrates the effectiveness of MELoRA.

MELoRA: Mini-Ensemble Low-Rank Adapters for Parameter-Efficient Fine-Tuning

acl-long.168

Poster

2402.17263

https://huggingface.co/papers/2402.17263

https://aclanthology.org/2024.acl-long.168/

https://aclanthology.org/2024.acl-long.169.bib

@inproceedings{tong-etal-2024-llms, title = "Can {LLM}s Learn from Previous Mistakes? Investigating {LLM}s{'} Errors to Boost for Reasoning", author = "Tong, Yongqi and Li, Dawei and Wang, Sizhe and Wang, Yujia and Teng, Fei and Shang, Jingbo", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.169", pages = "3065--3080", abstract = "Large language models (LLMs) have demonstrated striking reasoning capability. Recent works have shown the benefits to LLMs from fine-tuning golden-standard Chain-of-Thought (CoT) rationales or using them as correct examples in few-shot prompting. While humans can indeed imitate correct examples, learning from our mistakes is another vital aspect of human cognition. Hence, a question naturally arises: \textit{can LLMs learn and benefit from their mistakes, especially for their reasoning?}This study investigates this problem from both the prompting and model-tuning perspectives. We begin by introducing CoTErrorSet, a new benchmark with 609,432 questions, each designed with both correct and error references, and demonstrating the types and reasons for making such mistakes. To explore the effectiveness of those mistakes, we design two methods: (1) \textbf{Self-rethinking} prompting guides LLMs to rethink whether they have made similar previous mistakes; and (2) \textbf{Mistake tuning} involves finetuning models in both correct and incorrect reasoning domains, rather than only tuning models to learn ground truth in traditional methodology. We conduct a series of experiments to prove LLMs can obtain benefits from mistakes in both directions. Our two methods offer potentially cost-effective strategies by leveraging errors to enhance reasoning capabilities, which costs significantly less than creating meticulously hand-crafted golden references. We ultimately make a thorough analysis of the reasons behind LLMs{'} errors, which provides directions that future research needs to overcome. CoTErrorSet will be published soon on \url{https://github.com/YookiTong/Learn-from-Mistakes-CotErrorSet}.", }

Large language models (LLMs) have demonstrated striking reasoning capability. Recent works have shown the benefits to LLMs from fine-tuning golden-standard Chain-of-Thought (CoT) rationales or using them as correct examples in few-shot prompting. While humans can indeed imitate correct examples, learning from our mistakes is another vital aspect of human cognition. Hence, a question naturally arises: \textit{can LLMs learn and benefit from their mistakes, especially for their reasoning?}This study investigates this problem from both the prompting and model-tuning perspectives. We begin by introducing CoTErrorSet, a new benchmark with 609,432 questions, each designed with both correct and error references, and demonstrating the types and reasons for making such mistakes. To explore the effectiveness of those mistakes, we design two methods: (1) \textbf{Self-rethinking} prompting guides LLMs to rethink whether they have made similar previous mistakes; and (2) \textbf{Mistake tuning} involves finetuning models in both correct and incorrect reasoning domains, rather than only tuning models to learn ground truth in traditional methodology. We conduct a series of experiments to prove LLMs can obtain benefits from mistakes in both directions. Our two methods offer potentially cost-effective strategies by leveraging errors to enhance reasoning capabilities, which costs significantly less than creating meticulously hand-crafted golden references. We ultimately make a thorough analysis of the reasons behind LLMs{'} errors, which provides directions that future research needs to overcome. CoTErrorSet will be published soon on \url{https://github.com/YookiTong/Learn-from-Mistakes-CotErrorSet}.

Can LLMs Learn from Previous Mistakes? Investigating LLMs' Errors to Boost for Reasoning

acl-long.169

Poster

2403.20046

https://aclanthology.org/2024.acl-long.169/

https://aclanthology.org/2024.acl-long.170.bib

@inproceedings{yang-etal-2024-iterative, title = "An Iterative Associative Memory Model for Empathetic Response Generation", author = "Yang, Zhou and Ren, Zhaochun and Yufeng, Wang and Sun, Haizhou and Chen, Chao and Zhu, Xiaofei and Liao, Xiangwen", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.170", pages = "3081--3092", abstract = "Empathetic response generation aims to comprehend the cognitive and emotional states in dialogue utterances and generate proper responses. Psychological theories posit that comprehending emotional and cognitive states necessitates iteratively capturing and understanding associated words across dialogue utterances. However, existing approaches regard dialogue utterances as either a long sequence or independent utterances for comprehension, which are prone to overlook the associated words between them. To address this issue, we propose an Iterative Associative Memory Model (IAMM) for empathetic response generation. Specifically, we employ a novel second-order interaction attention mechanism to iteratively capture vital associated words between dialogue utterances and situations, dialogue history, and a memory module (for storing associated words), thereby accurately and nuancedly comprehending the utterances.We conduct experiments on the Empathetic-Dialogue dataset. Both automatic and human evaluations validate the efficacy of the model. Variant experiments on LLMs also demonstrate that attending to associated words improves empathetic comprehension and expression.", }

Empathetic response generation aims to comprehend the cognitive and emotional states in dialogue utterances and generate proper responses. Psychological theories posit that comprehending emotional and cognitive states necessitates iteratively capturing and understanding associated words across dialogue utterances. However, existing approaches regard dialogue utterances as either a long sequence or independent utterances for comprehension, which are prone to overlook the associated words between them. To address this issue, we propose an Iterative Associative Memory Model (IAMM) for empathetic response generation. Specifically, we employ a novel second-order interaction attention mechanism to iteratively capture vital associated words between dialogue utterances and situations, dialogue history, and a memory module (for storing associated words), thereby accurately and nuancedly comprehending the utterances.We conduct experiments on the Empathetic-Dialogue dataset. Both automatic and human evaluations validate the efficacy of the model. Variant experiments on LLMs also demonstrate that attending to associated words improves empathetic comprehension and expression.

An Iterative Associative Memory Model for Empathetic Response Generation

acl-long.170

Oral

2402.17959

https://aclanthology.org/2024.acl-long.170/

https://aclanthology.org/2024.acl-long.171.bib

@inproceedings{wang-etal-2024-detoxifying, title = "Detoxifying Large Language Models via Knowledge Editing", author = "Wang, Mengru and Zhang, Ningyu and Xu, Ziwen and Xi, Zekun and Deng, Shumin and Yao, Yunzhi and Zhang, Qishen and Yang, Linyi and Wang, Jindong and Chen, Huajun", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.171", pages = "3093--3118", abstract = "This paper investigates using knowledge editing techniques to detoxify Large Language Models (LLMs). We construct a benchmark, SafeEdit, which covers nine unsafe categories with various powerful attack prompts and equips comprehensive metrics for systematic evaluation. We conduct experiments with several knowledge editing approaches, indicating that knowledge editing has the potential to efficiently detoxify LLMs with limited impact on general performance. Then, we propose a simple yet effective baseline, dubbed Detoxifying with Intraoperative Neural Monitoring (DINM), to diminish the toxicity of LLMs within a few tuning steps via only one instance. We further provide an in-depth analysis of the internal mechanism for various detoxifying approaches, demonstrating that previous methods like SFT and DPO may merely suppress the activations of toxic parameters, while DINM mitigates the toxicity of the toxic parameters to a certain extent, making permanent adjustments. We hope that these insights could shed light on future work of developing detoxifying approaches and the underlying knowledge mechanisms of LLMs.", }

This paper investigates using knowledge editing techniques to detoxify Large Language Models (LLMs). We construct a benchmark, SafeEdit, which covers nine unsafe categories with various powerful attack prompts and equips comprehensive metrics for systematic evaluation. We conduct experiments with several knowledge editing approaches, indicating that knowledge editing has the potential to efficiently detoxify LLMs with limited impact on general performance. Then, we propose a simple yet effective baseline, dubbed Detoxifying with Intraoperative Neural Monitoring (DINM), to diminish the toxicity of LLMs within a few tuning steps via only one instance. We further provide an in-depth analysis of the internal mechanism for various detoxifying approaches, demonstrating that previous methods like SFT and DPO may merely suppress the activations of toxic parameters, while DINM mitigates the toxicity of the toxic parameters to a certain extent, making permanent adjustments. We hope that these insights could shed light on future work of developing detoxifying approaches and the underlying knowledge mechanisms of LLMs.

Detoxifying Large Language Models via Knowledge Editing

acl-long.171

Poster

2403.14472

https://huggingface.co/papers/2403.14472

https://aclanthology.org/2024.acl-long.171/

https://aclanthology.org/2024.acl-long.172.bib

@inproceedings{bai-etal-2024-longbench, title = "{L}ong{B}ench: A Bilingual, Multitask Benchmark for Long Context Understanding", author = "Bai, Yushi and Lv, Xin and Zhang, Jiajie and Lyu, Hongchang and Tang, Jiankai and Huang, Zhidian and Du, Zhengxiao and Liu, Xiao and Zeng, Aohan and Hou, Lei and Dong, Yuxiao and Tang, Jie and Li, Juanzi", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.172", pages = "3119--3137", abstract = "Although large language models (LLMs) demonstrate impressive performance for many language tasks, most of them can only handle texts a few thousand tokens long, limiting their applications on longer sequence inputs, such as books, reports, and codebases. Recent works have proposed methods to improve LLMs{'} long context capabilities by extending context windows and more sophisticated memory mechanisms. However, comprehensive benchmarks tailored for evaluating long context understanding are lacking. In this paper, we introduce LongBench, the first bilingual, multi-task benchmark for long context understanding, enabling a more rigorous evaluation of long context understanding. LongBench comprises 21 datasets across 6 task categories in both English and Chinese, with an average length of 6,711 words (English) and 13,386 characters (Chinese). These tasks cover key long-text application areas including single-doc QA, multi-doc QA, summarization, few-shot learning, synthetic tasks, and code completion. All datasets in LongBench are standardized into a unified format, allowing for effortless automatic evaluation of LLMs. Upon comprehensive evaluation of 8 LLMs on LongBench, we find that: (1) Commercial model (GPT-3.5-Turbo-16k) outperforms other open-sourced models, but still struggles on longer contexts. (2) Scaled position embedding and fine-tuning on longer sequences lead to substantial improvement on long context understanding. (3) Context compression technique such as retrieval brings improvement for model with weak ability on long contexts, but the performance still lags behind models that have strong long context understanding capability.", }

Although large language models (LLMs) demonstrate impressive performance for many language tasks, most of them can only handle texts a few thousand tokens long, limiting their applications on longer sequence inputs, such as books, reports, and codebases. Recent works have proposed methods to improve LLMs{'} long context capabilities by extending context windows and more sophisticated memory mechanisms. However, comprehensive benchmarks tailored for evaluating long context understanding are lacking. In this paper, we introduce LongBench, the first bilingual, multi-task benchmark for long context understanding, enabling a more rigorous evaluation of long context understanding. LongBench comprises 21 datasets across 6 task categories in both English and Chinese, with an average length of 6,711 words (English) and 13,386 characters (Chinese). These tasks cover key long-text application areas including single-doc QA, multi-doc QA, summarization, few-shot learning, synthetic tasks, and code completion. All datasets in LongBench are standardized into a unified format, allowing for effortless automatic evaluation of LLMs. Upon comprehensive evaluation of 8 LLMs on LongBench, we find that: (1) Commercial model (GPT-3.5-Turbo-16k) outperforms other open-sourced models, but still struggles on longer contexts. (2) Scaled position embedding and fine-tuning on longer sequences lead to substantial improvement on long context understanding. (3) Context compression technique such as retrieval brings improvement for model with weak ability on long contexts, but the performance still lags behind models that have strong long context understanding capability.

LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding

acl-long.172

Poster

2308.14508

https://huggingface.co/papers/2308.14508

https://aclanthology.org/2024.acl-long.172/

https://aclanthology.org/2024.acl-long.173.bib

@inproceedings{chen-etal-2024-dr, title = "{D}r.{A}cademy: A Benchmark for Evaluating Questioning Capability in Education for Large Language Models", author = "Chen, Yuyan and Yan, Songzhou and Liu, Panjun and Xiao, Yanghua", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.173", pages = "3138--3167", abstract = "Teachers are important to imparting knowledge and guiding learners, and the role of large language models (LLMs) as potential educators is emerging as an important area of study. Recognizing LLMs{'} capability to generate educational content can lead to advances in automated and personalized learning. While LLMs have been tested for their comprehension and problem-solving skills, their capability in teaching remains largely unexplored.In teaching, questioning is a key skill that guides students to analyze, evaluate, and synthesize core concepts and principles.Therefore, our research introduces a benchmark to evaluate the questioning capability in education as a teacher of LLMs through evaluating their generated educational questions, utilizing Anderson and Krathwohl{'}s taxonomy across general, monodisciplinary, and interdisciplinary domains. We shift the focus from LLMs as learners to LLMs as educators, assessing their teaching capability through guiding them to generate questions. We apply four metrics, including relevance, coverage, representativeness, and consistency, to evaluate the educational quality of LLMs{'} outputs. Our results indicate that GPT-4 demonstrates significant potential in teaching general, humanities, and science courses; Claude2 appears more apt as an interdisciplinary teacher. Furthermore, the automatic scores align with human perspectives.", }

Teachers are important to imparting knowledge and guiding learners, and the role of large language models (LLMs) as potential educators is emerging as an important area of study. Recognizing LLMs{'} capability to generate educational content can lead to advances in automated and personalized learning. While LLMs have been tested for their comprehension and problem-solving skills, their capability in teaching remains largely unexplored.In teaching, questioning is a key skill that guides students to analyze, evaluate, and synthesize core concepts and principles.Therefore, our research introduces a benchmark to evaluate the questioning capability in education as a teacher of LLMs through evaluating their generated educational questions, utilizing Anderson and Krathwohl{'}s taxonomy across general, monodisciplinary, and interdisciplinary domains. We shift the focus from LLMs as learners to LLMs as educators, assessing their teaching capability through guiding them to generate questions. We apply four metrics, including relevance, coverage, representativeness, and consistency, to evaluate the educational quality of LLMs{'} outputs. Our results indicate that GPT-4 demonstrates significant potential in teaching general, humanities, and science courses; Claude2 appears more apt as an interdisciplinary teacher. Furthermore, the automatic scores align with human perspectives.

Dr.Academy: A Benchmark for Evaluating Questioning Capability in Education for Large Language Models

acl-long.173

Poster

2408.10947

https://aclanthology.org/2024.acl-long.173/

https://aclanthology.org/2024.acl-long.174.bib

@inproceedings{pham-etal-2024-unibridge, title = "{U}ni{B}ridge: A Unified Approach to Cross-Lingual Transfer Learning for Low-Resource Languages", author = "Pham, Trinh and Le, Khoi and Luu, Anh Tuan", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.174", pages = "3168--3184", abstract = "In this paper, we introduce UniBridge (Cross-Lingual Transfer Learning with Optimized Embeddings and Vocabulary), a comprehensive approach developed to improve the effectiveness of Cross-Lingual Transfer Learning, particularly in languages with limited resources. Our approach tackles two essential elements of a language model: the initialization of embeddings and the optimal vocabulary size. Specifically, we propose a novel embedding initialization method that leverages both lexical and semantic alignment for a language. In addition, we present a method for systematically searching for the optimal vocabulary size, ensuring a balance between model complexity and linguistic coverage. Our experiments across multilingual datasets show that our approach greatly improves the F1-Score in several languages. UniBridge is a robust and adaptable solution for cross-lingual systems in various languages, highlighting the significance of initializing embeddings and choosing the right vocabulary size in cross-lingual environments.", }

In this paper, we introduce UniBridge (Cross-Lingual Transfer Learning with Optimized Embeddings and Vocabulary), a comprehensive approach developed to improve the effectiveness of Cross-Lingual Transfer Learning, particularly in languages with limited resources. Our approach tackles two essential elements of a language model: the initialization of embeddings and the optimal vocabulary size. Specifically, we propose a novel embedding initialization method that leverages both lexical and semantic alignment for a language. In addition, we present a method for systematically searching for the optimal vocabulary size, ensuring a balance between model complexity and linguistic coverage. Our experiments across multilingual datasets show that our approach greatly improves the F1-Score in several languages. UniBridge is a robust and adaptable solution for cross-lingual systems in various languages, highlighting the significance of initializing embeddings and choosing the right vocabulary size in cross-lingual environments.

UniBridge: A Unified Approach to Cross-Lingual Transfer Learning for Low-Resource Languages

acl-long.174

Poster

2406.09717

https://huggingface.co/papers/2406.09717

https://aclanthology.org/2024.acl-long.174/

https://aclanthology.org/2024.acl-long.175.bib

@inproceedings{zhou-etal-2024-vista, title = "{VISTA}: Visualized Text Embedding For Universal Multi-Modal Retrieval", author = "Zhou, Junjie and Liu, Zheng and Xiao, Shitao and Zhao, Bo and Xiong, Yongping", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.175", pages = "3185--3200", abstract = "Multi-modal retrieval becomes increasingly popular in practice. However, the existing retrievers are mostly text-oriented, which lack the capability to process visual information. Despite the presence of vision-language models like CLIP, the current methods are severely limited in representing the text-only and image-only data. In this work, we present a new embedding model VISTA for universal multi-modal retrieval. Our work brings forth threefold technical contributions. Firstly, we introduce a flexible architecture which extends a powerful text encoder with the image understanding capability by introducing visual token embeddings. Secondly, we develop two data generation strategies, which bring high-quality composed image-text to facilitate the training of the embedding model. Thirdly, we introduce a multi-stage training algorithm, which first aligns the visual token embedding with the text encoder using massive weakly labeled data, and then develops multi-modal representation capability using the generated composed image-text data. In our experiments, VISTA achieves superior performances across a variety of multi-modal retrieval tasks in both zero-shot and supervised settings. Our model, data, and source code are available at https://github.com/FlagOpen/FlagEmbedding.", }

Multi-modal retrieval becomes increasingly popular in practice. However, the existing retrievers are mostly text-oriented, which lack the capability to process visual information. Despite the presence of vision-language models like CLIP, the current methods are severely limited in representing the text-only and image-only data. In this work, we present a new embedding model VISTA for universal multi-modal retrieval. Our work brings forth threefold technical contributions. Firstly, we introduce a flexible architecture which extends a powerful text encoder with the image understanding capability by introducing visual token embeddings. Secondly, we develop two data generation strategies, which bring high-quality composed image-text to facilitate the training of the embedding model. Thirdly, we introduce a multi-stage training algorithm, which first aligns the visual token embedding with the text encoder using massive weakly labeled data, and then develops multi-modal representation capability using the generated composed image-text data. In our experiments, VISTA achieves superior performances across a variety of multi-modal retrieval tasks in both zero-shot and supervised settings. Our model, data, and source code are available at https://github.com/FlagOpen/FlagEmbedding.

VISTA: Visualized Text Embedding For Universal Multi-Modal Retrieval

acl-long.175

Poster

2406.04292

https://huggingface.co/papers/2406.04292

https://aclanthology.org/2024.acl-long.175/

https://aclanthology.org/2024.acl-long.176.bib

@inproceedings{cheng-etal-2024-black, title = "Black-Box Prompt Optimization: Aligning Large Language Models without Model Training", author = "Cheng, Jiale and Liu, Xiao and Zheng, Kehan and Ke, Pei and Wang, Hongning and Dong, Yuxiao and Tang, Jie and Huang, Minlie", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.176", pages = "3201--3219", abstract = "Large language models (LLMs) have shown impressive success in various applications. However, these models are often not well aligned with human intents, which calls for additional treatments on them; that is, the alignment problem. To make LLMs better follow user instructions, existing alignment methods primarily focus on further training them. However, the extra training of LLMs is usually expensive in terms of GPU computing; even worse, some LLMs are not accessible for user-demanded training, such as GPTs. In this work, we take a different perspective{---}Black-Box Prompt Optimization (BPO){---}to perform alignments. The idea is to optimize user prompts to suit LLMs{'} input understanding, so as to best realize users{'} intents without updating LLMs{'} parameters. BPO leverages human preferences to optimize prompts, thus making it superior to LLM (e.g., ChatGPT) as a prompt engineer. Moreover, BPO is model-agnostic, and the empirical results demonstrate that the BPO-aligned ChatGPT yields a 22{\%} increase in the win rate against its original version and 10{\%} for GPT-4. Notably, the BPO-aligned LLMs can outperform the same models aligned by PPO and DPO, and it also brings additional performance gains when combining BPO with PPO or DPO. Code and datasets are released at https://github.com/thu-coai/BPO.", }

Large language models (LLMs) have shown impressive success in various applications. However, these models are often not well aligned with human intents, which calls for additional treatments on them; that is, the alignment problem. To make LLMs better follow user instructions, existing alignment methods primarily focus on further training them. However, the extra training of LLMs is usually expensive in terms of GPU computing; even worse, some LLMs are not accessible for user-demanded training, such as GPTs. In this work, we take a different perspective{---}Black-Box Prompt Optimization (BPO){---}to perform alignments. The idea is to optimize user prompts to suit LLMs{'} input understanding, so as to best realize users{'} intents without updating LLMs{'} parameters. BPO leverages human preferences to optimize prompts, thus making it superior to LLM (e.g., ChatGPT) as a prompt engineer. Moreover, BPO is model-agnostic, and the empirical results demonstrate that the BPO-aligned ChatGPT yields a 22{\%} increase in the win rate against its original version and 10{\%} for GPT-4. Notably, the BPO-aligned LLMs can outperform the same models aligned by PPO and DPO, and it also brings additional performance gains when combining BPO with PPO or DPO. Code and datasets are released at https://github.com/thu-coai/BPO.

Black-Box Prompt Optimization: Aligning Large Language Models without Model Training

acl-long.176

Poster

2311.04155

https://huggingface.co/papers/2311.04155

https://aclanthology.org/2024.acl-long.176/

https://aclanthology.org/2024.acl-long.177.bib

@inproceedings{park-etal-2024-open, title = "Open {K}o-{LLM} Leaderboard: Evaluating Large Language Models in {K}orean with {K}o-H5 Benchmark", author = "Park, Chanjun and Kim, Hyeonwoo and Kim, Dahyun and Cho, SeongHwan and Kim, Sanghoon and Lee, Sukyung and Kim, Yungi and Lee, Hwalsuk", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.177", pages = "3220--3234", abstract = "This paper introduces the Open Ko-LLM Leaderboard and the Ko-H5 Benchmark as vital tools for evaluating Large Language Models (LLMs) in Korean. Incorporating private test sets while mirroring the English Open LLM Leaderboard, we establish a robust evaluation framework that has been well integrated in the Korean LLM community. We perform data leakage analysis that shows the benefit of private test sets along with a correlation study within the Ko-H5 benchmark and temporal analyses of the Ko-H5 score. Moreover, we present empirical support for the need to expand beyond set benchmarks. We hope the Open Ko-LLM Leaderboard sets precedent for expanding LLM evaluation to foster more linguistic diversity.", }

This paper introduces the Open Ko-LLM Leaderboard and the Ko-H5 Benchmark as vital tools for evaluating Large Language Models (LLMs) in Korean. Incorporating private test sets while mirroring the English Open LLM Leaderboard, we establish a robust evaluation framework that has been well integrated in the Korean LLM community. We perform data leakage analysis that shows the benefit of private test sets along with a correlation study within the Ko-H5 benchmark and temporal analyses of the Ko-H5 score. Moreover, we present empirical support for the need to expand beyond set benchmarks. We hope the Open Ko-LLM Leaderboard sets precedent for expanding LLM evaluation to foster more linguistic diversity.

Open Ko-LLM Leaderboard: Evaluating Large Language Models in Korean with Ko-H5 Benchmark

acl-long.177

Poster

2405.20574

https://aclanthology.org/2024.acl-long.177/

https://aclanthology.org/2024.acl-long.178.bib

@inproceedings{chen-etal-2024-unified-hallucination, title = "Unified Hallucination Detection for Multimodal Large Language Models", author = "Chen, Xiang and Wang, Chenxi and Xue, Yida and Zhang, Ningyu and Yang, Xiaoyan and Li, Qiang and Shen, Yue and Liang, Lei and Gu, Jinjie and Chen, Huajun", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.178", pages = "3235--3252", abstract = "Despite significant strides in multimodal tasks, Multimodal Large Language Models (MLLMs) are plagued by the critical issue of hallucination. The reliable detection of such hallucinations in MLLMs has, therefore, become a vital aspect of model evaluation and the safeguarding of practical application deployment. Prior research in this domain has been constrained by a narrow focus on singular tasks, an inadequate range of hallucination categories addressed, and a lack of detailed granularity. In response to these challenges, our work expands the investigative horizons of hallucination detection. We present a novel meta-evaluation benchmark, MHaluBench, meticulously crafted to facilitate the evaluation of advancements in hallucination detection methods. Additionally, we unveil a novel unified multimodal hallucination detection framework, UNIHD, which leverages a suite of auxiliary tools to validate the occurrence of hallucinations robustly. We demonstrate the effectiveness of UNIHD through meticulous evaluation and comprehensive analysis. We also provide strategic insights on the application of specific tools for addressing various categories of hallucinations.", }

Despite significant strides in multimodal tasks, Multimodal Large Language Models (MLLMs) are plagued by the critical issue of hallucination. The reliable detection of such hallucinations in MLLMs has, therefore, become a vital aspect of model evaluation and the safeguarding of practical application deployment. Prior research in this domain has been constrained by a narrow focus on singular tasks, an inadequate range of hallucination categories addressed, and a lack of detailed granularity. In response to these challenges, our work expands the investigative horizons of hallucination detection. We present a novel meta-evaluation benchmark, MHaluBench, meticulously crafted to facilitate the evaluation of advancements in hallucination detection methods. Additionally, we unveil a novel unified multimodal hallucination detection framework, UNIHD, which leverages a suite of auxiliary tools to validate the occurrence of hallucinations robustly. We demonstrate the effectiveness of UNIHD through meticulous evaluation and comprehensive analysis. We also provide strategic insights on the application of specific tools for addressing various categories of hallucinations.

Unified Hallucination Detection for Multimodal Large Language Models

acl-long.178

Poster

2402.03190

https://huggingface.co/papers/2402.03190

https://aclanthology.org/2024.acl-long.178/

https://aclanthology.org/2024.acl-long.179.bib

@inproceedings{ren-etal-2024-empowering, title = "Empowering Character-level Text Infilling by Eliminating Sub-Tokens", author = "Ren, Houxing and Zhan, Mingjie and Wu, Zhongyuan and Li, Hongsheng", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.179", pages = "3253--3267", abstract = "In infilling tasks, sub-tokens, representing instances where a complete token is segmented into two parts, often emerge at the boundaries of prefixes, middles, and suffixes. Traditional methods focused on training models at the token level, leading to sub-optimal performance in character-level infilling tasks during the inference stage. Alternately, some approaches considered character-level infilling, but they relied on predicting sub-tokens in inference, yet this strategy diminished ability in character-level infilling tasks due to the large perplexity of the model on sub-tokens. In this paper, we introduce FIM-SE, which stands for Fill-In-the-Middle with both Starting and Ending character constraints. The proposed method addresses character-level infilling tasks by utilizing a line-level format to avoid predicting any sub-token in inference. In addition, we incorporate two special tokens to signify the rest of the incomplete lines, thereby enhancing generation guidance. Extensive experiments demonstrate that our proposed approach surpasses previous methods, offering a significant advantage. Code is available at https://github.com/SenseLLM/FIM-SE.", }

In infilling tasks, sub-tokens, representing instances where a complete token is segmented into two parts, often emerge at the boundaries of prefixes, middles, and suffixes. Traditional methods focused on training models at the token level, leading to sub-optimal performance in character-level infilling tasks during the inference stage. Alternately, some approaches considered character-level infilling, but they relied on predicting sub-tokens in inference, yet this strategy diminished ability in character-level infilling tasks due to the large perplexity of the model on sub-tokens. In this paper, we introduce FIM-SE, which stands for Fill-In-the-Middle with both Starting and Ending character constraints. The proposed method addresses character-level infilling tasks by utilizing a line-level format to avoid predicting any sub-token in inference. In addition, we incorporate two special tokens to signify the rest of the incomplete lines, thereby enhancing generation guidance. Extensive experiments demonstrate that our proposed approach surpasses previous methods, offering a significant advantage. Code is available at https://github.com/SenseLLM/FIM-SE.

Empowering Character-level Text Infilling by Eliminating Sub-Tokens

acl-long.179

Poster

2405.17103

https://huggingface.co/papers/2405.17103

https://aclanthology.org/2024.acl-long.179/

https://aclanthology.org/2024.acl-long.180.bib

@inproceedings{luo-etal-2024-landmark, title = "Landmark Embedding: A Chunking-Free Embedding Method For Retrieval Augmented Long-Context Large Language Models", author = "Luo, Kun and Liu, Zheng and Xiao, Shitao and Zhou, Tong and Chen, Yubo and Zhao, Jun and Liu, Kang", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.180", pages = "3268--3281", abstract = "Retrieval augmentation is a promising approach to handle long-context language modeling. However, the existing retrieval methods usually work with the chunked context, which is prone to inferior quality of semantic representation and incomplete retrieval of useful information. In this work, we propose a new method for the retrieval augmentation of long-context language modeling, called Landmark Embedding. Our method is characterized by threefold technical contributions. Firstly, we introduce a \textit{chunking-free architecture}, which keeps the long context coherent such that high-quality embeddings can be generated for the fine-grained units within the context. Secondly, we present a position-aware objective function, which prioritizes the ultimate boundary for a consecutive span of information. By learning to discriminate such a special position, the useful information can be comprehensively retrieved for the query. Thirdly, we design a novel multi-stage learning algorithm, which makes the best use of readily available data and synthetic data for cost-effective training of the landmark embedding. In our experimental study, landmark embedding is able to substantially improve the performance for both LLaMA-2 and ChatGPT in a variety of long-context tasks; meanwhile, it also outperforms the existing retrieval methods with a notable advantage. Our model and source code will be made publicly available.", }

Retrieval augmentation is a promising approach to handle long-context language modeling. However, the existing retrieval methods usually work with the chunked context, which is prone to inferior quality of semantic representation and incomplete retrieval of useful information. In this work, we propose a new method for the retrieval augmentation of long-context language modeling, called Landmark Embedding. Our method is characterized by threefold technical contributions. Firstly, we introduce a \textit{chunking-free architecture}, which keeps the long context coherent such that high-quality embeddings can be generated for the fine-grained units within the context. Secondly, we present a position-aware objective function, which prioritizes the ultimate boundary for a consecutive span of information. By learning to discriminate such a special position, the useful information can be comprehensively retrieved for the query. Thirdly, we design a novel multi-stage learning algorithm, which makes the best use of readily available data and synthetic data for cost-effective training of the landmark embedding. In our experimental study, landmark embedding is able to substantially improve the performance for both LLaMA-2 and ChatGPT in a variety of long-context tasks; meanwhile, it also outperforms the existing retrieval methods with a notable advantage. Our model and source code will be made publicly available.

Landmark Embedding: A Chunking-Free Embedding Method For Retrieval Augmented Long-Context Large Language Models

acl-long.180

Oral

https://huggingface.co/papers/2402.11573

https://aclanthology.org/2024.acl-long.180/

https://aclanthology.org/2024.acl-long.181.bib

@inproceedings{ko-etal-2024-growover, title = "{G}row{OVER}: How Can {LLM}s Adapt to Growing Real-World Knowledge?", author = "Ko, Dayoon and Kim, Jinyoung and Choi, Hahyeon and Kim, Gunhee", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.181", pages = "3282--3308", abstract = "In the real world, knowledge is constantly evolving, which can render existing knowledge-based datasets outdated. This unreliability highlights the critical need for continuous updates to ensure both accuracy and relevance in knowledge-intensive tasks. To address this, we propose GrowOVER-QA and GrowOVER-Dialogue, dynamic open-domain QA and dialogue benchmarks that undergo a continuous cycle of updates, keeping pace with the rapid evolution of knowledge. Our research indicates that retrieval-augmented language models (RaLMs) struggle with knowledge that has not been trained on or recently updated. Consequently, we introduce a novel retrieval-interactive language model framework, where the language model evaluates and reflects on its answers for further re-retrieval. Our exhaustive experiments demonstrate that our training-free framework significantly improves upon existing methods, performing comparably to or even surpassing continuously trained language models.", }

In the real world, knowledge is constantly evolving, which can render existing knowledge-based datasets outdated. This unreliability highlights the critical need for continuous updates to ensure both accuracy and relevance in knowledge-intensive tasks. To address this, we propose GrowOVER-QA and GrowOVER-Dialogue, dynamic open-domain QA and dialogue benchmarks that undergo a continuous cycle of updates, keeping pace with the rapid evolution of knowledge. Our research indicates that retrieval-augmented language models (RaLMs) struggle with knowledge that has not been trained on or recently updated. Consequently, we introduce a novel retrieval-interactive language model framework, where the language model evaluates and reflects on its answers for further re-retrieval. Our exhaustive experiments demonstrate that our training-free framework significantly improves upon existing methods, performing comparably to or even surpassing continuously trained language models.

GrowOVER: How Can LLMs Adapt to Growing Real-World Knowledge?

acl-long.181

Poster

2406.05606

https://aclanthology.org/2024.acl-long.181/

https://aclanthology.org/2024.acl-long.182.bib

@inproceedings{slobodkin-etal-2024-attribute, title = "Attribute First, then Generate: Locally-attributable Grounded Text Generation", author = "Slobodkin, Aviv and Hirsch, Eran and Cattan, Arie and Schuster, Tal and Dagan, Ido", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.182", pages = "3309--3344", abstract = "Recent efforts to address hallucinations in Large Language Models (LLMs) have focused on attributed text generation, which supplements generated texts with citations of supporting sources for post-generation fact-checking and corrections. Yet, these citations often point to entire documents or paragraphs, burdening users with extensive verification work. In this paper, we introduce a locally-attributable text generation approach, prioritizing concise attributions. Our method, named {``}Attribute First, then Generate{``}, breaks down the conventional end-to-end generation process into three intuitive steps: content selection, sentence planning, and sequential sentence generation. By initially identifying relevant source segments ({``}select first{``}) and then conditioning the generation process on them ({``}then generate{``}), we ensure these segments also act as the output{'}s fine-grained attributions ({``}select{``} becomes {``}attribute{``}). Tested on Multi-document Summarization and Long-form Question-answering, our method not only yields more concise citations than the baselines but also maintains - and in some cases enhances - both generation quality and attribution accuracy. Furthermore, it significantly reduces the time required for fact verification by human assessors.", }

Recent efforts to address hallucinations in Large Language Models (LLMs) have focused on attributed text generation, which supplements generated texts with citations of supporting sources for post-generation fact-checking and corrections. Yet, these citations often point to entire documents or paragraphs, burdening users with extensive verification work. In this paper, we introduce a locally-attributable text generation approach, prioritizing concise attributions. Our method, named {``}Attribute First, then Generate{``}, breaks down the conventional end-to-end generation process into three intuitive steps: content selection, sentence planning, and sequential sentence generation. By initially identifying relevant source segments ({``}select first{``}) and then conditioning the generation process on them ({``}then generate{``}), we ensure these segments also act as the output{'}s fine-grained attributions ({``}select{``} becomes {``}attribute{``}). Tested on Multi-document Summarization and Long-form Question-answering, our method not only yields more concise citations than the baselines but also maintains - and in some cases enhances - both generation quality and attribution accuracy. Furthermore, it significantly reduces the time required for fact verification by human assessors.

Attribute First, then Generate: Locally-attributable Grounded Text Generation

acl-long.182

Poster

2403.17104

https://aclanthology.org/2024.acl-long.182/

https://aclanthology.org/2024.acl-long.183.bib

@inproceedings{yin-etal-2024-t2s, title = "{T}2{S}-{GPT}: Dynamic Vector Quantization for Autoregressive Sign Language Production from Text", author = "Yin, Aoxiong and Li, Haoyuan and Shen, Kai and Tang, Siliang and Zhuang, Yueting", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.183", pages = "3345--3356", abstract = "In this work, we propose a two-stage sign language production (SLP) paradigm that first encodes sign language sequences into discrete codes and then autoregressively generates sign language from text based on the learned codebook. However, existing vector quantization (VQ) methods are fixed-length encodings, overlooking the uneven information density in sign language, which leads to under-encoding of important regions and over-encoding of unimportant regions. To address this issue, we propose a novel dynamic vector quantization (DVA-VAE) model that can dynamically adjust the encoding length based on the information density in sign language to achieve accurate and compact encoding. Then, a GPT-like model learns to generate code sequences and their corresponding durations from spoken language text. Extensive experiments conducted on the PHOENIX14T dataset demonstrate the effectiveness of our proposed method. To promote sign language research, we propose a new large German sign language dataset, PHOENIX-News, which contains 486 hours of sign language videos, audio, and transcription texts. Experimental analysis on PHOENIX-News shows that the performance of our model can be further improved by increasing the size of the training data. Our project homepage is https://t2sgpt-demo.yinaoxiong.cn.", }

In this work, we propose a two-stage sign language production (SLP) paradigm that first encodes sign language sequences into discrete codes and then autoregressively generates sign language from text based on the learned codebook. However, existing vector quantization (VQ) methods are fixed-length encodings, overlooking the uneven information density in sign language, which leads to under-encoding of important regions and over-encoding of unimportant regions. To address this issue, we propose a novel dynamic vector quantization (DVA-VAE) model that can dynamically adjust the encoding length based on the information density in sign language to achieve accurate and compact encoding. Then, a GPT-like model learns to generate code sequences and their corresponding durations from spoken language text. Extensive experiments conducted on the PHOENIX14T dataset demonstrate the effectiveness of our proposed method. To promote sign language research, we propose a new large German sign language dataset, PHOENIX-News, which contains 486 hours of sign language videos, audio, and transcription texts. Experimental analysis on PHOENIX-News shows that the performance of our model can be further improved by increasing the size of the training data. Our project homepage is https://t2sgpt-demo.yinaoxiong.cn.

T2S-GPT: Dynamic Vector Quantization for Autoregressive Sign Language Production from Text

acl-long.183

Poster

2406.07119

https://aclanthology.org/2024.acl-long.183/

https://aclanthology.org/2024.acl-long.184.bib

@inproceedings{bi-etal-2024-oceangpt, title = "{O}cean{GPT}: A Large Language Model for Ocean Science Tasks", author = "Bi, Zhen and Zhang, Ningyu and Xue, Yida and Ou, Yixin and Ji, Daxiong and Zheng, Guozhou and Chen, Huajun", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.184", pages = "3357--3372", abstract = "Ocean science, which delves into the oceans that are reservoirs of life and biodiversity, is of great significance given that oceans cover over 70{\%} of our planet{'}s surface. Recently, advances in Large Language Models (LLMs) have transformed the paradigm in science. Despite the success in other domains, current LLMs often fall short in catering to the needs of domain experts like oceanographers, and the potential of LLMs for ocean science is under-explored. The intrinsic reason may be the immense and intricate nature of ocean data as well as the necessity for higher granularity and richness in knowledge. To alleviate these issues, we introduce OceanGPT, the first-ever LLM in the ocean domain, which is expert in various ocean science tasks. We propose DoInstruct, a novel framework to automatically obtain a large volume of ocean domain instruction data, which generates instructions based on multi-agent collaboration. Additionally, we construct the first oceanography benchmark, OceanBench, to evaluate the capabilities of LLMs in the ocean domain. Though comprehensive experiments, OceanGPT not only shows a higher level of knowledge expertise for oceans science tasks but also gains preliminary embodied intelligence capabilities in ocean technology.", }

Ocean science, which delves into the oceans that are reservoirs of life and biodiversity, is of great significance given that oceans cover over 70{\%} of our planet{'}s surface. Recently, advances in Large Language Models (LLMs) have transformed the paradigm in science. Despite the success in other domains, current LLMs often fall short in catering to the needs of domain experts like oceanographers, and the potential of LLMs for ocean science is under-explored. The intrinsic reason may be the immense and intricate nature of ocean data as well as the necessity for higher granularity and richness in knowledge. To alleviate these issues, we introduce OceanGPT, the first-ever LLM in the ocean domain, which is expert in various ocean science tasks. We propose DoInstruct, a novel framework to automatically obtain a large volume of ocean domain instruction data, which generates instructions based on multi-agent collaboration. Additionally, we construct the first oceanography benchmark, OceanBench, to evaluate the capabilities of LLMs in the ocean domain. Though comprehensive experiments, OceanGPT not only shows a higher level of knowledge expertise for oceans science tasks but also gains preliminary embodied intelligence capabilities in ocean technology.

OceanGPT: A Large Language Model for Ocean Science Tasks

acl-long.184

Poster

2310.02031

https://huggingface.co/papers/2310.02031

https://aclanthology.org/2024.acl-long.184/

https://aclanthology.org/2024.acl-long.185.bib

@inproceedings{zhu-etal-2024-beyond, title = "Beyond Memorization: The Challenge of Random Memory Access in Language Models", author = "Zhu, Tongyao and Liu, Qian and Pang, Liang and Jiang, Zhengbao and Kan, Min-Yen and Lin, Min", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.185", pages = "3373--3388", abstract = "Recent developments in Language Models (LMs) have shown their effectiveness in NLP tasks, particularly in knowledge-intensive tasks.However, the mechanisms underlying knowledge storage and memory access within their parameters remain elusive. In this paper, we investigate whether a generative LM (e.g., GPT-2) is able to access its memory sequentially or randomly. Through carefully-designed synthetic tasks, covering the scenarios of full recitation, selective recitation and grounded question answering, we reveal that LMs manage to sequentially access their memory while encountering challenges in randomly accessing memorized content. We find that techniques including recitation and permutation improve the random memory access capability of LMs. Furthermore, by applying this intervention to realistic scenarios of open-domain question answering, we validate that enhancing random access by recitation leads to notable improvements in question answering. The code to reproduce our experiments can be found at https://github.com/sail-sg/lm-random-memory-access.", }

Recent developments in Language Models (LMs) have shown their effectiveness in NLP tasks, particularly in knowledge-intensive tasks.However, the mechanisms underlying knowledge storage and memory access within their parameters remain elusive. In this paper, we investigate whether a generative LM (e.g., GPT-2) is able to access its memory sequentially or randomly. Through carefully-designed synthetic tasks, covering the scenarios of full recitation, selective recitation and grounded question answering, we reveal that LMs manage to sequentially access their memory while encountering challenges in randomly accessing memorized content. We find that techniques including recitation and permutation improve the random memory access capability of LMs. Furthermore, by applying this intervention to realistic scenarios of open-domain question answering, we validate that enhancing random access by recitation leads to notable improvements in question answering. The code to reproduce our experiments can be found at https://github.com/sail-sg/lm-random-memory-access.

Beyond Memorization: The Challenge of Random Memory Access in Language Models

acl-long.185

Oral

2403.07805

https://huggingface.co/papers/2403.07805

https://aclanthology.org/2024.acl-long.185/

https://aclanthology.org/2024.acl-long.186.bib

@inproceedings{kwon-etal-2024-biped, title = "{BIPED}: Pedagogically Informed Tutoring System for {ESL} Education", author = "Kwon, Soonwoo and Kim, Sojung and Park, Minju and Lee, Seunghyun and Kim, Kyuseok", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.186", pages = "3389--3414", abstract = "Large Language Models (LLMs) have a great potential to serve as readily available and cost-efficient Conversational Intelligent Tutoring Systems (CITS) for teaching L2 learners of English. Existing CITS, however, are designed to teach only simple concepts or lack the pedagogical depth necessary to address diverse learning strategies. To develop a more pedagogically informed CITS capable of teaching complex concepts, we construct a BIlingual PEDagogically-informed Tutoring Dataset (BIPED) of one-on-one, human-to-human English tutoring interactions. Through post-hoc analysis of the tutoring interactions, we come up with a lexicon of dialogue acts (34 tutor acts and 9 student acts), which we use to further annotate the collected dataset. Based on a two-step framework of first predicting the appropriate tutor act then generating the corresponding response, we implemented two CITS models using GPT-4 and SOLAR-KO, respectively. We experimentally demonstrate that the implemented models not only replicate the style of human teachers but also employ diverse and contextually appropriate pedagogical strategies.", }

Large Language Models (LLMs) have a great potential to serve as readily available and cost-efficient Conversational Intelligent Tutoring Systems (CITS) for teaching L2 learners of English. Existing CITS, however, are designed to teach only simple concepts or lack the pedagogical depth necessary to address diverse learning strategies. To develop a more pedagogically informed CITS capable of teaching complex concepts, we construct a BIlingual PEDagogically-informed Tutoring Dataset (BIPED) of one-on-one, human-to-human English tutoring interactions. Through post-hoc analysis of the tutoring interactions, we come up with a lexicon of dialogue acts (34 tutor acts and 9 student acts), which we use to further annotate the collected dataset. Based on a two-step framework of first predicting the appropriate tutor act then generating the corresponding response, we implemented two CITS models using GPT-4 and SOLAR-KO, respectively. We experimentally demonstrate that the implemented models not only replicate the style of human teachers but also employ diverse and contextually appropriate pedagogical strategies.

BIPED: Pedagogically Informed Tutoring System for ESL Education

acl-long.186

Poster

2406.03486

https://aclanthology.org/2024.acl-long.186/

https://aclanthology.org/2024.acl-long.187.bib

@inproceedings{chen-etal-2024-timeline, title = "Timeline-based Sentence Decomposition with In Context Learning for Temporal Fact Extraction", author = "Chen, Jianhao and Ouyang, Haoyuan and Ren, Junyang and Ding, Wentao and Hu, Wei and Qu, Yuzhong", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.187", pages = "3415--3432", abstract = "Facts extraction is pivotal for constructing knowledge graphs. Recently, the increasing demand for temporal facts in downstream tasks has led to the emergence of the task of temporal fact extraction. In this paper, we specifically address the extraction of temporal facts from natural language text. Previous studies fail to handle the challenge of establishing time-to-fact correspondences in complex sentences. To overcome this hurdle, we propose a timeline-based sentence decomposition strategy using large language models (LLMs) with in-context learning, ensuring a fine-grained understanding of the timeline associated with various facts. In addition, we evaluate the performance of LLMs for direct temporal fact extraction and get unsatisfactory results. To this end, we introduce TSDRE, a method that incorporates the decomposition capabilities of LLMs into the traditional fine-tuning of smaller pre-trained language models (PLMs). To support the evaluation, we construct ComplexTRED, a complex temporal fact extraction dataset. Our experiments show that TSDRE achieves state-of-the-art results on both HyperRED-Temporal and ComplexTRED datasets.", }

Facts extraction is pivotal for constructing knowledge graphs. Recently, the increasing demand for temporal facts in downstream tasks has led to the emergence of the task of temporal fact extraction. In this paper, we specifically address the extraction of temporal facts from natural language text. Previous studies fail to handle the challenge of establishing time-to-fact correspondences in complex sentences. To overcome this hurdle, we propose a timeline-based sentence decomposition strategy using large language models (LLMs) with in-context learning, ensuring a fine-grained understanding of the timeline associated with various facts. In addition, we evaluate the performance of LLMs for direct temporal fact extraction and get unsatisfactory results. To this end, we introduce TSDRE, a method that incorporates the decomposition capabilities of LLMs into the traditional fine-tuning of smaller pre-trained language models (PLMs). To support the evaluation, we construct ComplexTRED, a complex temporal fact extraction dataset. Our experiments show that TSDRE achieves state-of-the-art results on both HyperRED-Temporal and ComplexTRED datasets.

Timeline-based Sentence Decomposition with In Context Learning for Temporal Fact Extraction

acl-long.187

Poster

https://aclanthology.org/2024.acl-long.187/

https://aclanthology.org/2024.acl-long.188.bib

@inproceedings{aitken-etal-2024-collaboration, title = "Collaboration or Corporate Capture? Quantifying {NLP}{'}s Reliance on Industry Artifacts and Contributions", author = "Aitken, Will and Abdalla, Mohamed and Rudie, Karen and Stinson, Catherine", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.188", pages = "3433--3448", abstract = "Impressive performance of pre-trained models has garnered public attention and made news headlines in recent years. Almost always, these models are produced by or in collaboration with industry. Using them is critical for competing on natural language processing (NLP) benchmarks and correspondingly to stay relevant in NLP research. We surveyed 100 papers published at EMNLP 2022 to determine the degree to which researchers rely on industry models, other artifacts, and contributions to publish in prestigious NLP venues and found that the ratio of their citation is at least three times greater than what would be expected. Our work serves as a scaffold to enable future researchers to more accurately address whether: 1) Collaboration with industry is still collaboration in the absence of an alternative or 2) if NLP inquiry has been captured by the motivations and research direction of private corporations.", }

Impressive performance of pre-trained models has garnered public attention and made news headlines in recent years. Almost always, these models are produced by or in collaboration with industry. Using them is critical for competing on natural language processing (NLP) benchmarks and correspondingly to stay relevant in NLP research. We surveyed 100 papers published at EMNLP 2022 to determine the degree to which researchers rely on industry models, other artifacts, and contributions to publish in prestigious NLP venues and found that the ratio of their citation is at least three times greater than what would be expected. Our work serves as a scaffold to enable future researchers to more accurately address whether: 1) Collaboration with industry is still collaboration in the absence of an alternative or 2) if NLP inquiry has been captured by the motivations and research direction of private corporations.

Collaboration or Corporate Capture? Quantifying NLP's Reliance on Industry Artifacts and Contributions

acl-long.188

Poster

2312.03912

https://aclanthology.org/2024.acl-long.188/

https://aclanthology.org/2024.acl-long.189.bib

@inproceedings{datta-etal-2024-prompt, title = "Prompt Expansion for Adaptive Text-to-Image Generation", author = "Datta, Siddhartha and Ku, Alexander and Ramachandran, Deepak and Anderson, Peter", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.189", pages = "3449--3476", abstract = "Text-to-image generation models are powerful but difficult to use. Users craft specific prompts to get better images, though the images can be repetitive. This paper proposes the Prompt Expansion framework that helps users generate high-quality, diverse images with less effort. The Prompt Expansion model takes a text query as input and outputs a set of expanded text prompts that are optimized such that when passed to a text-to-image model, they generate a wider variety of appealing images. We conduct a human evaluation study that shows that images generated through Prompt Expansion are more aesthetically pleasing and diverse than those generated by baseline methods. Overall, this paper presents a novel and effective approach to improving the text-to-image generation experience.", }

Text-to-image generation models are powerful but difficult to use. Users craft specific prompts to get better images, though the images can be repetitive. This paper proposes the Prompt Expansion framework that helps users generate high-quality, diverse images with less effort. The Prompt Expansion model takes a text query as input and outputs a set of expanded text prompts that are optimized such that when passed to a text-to-image model, they generate a wider variety of appealing images. We conduct a human evaluation study that shows that images generated through Prompt Expansion are more aesthetically pleasing and diverse than those generated by baseline methods. Overall, this paper presents a novel and effective approach to improving the text-to-image generation experience.

Prompt Expansion for Adaptive Text-to-Image Generation

acl-long.189

Poster

2312.16720

https://huggingface.co/papers/2312.16720

https://aclanthology.org/2024.acl-long.189/

https://aclanthology.org/2024.acl-long.190.bib

@inproceedings{huang-etal-2024-progressively, title = "Progressively Modality Freezing for Multi-Modal Entity Alignment", author = "Huang, Yani and Zhang, Xuefeng and Zhang, Richong and Chen, Junfan and Kim, Jaein", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.190", pages = "3477--3489", abstract = "Multi-Modal Entity Alignment aims to discover identical entities across heterogeneous knowledge graphs. While recent studies have delved into fusion paradigms to represent entities holistically, the elimination of features irrelevant to alignment and modal inconsistencies is overlooked, which are caused by inherent differences in multi-modal features. To address these challenges, we propose a novel strategy of progressive modality freezing, called PMF, that focuses on alignment-relevant features and enhances multi-modal feature fusion. Notably, our approach introduces a pioneering cross-modal association loss to foster modal consistency.Empirical evaluations across nine datasets confirm PMF{'}s superiority, demonstrating state-of-the-art performance and the rationale for freezing modalities. Our code is available at https://github.com/ninibymilk/PMF-MMEA.", }

Multi-Modal Entity Alignment aims to discover identical entities across heterogeneous knowledge graphs. While recent studies have delved into fusion paradigms to represent entities holistically, the elimination of features irrelevant to alignment and modal inconsistencies is overlooked, which are caused by inherent differences in multi-modal features. To address these challenges, we propose a novel strategy of progressive modality freezing, called PMF, that focuses on alignment-relevant features and enhances multi-modal feature fusion. Notably, our approach introduces a pioneering cross-modal association loss to foster modal consistency.Empirical evaluations across nine datasets confirm PMF{'}s superiority, demonstrating state-of-the-art performance and the rationale for freezing modalities. Our code is available at https://github.com/ninibymilk/PMF-MMEA.

Progressively Modality Freezing for Multi-Modal Entity Alignment

acl-long.190

Poster

2407.16168

https://aclanthology.org/2024.acl-long.190/

https://aclanthology.org/2024.acl-long.191.bib

@inproceedings{li-etal-2024-llama2vec, title = "{L}lama2{V}ec: Unsupervised Adaptation of Large Language Models for Dense Retrieval", author = "Li, Chaofan and Liu, Zheng and Xiao, Shitao and Shao, Yingxia and Lian, Defu", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.191", pages = "3490--3500", abstract = "Dense retrieval calls for discriminative embeddings to represent the semantic relationship between query and document. It may benefit from the using of large language models (LLMs), given LLMs{'} strong capability on semantic understanding. However, the LLMs are learned by auto-regression, whose working mechanism is completely different from representing whole text as one discriminative embedding. Thus, it is imperative to study how to adapt LLMs properly so that they can be effectively initialized as the backbone encoder for dense retrieval. In this paper, we propose a novel approach, called \textbf{Llama2Vec}, which performs unsupervised adaptation of LLM for its dense retrieval application. Llama2Vec consists of two pretext tasks: EBAE (Embedding-Based Auto-Encoding) and EBAR (Embedding-Based Auto-Regression), where the LLM is prompted to \textit{reconstruct the input sentence} and \textit{predict the next sentence} based on its text embeddings. Llama2Vec is simple, lightweight, but highly effective. It is used to adapt LLaMA-2-7B on the Wikipedia corpus. With a moderate steps of adaptation, it substantially improves the model{'}s fine-tuned performances on a variety of dense retrieval benchmarks. Notably, it results in the new state-of-the-art performances on popular benchmarks, such as passage and document retrieval on MSMARCO, and zero-shot retrieval on BEIR. The model and source code will be made publicly available to facilitate the future research. Our model is available at https://github.com/FlagOpen/FlagEmbedding.", }

Dense retrieval calls for discriminative embeddings to represent the semantic relationship between query and document. It may benefit from the using of large language models (LLMs), given LLMs{'} strong capability on semantic understanding. However, the LLMs are learned by auto-regression, whose working mechanism is completely different from representing whole text as one discriminative embedding. Thus, it is imperative to study how to adapt LLMs properly so that they can be effectively initialized as the backbone encoder for dense retrieval. In this paper, we propose a novel approach, called \textbf{Llama2Vec}, which performs unsupervised adaptation of LLM for its dense retrieval application. Llama2Vec consists of two pretext tasks: EBAE (Embedding-Based Auto-Encoding) and EBAR (Embedding-Based Auto-Regression), where the LLM is prompted to \textit{reconstruct the input sentence} and \textit{predict the next sentence} based on its text embeddings. Llama2Vec is simple, lightweight, but highly effective. It is used to adapt LLaMA-2-7B on the Wikipedia corpus. With a moderate steps of adaptation, it substantially improves the model{'}s fine-tuned performances on a variety of dense retrieval benchmarks. Notably, it results in the new state-of-the-art performances on popular benchmarks, such as passage and document retrieval on MSMARCO, and zero-shot retrieval on BEIR. The model and source code will be made publicly available to facilitate the future research. Our model is available at https://github.com/FlagOpen/FlagEmbedding.

Llama2Vec: Unsupervised Adaptation of Large Language Models for Dense Retrieval

acl-long.191

Poster

https://aclanthology.org/2024.acl-long.191/

https://aclanthology.org/2024.acl-long.192.bib

@inproceedings{nguyen-etal-2024-democratizing, title = "Democratizing {LLM}s for Low-Resource Languages by Leveraging their {E}nglish Dominant Abilities with Linguistically-Diverse Prompts", author = "Nguyen, Xuan-Phi and Aljunied, Mahani and Joty, Shafiq and Bing, Lidong", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.192", pages = "3501--3516", abstract = "Large language models (LLMs) are known to effectively perform tasks by simply observing few exemplars. However, in low-resource languages, obtaining such hand-picked exemplars can still be challenging, where unsupervised techniques may be necessary. Moreover, competent generative capabilities of LLMs are observed only in high-resource languages, while their performances among under-represented languages fall behind due to pre-training data imbalance. To elicit LLMs{'} ability onto low-resource languages without any supervised data, we propose to assemble synthetic exemplars from a diverse set of high-resource languages to prompt the LLMs to translate from any language into English. These prompts are then used to create intra-lingual exemplars to perform tasks in the target languages. Our unsupervised prompting method performs on par with supervised few-shot learning in LLMs of different sizes for translations between English and 13 Indic and 21 African low-resource languages. We also show that fine-tuning a 7B model on data generated from our method helps it perform competitively with a 175B model. In non-English translation tasks, our method even outperforms supervised prompting by up to 3 chrF++ in many low-resource languages. When evaluated on zero-shot multilingual summarization, our method surpasses other English-pivoting baselines by up to 4 ROUGE-L and is also favored by GPT-4.", }

Large language models (LLMs) are known to effectively perform tasks by simply observing few exemplars. However, in low-resource languages, obtaining such hand-picked exemplars can still be challenging, where unsupervised techniques may be necessary. Moreover, competent generative capabilities of LLMs are observed only in high-resource languages, while their performances among under-represented languages fall behind due to pre-training data imbalance. To elicit LLMs{'} ability onto low-resource languages without any supervised data, we propose to assemble synthetic exemplars from a diverse set of high-resource languages to prompt the LLMs to translate from any language into English. These prompts are then used to create intra-lingual exemplars to perform tasks in the target languages. Our unsupervised prompting method performs on par with supervised few-shot learning in LLMs of different sizes for translations between English and 13 Indic and 21 African low-resource languages. We also show that fine-tuning a 7B model on data generated from our method helps it perform competitively with a 175B model. In non-English translation tasks, our method even outperforms supervised prompting by up to 3 chrF++ in many low-resource languages. When evaluated on zero-shot multilingual summarization, our method surpasses other English-pivoting baselines by up to 4 ROUGE-L and is also favored by GPT-4.

Democratizing LLMs for Low-Resource Languages by Leveraging their English Dominant Abilities with Linguistically-Diverse Prompts

acl-long.192

Poster

2306.11372

https://huggingface.co/papers/2306.11372

https://aclanthology.org/2024.acl-long.192/

https://aclanthology.org/2024.acl-long.193.bib

@inproceedings{tong-etal-2024-metaphor, title = "Metaphor Understanding Challenge Dataset for {LLM}s", author = "Tong, Xiaoyu and Choenni, Rochelle and Lewis, Martha and Shutova, Ekaterina", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.193", pages = "3517--3536", abstract = "Metaphors in natural language are a reflection of fundamental cognitive processes such as analogical reasoning and categorisation, and are deeply rooted in everyday communication. Metaphor understanding is therefore an essential task for large language models (LLMs). We release the Metaphor Understanding Challenge Dataset (MUNCH), designed to evaluate the metaphor understanding capabilities of LLMs. The dataset provides over 10k paraphrases for sentences containing metaphor use, as well as 1.5k instances containing inapt paraphrases. The inapt paraphrases were carefully selected to serve as control to determine whether the model indeed performs full metaphor interpretation or rather resorts to lexical similarity. All apt and inapt paraphrases were manually annotated. The metaphorical sentences cover natural metaphor uses across 4 genres (academic, news, fiction, and conversation), and they exhibit different levels of novelty. Experiments with LLaMA and GPT-3.5 demonstrate that MUNCH presents a challenging task for LLMs. The dataset is freely accessible at https://github.com/xiaoyuisrain/metaphor-understanding-challenge.", }

Metaphors in natural language are a reflection of fundamental cognitive processes such as analogical reasoning and categorisation, and are deeply rooted in everyday communication. Metaphor understanding is therefore an essential task for large language models (LLMs). We release the Metaphor Understanding Challenge Dataset (MUNCH), designed to evaluate the metaphor understanding capabilities of LLMs. The dataset provides over 10k paraphrases for sentences containing metaphor use, as well as 1.5k instances containing inapt paraphrases. The inapt paraphrases were carefully selected to serve as control to determine whether the model indeed performs full metaphor interpretation or rather resorts to lexical similarity. All apt and inapt paraphrases were manually annotated. The metaphorical sentences cover natural metaphor uses across 4 genres (academic, news, fiction, and conversation), and they exhibit different levels of novelty. Experiments with LLaMA and GPT-3.5 demonstrate that MUNCH presents a challenging task for LLMs. The dataset is freely accessible at https://github.com/xiaoyuisrain/metaphor-understanding-challenge.

Metaphor Understanding Challenge Dataset for LLMs

acl-long.193

Poster

2403.11810

https://aclanthology.org/2024.acl-long.193/

https://aclanthology.org/2024.acl-long.194.bib

@inproceedings{zhang-etal-2024-multi-task, title = "A Multi-Task Embedder For Retrieval Augmented {LLM}s", author = "Zhang, Peitian and Liu, Zheng and Xiao, Shitao and Dou, Zhicheng and Nie, Jian-Yun", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.194", pages = "3537--3553", abstract = "LLMs confront inherent limitations in terms of its knowledge, memory, and action. The retrieval augmentation stands as a vital mechanism to address these limitations, which brings in useful information from external sources to augment the LLM. However, existing retrieval methods encounter two pressing issues. On one hand, the general retrievers are not properly optimized for retrieval augmentation hence exhibit limited effectiveness; on the other hand, the task-specific retrievers excel in the targeted retrieval augmentation scenario, while lack the versatility to handle diverse scenarios. In this work, we propose LLM-Embedder for the unified support of diverse retrieval augmentation scenarios. Our method presents three technical contributions. Firstly, we introduce a new reward formulation, namely rank-aware reward. It exploits the ranking position of the desired output among N sampled outputs from the LLM, which leads to fine-grained and robust computation of reward from the LLM{'}s feedback. Secondly, we design a novel distillation objective, called graded distillation. It incorporates both the absolute value and the relative order of the reward for more sufficient utilization of the LLM{'}s feedback. Thirdly, we systematically optimize the multi-task learning, which effectively unifies the multiple retrieval functionalities into one model. In our experiment, LLM-Embedder substantially improves the LLM{'}s performances in various downstream tasks, while introducing superior retrieval augmentation{'}s effect over both general and task-specifc retrievers. Our data, code, and model have been released at https://github.com/FlagOpen/FlagEmbedding.", }

LLMs confront inherent limitations in terms of its knowledge, memory, and action. The retrieval augmentation stands as a vital mechanism to address these limitations, which brings in useful information from external sources to augment the LLM. However, existing retrieval methods encounter two pressing issues. On one hand, the general retrievers are not properly optimized for retrieval augmentation hence exhibit limited effectiveness; on the other hand, the task-specific retrievers excel in the targeted retrieval augmentation scenario, while lack the versatility to handle diverse scenarios. In this work, we propose LLM-Embedder for the unified support of diverse retrieval augmentation scenarios. Our method presents three technical contributions. Firstly, we introduce a new reward formulation, namely rank-aware reward. It exploits the ranking position of the desired output among N sampled outputs from the LLM, which leads to fine-grained and robust computation of reward from the LLM{'}s feedback. Secondly, we design a novel distillation objective, called graded distillation. It incorporates both the absolute value and the relative order of the reward for more sufficient utilization of the LLM{'}s feedback. Thirdly, we systematically optimize the multi-task learning, which effectively unifies the multiple retrieval functionalities into one model. In our experiment, LLM-Embedder substantially improves the LLM{'}s performances in various downstream tasks, while introducing superior retrieval augmentation{'}s effect over both general and task-specifc retrievers. Our data, code, and model have been released at https://github.com/FlagOpen/FlagEmbedding.

A Multi-Task Embedder For Retrieval Augmented LLMs

acl-long.194

Poster

https://aclanthology.org/2024.acl-long.194/

https://aclanthology.org/2024.acl-long.195.bib

@inproceedings{lee-lim-2024-language, title = "Language Models Don{'}t Learn the Physical Manifestation of Language", author = "Lee, Bruce and Lim, Jaehyuk", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.195", pages = "3554--3579", abstract = "We argue that language-only models don{'}t learn the physical manifestation of language. We present an empirical investigation of visual-auditory properties of language through a series of tasks, termed H-Test.These tasks highlight a fundamental gap between human linguistic understanding and the sensory-deprived linguistic understanding of LLMs. In support of our hypothesis, 1. deliberate reasoning (Chain-of-Thought), 2. few-shot examples, or 3. stronger LLM from the same model family (LLaMA 2 13B -{\textgreater} LLaMA 2 70B) has no significant effect on H-Test performance. We bring in the philosophical case of Mary, who learns about the world in a sensory-deprived environment as a useful conceptual framework to understand how language-only models learn about the world (Jackson, 1986). Our experiments show that some of the strongest proprietary LLMs stay near random chance baseline accuracy of 50{\%}, highlighting the limitations of linguistic knowledge acquired in the absence of sensory experience. Our code and data are available at {\textless}github.com/brucewlee/h-test{\textgreater}.", }

We argue that language-only models don{'}t learn the physical manifestation of language. We present an empirical investigation of visual-auditory properties of language through a series of tasks, termed H-Test.These tasks highlight a fundamental gap between human linguistic understanding and the sensory-deprived linguistic understanding of LLMs. In support of our hypothesis, 1. deliberate reasoning (Chain-of-Thought), 2. few-shot examples, or 3. stronger LLM from the same model family (LLaMA 2 13B -{\textgreater} LLaMA 2 70B) has no significant effect on H-Test performance. We bring in the philosophical case of Mary, who learns about the world in a sensory-deprived environment as a useful conceptual framework to understand how language-only models learn about the world (Jackson, 1986). Our experiments show that some of the strongest proprietary LLMs stay near random chance baseline accuracy of 50{\%}, highlighting the limitations of linguistic knowledge acquired in the absence of sensory experience. Our code and data are available at {\textless}github.com/brucewlee/h-test{\textgreater}.

Language Models Don't Learn the Physical Manifestation of Language

acl-long.195

Poster

2402.11349

https://aclanthology.org/2024.acl-long.195/

https://aclanthology.org/2024.acl-long.196.bib

@inproceedings{feng-etal-2024-bot, title = "What Does the Bot Say? Opportunities and Risks of Large Language Models in Social Media Bot Detection", author = "Feng, Shangbin and Wan, Herun and Wang, Ningnan and Tan, Zhaoxuan and Luo, Minnan and Tsvetkov, Yulia", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.196", pages = "3580--3601", abstract = "Social media bot detection has always been an arms race between advancements in machine learning bot detectors and adversarial bot strategies to evade detection. In this work, we bring the arms race to the next level by investigating the opportunities and risks of state-of-the-art large language models (LLMs) in social bot detection. To investigate the opportunities, we design novel LLM-based bot detectors by proposing a mixture-of-heterogeneous-experts framework to divide and conquer diverse user information modalities. To illuminate the risks, we explore the possibility of LLM-guided manipulation of user textual and structured information to evade detection. Extensive experiments with three LLMs on two datasets demonstrate that instruction tuning on merely 1,000 annotated examples produces specialized LLMs that outperform state-of-the-art baselines by up to 9.1{\%} on both datasets, while LLM-guided manipulation strategies could significantly bring down the performance of existing bot detectors by up to 29.6{\%} and harm the calibration and reliability of bot detection systems.", }

Social media bot detection has always been an arms race between advancements in machine learning bot detectors and adversarial bot strategies to evade detection. In this work, we bring the arms race to the next level by investigating the opportunities and risks of state-of-the-art large language models (LLMs) in social bot detection. To investigate the opportunities, we design novel LLM-based bot detectors by proposing a mixture-of-heterogeneous-experts framework to divide and conquer diverse user information modalities. To illuminate the risks, we explore the possibility of LLM-guided manipulation of user textual and structured information to evade detection. Extensive experiments with three LLMs on two datasets demonstrate that instruction tuning on merely 1,000 annotated examples produces specialized LLMs that outperform state-of-the-art baselines by up to 9.1{\%} on both datasets, while LLM-guided manipulation strategies could significantly bring down the performance of existing bot detectors by up to 29.6{\%} and harm the calibration and reliability of bot detection systems.

What Does the Bot Say? Opportunities and Risks of Large Language Models in Social Media Bot Detection

acl-long.196

Poster

2402.00371

https://aclanthology.org/2024.acl-long.196/

https://aclanthology.org/2024.acl-long.197.bib

@inproceedings{zhang-etal-2024-self-contrast, title = "Self-Contrast: Better Reflection Through Inconsistent Solving Perspectives", author = "Zhang, Wenqi and Shen, Yongliang and Wu, Linjuan and Peng, Qiuying and Wang, Jun and Zhuang, Yueting and Lu, Weiming", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.197", pages = "3602--3622", abstract = "The reflection capacity of Large Language Model (LLM) has garnered extensive attention. A post-hoc prompting strategy, e.g., reflexion and self-refine, refines LLM{'}s response based on self-evaluated or external feedback. However, recent research indicates without external feedback, LLM{'}s intrinsic reflection is unstable. Our investigation unveils that the key bottleneck is the quality of the self-evaluated feedback. We find LLMs often exhibit overconfidence or high randomness when self-evaluate, offering stubborn or inconsistent feedback, which causes poor reflection. To remedy this, we advocate Self-Contrast: It adaptively explores diverse solving perspectives tailored to the request, contrasts the differences, and summarizes these discrepancies into a checklist which could be used to re-examine and eliminate discrepancies. Our method endows LLM with diverse perspectives to alleviate stubborn biases. Moreover, their discrepancies indicate potential errors or inherent uncertainties that LLM often overlooks. Reflecting upon these can catalyze more accurate and stable reflection. Experiments conducted on a series of reasoning and translation tasks with different LLMs serve to underscore the effectiveness and generality of our strategy.", }

The reflection capacity of Large Language Model (LLM) has garnered extensive attention. A post-hoc prompting strategy, e.g., reflexion and self-refine, refines LLM{'}s response based on self-evaluated or external feedback. However, recent research indicates without external feedback, LLM{'}s intrinsic reflection is unstable. Our investigation unveils that the key bottleneck is the quality of the self-evaluated feedback. We find LLMs often exhibit overconfidence or high randomness when self-evaluate, offering stubborn or inconsistent feedback, which causes poor reflection. To remedy this, we advocate Self-Contrast: It adaptively explores diverse solving perspectives tailored to the request, contrasts the differences, and summarizes these discrepancies into a checklist which could be used to re-examine and eliminate discrepancies. Our method endows LLM with diverse perspectives to alleviate stubborn biases. Moreover, their discrepancies indicate potential errors or inherent uncertainties that LLM often overlooks. Reflecting upon these can catalyze more accurate and stable reflection. Experiments conducted on a series of reasoning and translation tasks with different LLMs serve to underscore the effectiveness and generality of our strategy.

Self-Contrast: Better Reflection Through Inconsistent Solving Perspectives

acl-long.197

Poster

2401.02009

https://huggingface.co/papers/2401.02009

https://aclanthology.org/2024.acl-long.197/

https://aclanthology.org/2024.acl-long.198.bib

@inproceedings{zhou-etal-2024-relying, title = "Relying on the Unreliable: The Impact of Language Models{'} Reluctance to Express Uncertainty", author = "Zhou, Kaitlyn and Hwang, Jena and Ren, Xiang and Sap, Maarten", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.198", pages = "3623--3643", abstract = "As natural language becomes the default interface for human-AI interaction, there is a need for LMs to appropriately communicate uncertainties in downstream applications. In this work, we investigate how LMs incorporate confidence in responses via natural language and how downstream users behave in response to LM-articulated uncertainties. We examine publicly deployed models and find that LMs are reluctant to express uncertainties when answering questions even when they produce incorrect responses. LMs can be explicitly prompted to express confidences, but tend to be overconfident, resulting in high error rates (an average of 47{\%}) among confident responses. We test the risks of LM overconfidence by conducting human experiments and show that users rely heavily on LM generations, whether or not they are marked by certainty. Lastly, we investigate the preference-annotated datasets used in post training alignment and find that humans are biased against texts with uncertainty. Our work highlights new safety harms facing human-LM interactions and proposes design recommendations and mitigating strategies moving forward.", }

As natural language becomes the default interface for human-AI interaction, there is a need for LMs to appropriately communicate uncertainties in downstream applications. In this work, we investigate how LMs incorporate confidence in responses via natural language and how downstream users behave in response to LM-articulated uncertainties. We examine publicly deployed models and find that LMs are reluctant to express uncertainties when answering questions even when they produce incorrect responses. LMs can be explicitly prompted to express confidences, but tend to be overconfident, resulting in high error rates (an average of 47{\%}) among confident responses. We test the risks of LM overconfidence by conducting human experiments and show that users rely heavily on LM generations, whether or not they are marked by certainty. Lastly, we investigate the preference-annotated datasets used in post training alignment and find that humans are biased against texts with uncertainty. Our work highlights new safety harms facing human-LM interactions and proposes design recommendations and mitigating strategies moving forward.

Relying on the Unreliable: The Impact of Language Models' Reluctance to Express Uncertainty

acl-long.198

Poster

2401.06730

https://aclanthology.org/2024.acl-long.198/

https://aclanthology.org/2024.acl-long.199.bib

@inproceedings{wang-etal-2024-unity, title = "Unity in Diversity: Collaborative Pre-training Across Multimodal Medical Sources", author = "Wang, Xiaochen and Luo, Junyu and Wang, Jiaqi and Zhong, Yuan and Zhang, Xiaokun and Wang, Yaqing and Bhatia, Parminder and Xiao, Cao and Ma, Fenglong", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.199", pages = "3644--3656", abstract = "Although pre-training has become a prevalent approach for addressing various biomedical tasks, the current efficacy of pre-trained models is hindered by their reliance on a limited scope of medical sources. This limitation results in data scarcity during pre-training and restricts the range of applicable downstream tasks. In response to these challenges, we develop MedCSP, a new pre-training strategy designed to bridge the gap between multimodal medical sources. MedCSP employs modality-level aggregation to unify patient data within individual sources. Additionally, leveraging temporal information and diagnosis history, MedCSP effectively captures explicit and implicit correlations between patients across different sources. To evaluate the proposed strategy, we conduct comprehensive experiments, where the experiments are based on 6 modalities from 2 real-world medical data sources, and MedCSP is evaluated on 4 tasks against 19 baselines, marking an initial yet essential step towards cross-source modeling in the medical domain.", }

Although pre-training has become a prevalent approach for addressing various biomedical tasks, the current efficacy of pre-trained models is hindered by their reliance on a limited scope of medical sources. This limitation results in data scarcity during pre-training and restricts the range of applicable downstream tasks. In response to these challenges, we develop MedCSP, a new pre-training strategy designed to bridge the gap between multimodal medical sources. MedCSP employs modality-level aggregation to unify patient data within individual sources. Additionally, leveraging temporal information and diagnosis history, MedCSP effectively captures explicit and implicit correlations between patients across different sources. To evaluate the proposed strategy, we conduct comprehensive experiments, where the experiments are based on 6 modalities from 2 real-world medical data sources, and MedCSP is evaluated on 4 tasks against 19 baselines, marking an initial yet essential step towards cross-source modeling in the medical domain.

Unity in Diversity: Collaborative Pre-training Across Multimodal Medical Sources

acl-long.199

Poster

https://aclanthology.org/2024.acl-long.199/

https://aclanthology.org/2024.acl-long.200.bib

@inproceedings{papi-etal-2024-good, title = "When Good and Reproducible Results are a Giant with Feet of Clay: The Importance of Software Quality in {NLP}", author = "Papi, Sara and Gaido, Marco and Pilzer, Andrea and Negri, Matteo", editor = "Ku, Lun-Wei and Martins, Andre and Srikumar, Vivek", booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = aug, year = "2024", address = "Bangkok, Thailand", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.acl-long.200", pages = "3657--3672", abstract = "Despite its crucial role in research experiments, code correctness is often presumed solely based on the perceived quality of results. This assumption, however, comes with the risk of erroneous outcomes and, in turn, potentially misleading findings. To mitigate this risk, we posit that the current focus on reproducibility should go hand in hand with the emphasis on software quality. We support our arguments with a case study in which we identify and fix three bugs in widely used implementations of the state-of-the-art Conformer architecture. Through experiments on speech recognition and translation in various languages, we demonstrate that the presence of bugs does not prevent the achievement of good and reproducible results, which however can lead to incorrect conclusions that potentially misguide future research. As countermeasures, we release pangoliNN, a library dedicated to testing neural models, and propose a Code-quality Checklist, with the goal of promoting coding best practices and improving software quality within the NLP community.", }

Despite its crucial role in research experiments, code correctness is often presumed solely based on the perceived quality of results. This assumption, however, comes with the risk of erroneous outcomes and, in turn, potentially misleading findings. To mitigate this risk, we posit that the current focus on reproducibility should go hand in hand with the emphasis on software quality. We support our arguments with a case study in which we identify and fix three bugs in widely used implementations of the state-of-the-art Conformer architecture. Through experiments on speech recognition and translation in various languages, we demonstrate that the presence of bugs does not prevent the achievement of good and reproducible results, which however can lead to incorrect conclusions that potentially misguide future research. As countermeasures, we release pangoliNN, a library dedicated to testing neural models, and propose a Code-quality Checklist, with the goal of promoting coding best practices and improving software quality within the NLP community.

When Good and Reproducible Results are a Giant with Feet of Clay: The Importance of Software Quality in NLP

acl-long.200

Oral

2303.16166

https://huggingface.co/papers/2303.16166

https://aclanthology.org/2024.acl-long.200/