arxiv_id
stringlengths 9
12
| abstract
stringlengths 431
13.8k
|
|---|---|
2112.09332 | We fine-tune GPT-3 to answer long-form questions using a text-based web-browsing environment, which allows the model to search and navigate the web. By setting up the task so that it can be performed by humans, we are able to train models on the task using imitation learning, and then optimize answer quality with human feedback. To make human evaluation of factual accuracy easier, models must collect references while browsing in support of their answers. We train and evaluate our models on ELI5, a dataset of questions asked by Reddit users. Our best model is obtained by fine-tuning GPT-3 using behavior cloning, and then performing rejection sampling against a reward model trained to predict human preferences. This model’s answers are preferred by humans 56% of the time to those of our human demonstrators, and 69% of the time to the highest-voted answer from Reddit. |
1911.02150 | Multi-head attention layers, as used in the Transformer neural sequence model, are a powerful alternative to RNNs for moving information across and between sequences. While training these layers is generally fast and simple, due to parallelizability across the length of the sequence, incremental inference (where such paralleization is impossible) is often slow, due to the memory-bandwidth cost of repeatedly loading the large "keys" and "values" tensors. We propose a variant called multi-query attention, where the keys and values are shared across all of the different attention "heads", greatly reducing the size of these tensors and hence the memory bandwidth requirements of incremental decoding. We verify experimentally that the resulting models can indeed be much faster to decode, and incur only minor quality degradation from the baseline. |
2209.00588 | Deep reinforcement learning agents are notoriously sample inefficient, which considerably limits their application to real-world problems. Recently, many model-based methods have been designed to address this issue, with learning in the imagination of a world model being one of the most prominent approaches. However, while virtually unlimited interaction with a simulated environment sounds appealing, the world model has to be accurate over extended periods of time. Motivated by the success of Transformers in sequence modeling tasks, we introduceiris, a data-efficient agent that learns in a world model composed of a discrete autoencoder and an autoregressive Transformer. With the equivalent of only two hours of gameplay in the Atari 100k benchmark,irisachieves a mean human normalized score of 1.046, and outperforms humans on 10 out of 26 games, setting a new state of the art for methods without lookahead search. To foster future research on Transformers and world models for sample-efficient reinforcement learning, we release our code and models athttps://github.com/eloialonso/iris. |
2209.00626 | In coming decades, artificial general intelligence (AGI) may surpass human capabilities at many critical tasks. We argue that, without substantial effort to prevent it, AGIs could learn to pursue goals that conflict (i.e., are misaligned) with human interests. If trained like today’s most capable models, AGIs could learn to act deceptively to receive higher reward, learn internally-represented goals which generalize beyond their fine-tuning distributions, and pursue those goals using power-seeking strategies. We review emerging evidence for these properties. AGIs with these properties would be difficult to align and may appear aligned even when they are not. We outline how the deployment of misaligned AGIs might irreversibly undermine human control over the world, and briefly review research directions aimed at preventing this outcome. |
2101.00190 | Fine-tuning is the de facto way to leverage large pretrained language models to perform downstream tasks.
However, it modifies all the language model parameters and therefore necessitates storing a full copy for each task.
In this paper, we propose prefix-tuning, a lightweight alternative to fine-tuning for natural language generation tasks, which keeps language model parameters frozen, but optimizes a smallcontinuous task-specificvector (called the prefix) .
Prefix-tuning draws inspiration from prompting, allowing subsequent tokens to attend to this prefix as if it were “virtual tokens”.
We apply prefix-tuning to GPT-2 for table-to-text generation and to BART for summarization.
We find that by learning only 0.1% of the parameters, prefix-tuning obtains comparable performance in the full data setting, outperforms fine-tuning in low-data settings, and extrapolates better to examples with topics unseen during training. |
2303.11366v1 | Large language models (LLMs) have been increasingly used to interact with external
environments (e.g., games, compilers, APIs) as goal-driven agents. However, it
remains challenging for these language agents to quickly and efficiently learn from
trial-and-error as traditional reinforcement learning methods require extensive
training samples and expensive model fine-tuning. We proposeReflexion, a
novel framework to reinforce language agents not by updating weights, but instead
through linguistic feedback. Concretely, Reflexion agents verbally reflect on
task feedback signals, then maintain their own reflective text in an episodic
memory buffer to induce better decision-making in subsequent trials. Reflexion is
flexible enough to incorporate various types (scalar values or free-form
language) and sources (external or internally simulated) of feedback signals,
and obtains significant improvements over a baseline agent across diverse tasks
(sequential decision-making, coding, language reasoning) . For example, Reflexion
achieves a 91% pass@1 accuracy on the HumanEval coding benchmark, surpassing
the previous state-of-the-art GPT-4 that achieves 80%. We also conduct ablation
and analysis studies using different feedback signals, feedback incorporation
methods, and agent types, and provide insights into how they affect performance.
We release all code, demos, and datasets athttps://github.com/noahshinn024/reflexion. |
2110.07602 | Prompt tuning, which only tunes continuous prompts with a frozen language model, substantially reduces per-task storage and memory usage at training. However, in the context of NLU, prior work reveals that prompt tuning does not perform well for normal-sized pretrained models. We also find that existing methods of prompt tuning cannot handle hard sequence labeling tasks, indicating a lack of universality. We present a novel empirical finding that properly optimized prompt tuning can be universally effective across a wide range of model scales and NLU tasks. It matches the performance of finetuning while having only 0.1%-3% tuned parameters. Our method P-Tuning v2 is an implementation of Deep Prompt Tuningoptimized and adapted for NLU. Given the universality and simplicity of P-Tuning v2, we believe it can serve as an alternative to finetuning and a strong baseline for future research.111Our code and data are released athttps://github.com/THUDM/P-tuning-v2. |
2103.10385 | Prompting a pretrained language model with natural language patterns has been proved effective for natural language understanding (NLU) . However,
our preliminary study reveals that manual discrete prompts often lead to unstable performance—e.g., changing a single word in the prompt might result in substantial performance drop. We propose a novel method P-Tuning that employs trainable continuous prompt embeddings in concatenation with discrete prompts. Empirically, P-Tuning not only stabilizes training by minimizing the gap between various discrete prompts, but also improves performance by a sizeable margin on a wide range of NLU tasks including LAMA and SuperGLUE. P-Tuning is generally effective for both frozen and tuned language models, under both the fully-supervised and few-shot settings. |
2104.08691 | In this work, we explore “prompt tuning,” a simple yet effective mechanism for learning “soft prompts” to condition frozen language models to perform specific downstream tasks.
Unlike the discrete text prompts used byGPT-3, soft prompts are learned through backpropagation and can be tuned to incorporate signals from any number of labeled examples.
Our end-to-end learned approach outperformsGPT-3’sfew-shot learning by a large margin.
More remarkably, through ablations on model size using T5, we show that prompt tuning becomes more competitive with scale: as models exceed billions of parameters, our method “closes the gap” and matches the strong performance of model tuning (where all model weights are tuned) .
This finding is especially relevant because large models are costly to share and serve and the ability to reuse one frozen model for multiple downstream tasks can ease this burden.
Our method can be seen as a simplification of the recently proposed “prefix tuning” ofand we provide a comparison to this and other similar approaches.
Finally, we show that conditioning a frozen model with soft prompts confers benefits in robustness to domain transfer and enables efficient “prompt ensembling.” |
2302.00482 | Continuous normalizing flows (CNFs) are an attractive generative modeling technique, but they have been held back by limitations in their simulation-based maximum likelihood training. We introduce the generalized *conditional flow matching* (CFM) technique, a family of simulation-free training objectives for CNFs. CFM features a stable regression objective like that used to train the stochastic flow in diffusion models but enjoys the efficient inference of deterministic flow models. In contrast to both diffusion models and prior CNF training algorithms, CFM does not require the source distribution to be Gaussian or require evaluation of its density. A variant of our objective is *optimal transport CFM* (OT-CFM), which creates simpler flows that are more stable to train and lead to faster inference, as evaluated in our experiments. Furthermore, we show that when the true OT plan is available, our OT-CFM method approximates dynamic OT. Training CNFs with CFM improves results on a variety of conditional and unconditional generation tasks, such as inferring single cell dynamics, unsupervised image translation, and Schrödinger bridge inference. The Python code is available at https://github.com/atong01/conditional-flow-matching. |
2212.10496 | While dense retrieval has been shown effective and efficient across tasks and languages, it remains difficult to create effective fully zero-shot dense retrieval systems when no relevance label is available.
In this paper, we recognize the difficulty of zero-shot learning and encoding relevance. Instead, we propose to pivot throughHypotheticalDocumentEmbeddings (HyDE) .
Given a query,HyDEfirst zero-shot instructs an instruction-following language model (e.g.InstructGPT) to generate ahypotheticaldocument.
The document captures relevance patterns but is unreal and may contain false details. Then, an unsupervised contrastively learned encoder (e.g.Contriever) encodes the document into an embedding vector. This vector identifies a neighborhood in the corpus embedding space, where similarrealdocuments are retrieved based on vector similarity. This second step ground the generated document to the actual corpus, with the encoder’s dense bottleneck filtering out the incorrect details.
Our experiments show thatHyDEsignificantly outperforms the state-of-the-art unsupervised dense retrieverContrieverand shows strong performance comparable to fine-tuned retrievers, across various tasks (e.g. web search, QA, fact verification) and languages (e.g. sw, ko, ja) .111No models were trained or fine-tuned in making this pre-print. Our open source code is available athttps://github.com/texttron/hyde. |
2303.07295v1 | Most language models (LMs) are trained and applied in an autoregressive
left-to-right fashion,
assuming that the next token only depends on the preceding ones. However, this assumption
ignores the potential benefits of using the full sequence information during training, and the
possibility of having context from both sides during inference.
In this paper, we propose a new pre-training paradigm with techniques
that jointly improve the training data efficiency and the
capabilities of the LMs in the infilling task.
The first is a training objective that aligns the predictions of a left-to-right LM with those of a right-to-left LM, trained on the same data but in reverse order. The second is a bidirectional inference procedure that enables both LMs to meet in the middle. We show the effectiveness of our pre-training paradigm with extensive experiments on both programming and natural language models, outperforming strong baselines.111Code and models available athttps://github.com/microsoft/Meet-in-the-Middle |
2303.09752 | Many natural language processing tasks benefit from long inputs, but processing long documents with Transformers is expensive – not only due to quadratic attention complexity but also from applying feedforward and projection layers to every token. However, not all tokens are equally important, especially for longer documents. We proposeCoLT5, a long-input Transformer model that builds on this intuition by employing conditional computation, devoting more resources to important tokens in both feedforward and attention layers. We show thatCoLT5achieves stronger performance thanLongT5with much faster training and inference, achieving SOTA on the long-input SCROLLS benchmark. Moreover,CoLT5can effectively and tractably make use of extremely long inputs, showing strong gains up to 64k input length. |
2106.06981 | What is the computational model behind a Transformer?
Where recurrent neural networks have direct parallels in finite state machines, allowing clear discussion and thought around architecture variants or trained models, Transformers have no such familiar parallel.
In this paper we aim to change that, proposing a computational model for the transformer-encoder in the form of a programming language.
We map the basic components of a transformer-encoder—attention and feed-forward computation—into simple primitives, around which we form a programming language: the Restricted Access Sequence Processing Language (RASP) .
We show how RASP can be used to program solutions to tasks that could conceivably be learned by a Transformer, and how a Transformer can be trained to mimic a RASP solution.
In particular, we provide RASP programs for histograms, sorting, and Dyck-languages.
We further use our model to relate their difficulty in terms of the number of required layers and attention heads:
analyzing a RASP program implies a maximum number of heads and layers necessary to encode a task in a transformer.
Finally, we see how insights gained from our abstraction might be used to explain phenomena seen in recent works. |
2009.01325v3 | As language models become more powerful, training and evaluation are increasingly bottlenecked by the data and metrics used for a particular task. For example, summarization models are often trained to predict human reference summaries and evaluated using ROUGE, but both of these metrics are rough proxies for what we really care about—summary quality. In this work, we show that it is possible to significantly improve summary quality by training a model to optimize for human preferences. We collect a large, high-quality dataset of human comparisons between summaries, train a model to predict the human-preferred summary, and use that model as a reward function to fine-tune a summarization policy using reinforcement learning. We apply our method to a version of the TL;DR dataset of Reddit postsand find that our models significantly outperform both human reference summaries and much larger models fine-tuned with supervised learning alone. Our models also transfer to CNN/DM news articles, producing summaries nearly as good as the human reference without any news-specific fine-tuning.111Samples from all of our models can be viewedon our website.We conduct extensive analyses to understand our human feedback dataset and fine-tuned models.222We provide inference code for our 1.3B models and baselines, as well as a model card and our human feedback dataset with over 64k summary comparisons,here.We establish that our reward model generalizes to new datasets, and that optimizing our reward model results in better summaries than optimizing ROUGE according to humans.
We hope the evidence from our paper motivates machine learning researchers to pay closer attention to how their training loss affects the model behavior they actually want. |
2303.12712 | Artificial intelligence (AI) researchers have been developing and refining large language models (LLMs) that exhibit remarkable capabilities across a variety of domains and tasks, challenging our understanding of learning and cognition. The latest model developed by OpenAI, GPT-4 [Ope23], was trained using an unprecedented scale of compute and data. In this paper, we report on our investigation of an early version of GPT-4, when it was still in active development by OpenAI. We contend that (this early version of) GPT4 is part of a new cohort of LLMs (along with ChatGPT and Google's PaLM for example) that exhibit more general intelligence than previous AI models. We discuss the rising capabilities and implications of these models. We demonstrate that, beyond its mastery of language, GPT-4 can solve novel and difficult tasks that span mathematics, coding, vision, medicine, law, psychology and more, without needing any special prompting. Moreover, in all of these tasks, GPT-4's performance is strikingly close to human-level performance, and often vastly surpasses prior models such as ChatGPT. Given the breadth and depth of GPT-4's capabilities, we believe that it could reasonably be viewed as an early (yet still incomplete) version of an artificial general intelligence (AGI) system. In our exploration of GPT-4, we put special emphasis on discovering its limitations, and we discuss the challenges ahead for advancing towards deeper and more comprehensive versions of AGI, including the possible need for pursuing a new paradigm that moves beyond next-word prediction. We conclude with reflections on societal influences of the recent technological leap and future research directions. |
2205.05131 | Existing pre-trained models are generally geared towards a particular class of problems. To date, there seems to be still no consensus on what the right architecture and pre-training setup should be. This paper presents a unified framework for pre-training models that are universally effective across datasets and setups.
We begin by disentangling architectural archetypes with pre-training objectives – two concepts that are commonly conflated. Next, we present a generalized and unified perspective for self-supervision in NLP and show how different pre-training objectives can be cast as one another and how interpolating between different objectives can be effective. We then propose Mixture-of-Denoisers (MoD) , a pre-training objective that combines diverse pre-training paradigms together. We furthermore introduce a notion of mode switching, wherein downstream fine-tuning is associated with specific pre-training schemes. We conduct extensive ablative experiments to compare multiple pre-training objectives and find that our method pushes the Pareto-frontier by outperforming T5 and/or GPT-like models across multiple diverse setups. Finally, by scaling our model up to 20B parameters, we achieve SOTA performance on 50 well-established supervised NLP tasks ranging from language generation (with automated and human evaluation) , language understanding, text classification, question answering, commonsense reasoning, long text reasoning, structured knowledge grounding and information retrieval. Our model also achieve strong results at in-context learning, outperforming 175B GPT-3 (published paper results) on zero-shot SuperGLUE and tripling the performance of T5-XXL on one-shot summarization. On zero-shot MMLU, UL2 20B outperforms T0 and T5 models. Additionally, we show that UL2 20B works well with chain-of-thought prompting and reasoning, making it an appealing choice for research into reasoning at a small to medium scale of 20B parameters. Finally, we apply FLAN instruction tuning to the UL2 20B model, achieving MMLU and Big-Bench scores competitive to FLAN-PaLM 62B. We release Flax-based T5X model checkpoints for the UL2 20B model and Flan-UL2 20B model athttps://github.com/google-research/google-research/tree/master/ul2. |
2210.13382 | Language models show a surprising range of capabilities, but the source of their apparent competence is unclear. Do these networks just memorize a collection of surface statistics, or do they rely on internal representations of the process that generates the sequences they see? We investigate this question in a synthetic setting by applying a variant of the GPT model to the task of predicting legal moves in a simple board game, Othello. Although the network has no a priori knowledge of the game or its rules, we uncover evidence of an emergent nonlinear internal representation of the board state. Interventional experiments indicate this representation can be used to control the output of the network. By leveraging these intervention techniques, we produce “latent saliency maps” that help explain predictions.111Codes athttps://github.com/likenneth/othello_world |
1904.10509 | Transformers are powerful sequence models, but require time and memory that grows quadratically with the sequence length. In this paper we introduce sparse factorizations of the attention matrix which reduce this to. We also introduce a) a variation on architecture and initialization to train deeper networks, b) the recomputation of attention matrices to save memory, and c) fast attention kernels for training. We call networks with these changes Sparse Transformers, and show they can model sequences tens of thousands of timesteps long using hundreds of layers. We use the same architecture to model images, audio, and text from raw bytes, setting a new state of the art for density modeling of Enwik8, CIFAR-10, and ImageNet-64. We generate unconditional samples that demonstrate global coherence and great diversity, and show it is possible in principle to use self-attention to model sequences of length one million or more. |
2303.11156v1 | The rapid progress of large language models (LLMs) has made them capable of performing astonishingly well on various tasks including document completion and question answering. The unregulated use of these models, however, can potentially lead to malicious consequences such as plagiarism, generating fake news, spamming, etc. Therefore, reliable detection of AI-generated text can be critical to ensure the responsible use of LLMs. Recent works attempt to tackle this problem either using certain model signatures present in the generated text outputs or by applying watermarking techniques that imprint specific patterns onto them. In this paper, both empirically and theoretically, we show that these detectors are not reliable in practical scenarios. Empirically, we show thatparaphrasing attacks, where a light paraphraser is applied on top of the generative text model, can break a whole range of detectors, including the ones using the watermarking schemes as well as neural network-based detectors and zero-shot classifiers.
Our experiments demonstrate that retrieval-based detectors, designed to evade paraphrasing attacks, are still vulnerable againstrecursiveparaphrasing. We then provide a theoreticalimpossibility resultindicating that as language models become more sophisticated and better at emulating human text, the performance of even the best-possible detector decreases.
For a sufficiently advanced language model seeking to imitate human text, even the best-possible detector may only perform marginally better than a random classifier.
Our result is general enough to capture specific scenarios such as particular writing styles, clever prompt design, or text paraphrasing.
We also extend the impossibility result to include the case wherepseudorandomnumber generators are used for AI-text generation instead of true randomness.
We show that the same result holds with a negligible correction term for all polynomial-time computable detectors.
Finally, we show that even LLMs protected by watermarking schemes can be vulnerable againstspoofing attackswhereadversarial humanscan infer hidden LLM text signatures and add them to human-generated text to be detected as text generated by the LLMs, potentially causing reputational damage to their developers. We believe these results can open an honest conversation in the community regarding the ethical and reliable use of AI-generated text. Our code is publicly available athttps://github.com/vinusankars/Reliability-of-AI-text-detectors. |
2203.00555 | In this paper, we propose a simple yet effective method to stabilize extremely deep Transformers. Specifically, we introduce a new normalization function (DeepNorm) to modify the residual connection in Transformer, accompanying with theoretically derived initialization.
In-depth theoretical analysis shows that model updates can be bounded in a stable way.
The proposed method combines the best of two worlds, i.e., good performance of Post-LN and stable training of Pre-LN, makingDeepNorma preferred alternative.
We successfully scale Transformers up to 1,000 layers (i.e., 2,500 attention and feed-forward network sublayers) without difficulty, which is one order of magnitude deeper than previous deep Transformers. Remarkably, on a multilingual benchmark with 7,482 translation directions, our 200-layer model with 3.2B parameters significantly outperforms the 48-layer state-of-the-art model with 12B parameters by 5 BLEU points, which indicates a promising scaling direction. |
2210.06423 | A big convergence of model architectures across language, vision, speech, and multimodal is emerging. However, under the same name “Transformers”, the above areas use different implementations for better performance, e.g., Post-LayerNorm for BERT, and Pre-LayerNorm for GPT and vision Transformers.
We call for the development ofFoundation Transformerfortrue general-purpose modeling, which serves as a go-to architecture for various tasks and modalities with guaranteed training stability.
In this work, we introduce a Transformer variant, namedMagneto, to fulfill the goal.
Specifically, we propose Sub-LayerNorm for good expressivity, and the initialization strategy theoretically derived from DeepNetfor stable scaling up.
Extensive experiments demonstrate its superior performance and better stability than the de facto Transformer variants designed for various applications, including language modeling (i.e., BERT, and GPT) , machine translation, vision pretraining (i.e., BEiT) , speech recognition, and multimodal pretraining (i.e., BEiT-3) . |
2212.10554 | Position modeling plays a critical role in Transformers.
In this paper, we focus on length extrapolation, i.e., training on short texts while evaluating longer sequences. We defineattention resolutionas an indicator of extrapolation. Then we propose two designs to improve the above metric of Transformers. Specifically, we introduce a relative position embedding to explicitly maximize attention resolution.
Moreover, we use blockwise causal attention during inference for better resolution. We evaluate different Transformer variants with language modeling.
Experimental results show that our model achieves strong performance in both interpolation and extrapolation settings.
The code will be available athttps://aka.ms/LeX-Transformer. |
2302.13007 | Text data augmentation is an effective strategy for overcoming the challenge of limited sample sizes in many natural language processing (NLP) tasks. This challenge is especially prominent in the few-shot learning scenario, where the data in the target domain is generally much scarcer and of lowered quality. A natural and widely-used strategy to mitigate such challenges is to perform data augmentation to better capture the data invariance and increase the sample size. However, current text data augmentation methods either can’t ensure the correct labeling of the generated data (lacking faithfulness) or can’t ensure sufficient diversity in the generated data (lacking compactness) , or both. Inspired by the recent success of large language models, especially the development of ChatGPT, which demonstrated improved language comprehension abilities, in this work, we propose a text data augmentation approach based on ChatGPT (named AugGPT) . AugGPT rephrases each sentence in the training samples into multiple conceptually similar but semantically different samples. The augmented samples can then be used in downstream model training. Experiment results on few-shot learning text classification tasks show the superior performance of the proposed AugGPT approach over state-of-the-art text data augmentation methods in terms of testing accuracy and distribution of the augmented samples. |
2302.06476 | Spurred by advancements in scale, large language models (LLMs) have demonstrated the ability to perform a variety of natural language processing (NLP) tasks zero-shot—i.e., without adaptation on downstream data. Recently, the debut of ChatGPT111https://chat.openai.com/has drawn a great deal of attention from the natural language processing (NLP) community due to the fact that it can generate high-quality responses to human input and self-correct previous mistakes based on subsequent conversations. However, it is not yet known whether ChatGPT can serve as a generalist model that can perform many NLP tasks zero-shot. In this work, we empirically analyze the zero-shot learning ability of ChatGPT by evaluating it on 20 popular NLP datasets covering 7 representative task categories. With extensive empirical studies, we demonstrate both the effectiveness and limitations of the current version of ChatGPT. We find that ChatGPT performs well on many tasks favoring reasoning capabilities (e.g.,arithmetic reasoning) while it still faces challenges when solving specific tasks such as sequence tagging. We additionally provide in-depth analysis through qualitative case studies. |
2302.00923 | Large language models (LLMs) have shown impressive performance on complex reasoning by leveraging chain-of-thought (CoT) prompting to generate intermediate reasoning chains as the rationale to infer the answer. However, existing CoT studies have focused on the language modality. We propose Multimodal-CoT that incorporates language (text) and vision (images) modalities into a two-stage framework that separates rationale generation and answer inference.
In this way, answer inference can leverage better generated rationales that are based on multimodal information.
With Multimodal-CoT, our model under 1 billion parameters outperforms the previous state-of-the-art LLM (GPT-3.5) by 16 percentage points (75.17%91.68% accuracy) and even surpasses human performance on the ScienceQA benchmark.
Code is publicly available.111https://github.com/amazon-science/mm-cot |
2303.09556 | Denoising diffusion models have been a mainstream approach for image generation, however, training these models often suffers from slow convergence. In this paper, we discovered that the slow convergence is partly due to conflicting optimization directions between timesteps. To address this issue, we treat the diffusion training as a multi-task learning problem, and introduce a simple yet effective approach referred to as Min-SNR-. This method adapts loss weights of timesteps based on clamped signal-to-noise ratios, which effectively balances the conflicts among timesteps. Our results demonstrate a significant improvement in converging speed, 3.4faster than previous weighting strategies. It is also more effective, achieving a new record FID score of 2.06 on the ImageNetbenchmark using smaller architectures than that employed in previous state-of-the-art. The code is available athttps://github.com/TiankaiHang/Min-SNR-Diffusion-Training. |
1910.10683 | Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP) .
The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice.
In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts all text-based language problems into a text-to-text format.
Our systematic study compares pre-training objectives, architectures, unlabeled data sets, transfer approaches, and other factors on dozens of language understanding tasks.
By combining the insights from our exploration with scale and our new “Colossal Clean Crawled Corpus”, we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more.
To facilitate future work on transfer learning for NLP, we release our data set, pre-trained models, and code.111https://github.com/google-research/text-to-text-transfer-transformer |
2210.11399 | Scaling language models improves performance but comes with significant computational costs.
This paper proposes UL2R, a method that substantially improves existing language models and their scaling curves with a relatively tiny amount of extra compute. The key idea is to continue training a state-of-the-art large language model (e.g., PaLM) on a few more steps with UL2’s mixture-of-denoiser objective. We show that, with almost negligible extra computational costs and no new sources of data, we are able to substantially improve the scaling properties of large language models on downstream metrics. In this paper, we continue training PaLM with UL2R, introducing a new set of models at 8B, 62B, and 540B scale which we call U-PaLM. Impressively, at 540B scale, we show an approximately 2x computational savings rate where U-PaLM achieves the same performance as the final PaLM 540B model at around half its computational budget (i.e., saving4.4 million TPUv4 hours) . We further show that this improved scaling curve leads to “emergent abilities” on challenging BIG-Bench tasks—for instance, U-PaLM does much better than PaLM on some tasks or demonstrates better quality at much smaller scale (62B as opposed to 540B) . Overall, we show that U-PaLM outperforms PaLM on many few-shot setups, i.e., English NLP tasks (e.g., commonsense reasoning, question answering) , reasoning tasks with chain-of-thought (e.g., GSM8K) , multilingual tasks (MGSM, TydiQA) , MMLU and challenging BIG-Bench tasks. Finally, we provide qualitative examples showing the new capabilities of U-PaLM for single and multi-span infilling. |
1912.02164 | Large transformer-based language models (LMs) trained on huge text corpora have shown unparalleled generation capabilities.
However, controlling attributes of the generated language (e.g. switching topic or sentiment) is difficult without modifying the model architecture or fine-tuning on attribute-specific data and entailing the significant cost of retraining.
We propose a simple alternative: the Plug and Play Language Model (PPLM) for controllable language generation, which combines a pretrained LM with one or more simple attribute classifiers that guide text generation without any further training of the LM.
In the canonical scenario we present, the attribute models are simple classifiers consisting of a user-specified bag of words or a single learned layer with 100,000 times fewer parameters than the LM.
Sampling entails a forward and backward pass in which gradients from the attribute model push the LM’s hidden activations and thus guide the generation.
Model samples demonstrate control over a range of topics and sentiment styles, and extensive automated and human annotated evaluations show attribute alignment and fluency.
PPLMs are flexible in that any combination of differentiable attribute models may be used to steer text generation, which will allow for diverse and creative applications beyond the examples given in this paper. |
2009.06367 | While large-scale language models (LMs) are able to imitate the distribution of natural language well enough to generate realistic text, it is difficult to control which regions of the distribution they generate. This is especially problematic because datasets used for training large LMs usually contain significant toxicity, hate, bias, and negativity. We propose GeDi as an efficient method for using smaller LMs as generative discriminators to guide generation from large LMs to make them safer and more controllable. GeDi guides generation at each step by computing classification probabilities for all possible next tokens via Bayes rule by normalizing over two class-conditional distributions; one conditioned on the desired attribute, orcontrol code, and another conditioned on the undesired attribute, oranti control code. We find that GeDi gives stronger controllability than the state of the art method while also achieving generation speeds more thantimes faster. Additionally, training GeDi on only four topics allows us to controllably generate new topics zero-shot from just a keyword, unlocking a new capability that previous controllable generation methods do not have. Lastly, we show that GeDi can make GPT-2 (1.5B parameters) significantly less toxic without sacrificing linguistic quality, making it by far the most practical existing method for detoxifying large language models while maintaining a fast generation speed.111Code available athttps://github.com/salesforce/GeDi, includes GeDi-guided GPT-3 generation using OpenAI API. |
2210.11610 | Large Language Models (LLMs) have achieved excellent performances in various tasks. However, fine-tuning an LLM requires extensive supervision. Human, on the other hand, may improve their reasoning abilities by self-thinking without external inputs. In this work, we demonstrate that an LLM is also capable of self-improving with only unlabeled datasets.
We use a pre-trained LLM to generate “high-confidence” rationale-augmented answers for unlabeled questions using Chain-of-Thought prompting and self-consistency, and fine-tune the LLM using those self-generated solutions as target outputs. We show that our approach improves the general reasoning ability of a 540B-parameter LLM (74.4%82.1% on GSM8K, 78.2%83.0% on DROP, 90.0%94.4% on OpenBookQA, and 63.4%67.9% on ANLI-A3) and achieves state-of-the-art-level performance, without any ground truth label. We conduct ablation studies and show that fine-tuning on reasoning is critical for self-improvement. |
2106.09685 | An important paradigm of natural language processing consists of large-scale pre-training on general domain data and adaptation to particular tasks or domains.
As we pre-train larger models, full fine-tuning, which retrains all model parameters, becomes less feasible.
Using GPT-3 175B as an example – deploying independent instances of fine-tuned models, each with 175B parameters, is prohibitively expensive.
We proposeLow-RankAdaptation, or LoRA, which freezes the pre-trained model weights and injects trainable rank decomposition matrices into each layer of the Transformer architecture, greatly reducing the number of trainable parameters for downstream tasks.
Compared to GPT-3 175B fine-tuned with Adam, LoRA can reduce the number of trainable parameters by 10,000 times and the GPU memory requirement by 3 times.
LoRA performs on-par or better than fine-tuning in model quality on RoBERTa, DeBERTa, GPT-2, and GPT-3, despite having fewer trainable parameters, a higher training throughput, and, unlike adapters,no additional inference latency.
We also provide an empirical investigation into rank-deficiency in language model adaptation, which sheds light on the efficacy of LoRA.
We release a package that facilitates the integration of LoRA with PyTorch models and provide our implementations and model checkpoints for RoBERTa, DeBERTa, and GPT-2 athttps://github.com/microsoft/LoRA. |
2105.14103 | We introduce Attention Free Transformer (AFT) , an efficient variant of Transformersthat eliminates the need for dot product self attention. In an AFT layer, the key and value are first combined with a set of learned position biases, the result of which is multiplied with the query in an element-wise fashion. This new operation has a memory complexity linear w.r.t. both the context size and the dimension of features, making it compatible to both large input and model sizes. We also introduce AFT-local and AFT-conv, two model variants that take advantage of the idea of locality and spatial weight sharing while maintaining global connectivity. We conduct extensive experiments on two autoregressive modeling tasks (CIFAR10 and Enwik8) as well as an image recognition task (ImageNet-1K classification) . We show that AFT demonstrates competitive performance on all the benchmarks, while providing excellent efficiency at the same time. |
2406.03880 | Deep Learning (DL) powered by Deep Neural Networks (DNNs) has revolutionized various domains,
yet understanding the intricacies of DNN decision-making and learning processes remains a significant challenge.
Recent investigations have uncovered an interesting memorization phenomenon in which DNNs tend to memorize specific details from examples rather than learning general patterns,
affecting model generalization, security, and privacy.
This raises critical questions about the nature of generalization in DNNs and their susceptibility to security breaches.
In this survey,
we present a systematic framework to organize memorization definitions based on the generalization and security/privacy domains and summarize memorization evaluation methods at both the example and model levels.
Through a comprehensive literature review,
we explore DNN memorization behaviors and their impacts on security and privacy.
We also introduce privacy vulnerabilities caused by memorization
and the phenomenon of forgetting and explore its connection with memorization.
Furthermore,
we spotlight various applications leveraging memorization and forgetting mechanisms, including noisy label learning, privacy preservation, and model enhancement.
This survey offers the first-in-kind understanding of memorization in DNNs,
providing insights into its challenges and opportunities for enhancing AI development while addressing critical ethical concerns. |
1904.12848v6 | Semi-supervised learning lately has shown much promise in improving deep learning models when labeled data is scarce. Common among recent approaches is the use of consistency training on a large amount of unlabeled data to constrain model predictions to be invariant to input noise. In this work, we present a new perspective on how to effectively noise unlabeled examples and argue that the quality of noising, specifically those produced by advanced data augmentation methods, plays a crucial role in semi-supervised learning. By substituting simple noising operations with advanced data augmentation methods such as RandAugment and back-translation, our method brings substantial improvements across six language and three vision tasks under the same consistency training framework. On the IMDb text classification dataset, with only 20 labeled examples, our method achieves an error rate of 4.20, outperforming the state-of-the-art model trained on 25,000 labeled examples. On a standard semi-supervised learning benchmark, CIFAR-10, our method outperforms all previous approaches and achieves an error rate of 5.43 with only 250 examples. Our method also combines well with transfer learning, e.g., when finetuning from BERT, and yields improvements in high-data regime, such as ImageNet, whether when there is only 10% labeled data or when a full labeled set with 1.3M extra unlabeled examples is used.111Code is available athttps://github.com/google-research/uda. |
2403.05045 | This study analyzes changes in the attention mechanisms of large language models (LLMs) when used to understand natural conversations between humans (human-human) . We analyze three use cases of LLMs: interactions over web content, code, and mathematical texts. By analyzing attention distance, dispersion, and interdependency across these domains, we highlight the unique challenges posed by conversational data. Notably, conversations require nuanced handling of long-term contextual relationships and exhibit higher complexity through their attention patterns. Our findings reveal that while language models exhibit domain-specific attention behaviors, there is a significant gap in their ability to specialize in human conversations. Through detailed attention entropy analysis and t-SNE visualizations, we demonstrate the need for models trained with a diverse array of high-quality conversational data to enhance understanding and generation of human-like dialogue. This research highlights the importance of domain specialization in language models and suggests pathways for future advancement in modeling human conversational nuances. |
2105.05720v5 | Recent trends towards large machine learning models require both training and inference tasks to be distributed.
Considering the huge cost of training these models, it is imperative to unlock optimizations in computation and communication to obtain best performance.
However, the current logical separation between computation and communication kernels in machine learning
frameworks misses optimization opportunities across this barrier.
Breaking this abstraction can provide many optimizations to improve the performance
of distributed workloads.
However, manually applying these optimizations requires modifying the underlying computation and communication libraries for each scenario, which is both time consuming and error-prone. Therefore, we presentCoCoNet, which
contains (i) a domain specific language to express a distributed machine learning program in the form of computation and communication operations, (ii) a set of semantics preserving transformations to optimize the program, and (iii) a compiler to generate jointly optimized communication and computation GPU kernels.
Providing both computation and communication as first class constructs allows users to work on a high-level abstraction and
apply powerful optimizations, such as fusion or overlapping of communication and computation.CoCoNetenabled us to optimize data-, model- and pipeline-parallel workloads in large language models with only a few lines of code.
Our experiments show thatCoCoNetsignificantly outperforms state-of-the-art distributed machine learning implementations. |
2406.01297 | Self-correction is an approach to improving responses from large language models (LLMs) by refining the responses using LLMs during inference. Prior work has proposed various self-correction frameworks using different sources of feedback, including self-evaluation and external feedback. However, there is still no consensus on the question ofwhen LLMs can correct their own mistakes, as recent studies also report negative results. In this work, we critically survey broad papers and discuss the conditions required for successful self-correction. We first find that prior studies often do not define their research questions in detail and involve impractical frameworks or unfair evaluations that over-evaluate self-correction. To tackle these issues, we categorize research questions in self-correction research and provide a checklist for designing appropriate experiments. Our critical survey based on the newly categorized research questions shows that (1) no prior work demonstrates successful self-correction with feedback from prompted LLMs in general tasks, (2) self-correction works well in tasks that can use reliable external feedback, and (3) large-scale fine-tuning enables self-correction. |
2402.05808 | In this paper, we proposeR3: LearningReasoning throughReverse CurriculumReinforcement Learning (RL) , a novel method that employs only outcome supervision to achieve the benefits of process supervision for large language models. The core challenge in applying RL to complex reasoning is to identify a sequence of actions that result in positive rewards and provide appropriate supervision for optimization. Outcome supervision provides sparse rewards for final results without identifying error locations, whereas process supervision offers step-wise rewards but requires extensive manual annotation.R3overcomes these limitations by learning from correct demonstrations. Specifically,R3progressively slides the start state of reasoning from a demonstration’s end to its beginning, facilitating easier model exploration at all stages. Thus,R3establishes a step-wise curriculum, allowing outcome supervision to offer step-level signals and precisely pinpoint errors. Using Llama2-7B, our method surpasses RL baseline on eight reasoning tasks bypoints on average. Notebaly, in program-based reasoning on GSM8K, it exceeds the baseline bypoints across three backbone models, and without any extra data, Codellama-7B +R3performs comparable to larger models or closed-source models.111Our codes and data are available at Github :https://github.com/WooooDyy/LLM-Reverse-Curriculum-RL. |
2407.11062 | Large language models (LLMs) are integral to modern natural language processing and artificial intelligence. However, they face challenges in managing their significant memory requirements. Although quantization-aware training (QAT) offers a solution by reducing memory consumption through low-bit representations with minimal accuracy loss, it demands substantial training resources to optimize model weights and quantization parameters.
To address this, we propose Efficient Quantization-Aware Training (EfficientQAT) , a novel quantization technique for compressing LLMs. EfficientQAT involves two consecutive phases: Block-wise training of all parameters (Block-AP) and end-to-end training of quantization parameters (E2E-QP) . Block-AP sequentially conducts quantization-aware training for all parameters in each transformer block with block-wise reconstruction, maintaining efficiency by avoiding training the entire LLM. Initialized with quantized model, E2E-QP then trains only quantization parameters (step sizes) end-to-end, enhancing efficiency with a fixed quantized backbone and reduced trainable parameter count.
Extensive experiments demonstrate that EfficientQAT outperforms previous quantization methods across a range of models, including base LLMs, instruction-tuned LLMs, and multimodal LLMs, with scales from 7B to 70B parameters at various quantization bits.
For instance, EfficientQAT obtains a 2-bit Llama-2-70B model on a single A100-80GB GPU in 41 hours, with less than 3% accuracy degradation compared to the full precision (69.48 vs. 72.41) . Notably, this INT2 quantized 70B model obtains a 1.67 accuracy gain over the Llama-2-13B model (69.48 vs. 67.81) while requiring less memory (19.2GB vs. 24.2GB) . Code is available athttps://github.com/OpenGVLab/EfficientQAT. |
2108.06084 | Recent works have demonstrated great success in pre-training large-scale autoregressive language models (e.g., GPT-3) on massive GPUs. To reduce the wall-clock training time, a common practice is to increase the batch size and learning rate. However, such practice is often brittle and leads to a so-called stability-efficiency dilemma: increasing the batch sizes and learning rates leads to better training efficiency but can also result in training instability, leading to poor generalization accuracy or failed runs. To better understand this phenomenon, we conduct an in-depth analysis on large-scale pre-training experiments replicating the GPT-2 model with public dataset. We find that there is a strong correlation between training instability and extreme values of gradient variance. We further identify that samples with long sequence lengths contribute to these extreme gradient variance values, especially at the beginning of the training, indicating that long sequence length can be a main source of training instability. Based on the analysis, we present a simple yet effective Sequence Length Warmup method that aims to solve the training stability-efficiency dilemma by avoiding extreme gradient variance values. Moreover, we present a lightweight tuning strategy that allows us to tune our method with just a small portion of the expensive full training. Experiments replicating GPT-2 models (117M and 1.5B) show that our approach enables stable training with 8x larger batch size and 4x larger learning rate, whereas the baseline approach struggles with training instability. To achieve the same or better zero-shot evaluation results, our method reduces the required number of training tokens and wall clock time by up to 2.2x and 3.7x, respectively. Experiments replicating GPT-3 model (125M) show that our approach enables stable training with 8x larger batch size and 40x larger learning rate, and retains 99% of the zero-shot accuracy on 11 tasks using 10x less data and 17x less time compared to the original GPT-3 training recipe, while the baseline diverges under the same settings and only retain 95% of accuracy under lower learning rate. |
2312.08935 | In this paper, we present an innovative process-oriented math process reward model called\methodname, which assigns a reward score to each step of math problem solutions.
The training of\methodnameis achieved using automatically constructed process-wise supervision data, breaking the bottleneck of heavy reliance on manual annotation in existing work.
We explore the effectiveness of\methodnamein two scenarios:
1) Verification:\methodnameis utilized for reranking multiple outputs generated by Large Language Models (LLMs) ;
2) Reinforcement Learning:\methodnameis employed to reinforce LLMs with step-by-step Proximal Policy Optimization (PPO) .
With\methodname, a series of open-source LLMs demonstrates exceptional performance.
For instance, the step-by-step PPO with\methodnamesignificantly improves the accuracy of Mistral-7B (77.9%84.1% on GSM8K and 28.6%33.0% on MATH) .
The accuracy can be further enhanced to 89.1% and 43.5% on GSM8K and MATH with the verification of\methodname, respectively.
We believe that automatic process supervision holds significant potential for the future evolution of LLMs. |
2309.07314 | Audio super-resolution is a fundamental task that predicts high-frequency components for low-resolution audio, enhancing audio quality in digital applications. Previous methods have limitations such as the limited scope of audio types (e.g., music, speech) and specific bandwidth settings they can handle (e.g.,kHz tokHz) . In this paper, we introduce a diffusion-based generative model,AudioSR, that is capable of performing robust audio super-resolution on versatile audio types, including sound effects, music, and speech. Specifically,AudioSRcan upsample any input audio signal within the bandwidth range ofkHz tokHz to a high-resolution audio signal atkHz bandwidth with a sampling rate ofkHz. Extensive objective evaluation on various audio super-resolution benchmarks demonstrates the strong result achieved by the proposed model. In addition, our subjective evaluation shows thatAudioSRcan acts as a plug-and-play module to enhance the generation quality of a wide range of audio generative models, including AudioLDM, Fastspeech2, and MusicGen. Our code and demo are available athttps://audioldm.github.io/audiosr. |
2406.10774 | As the demand for long-context large language models (LLMs) increases, models with context windows of up to 128K or 1M tokens are becoming increasingly prevalent. However, long-context LLM inference is challenging since the inference speed decreases significantly as the sequence length grows. This slowdown is primarily caused by loading a large KV cache during self-attention.
Previous works have shown that a small portion of critical tokens will dominate the attention outcomes. However, we observe the criticality of a token highly depends on the query. To this end, we propose Quest, a query-aware KV cache selection algorithm. Quest keeps track of the minimal and maximal Key values in KV cache pages and estimates the criticality of a given page using Query vectors. By only loading the Top-K critical KV cache pages for attention, Quest significantly speeds up self-attention without sacrificing accuracy. We show that Quest can achieve up toself-attention speedup, which reduces inference latency bywhile performing well on tasks with long dependencies with negligible accuracy loss. Code is available athttps://github.com/mit-han-lab/Quest. |
2402.01912 | Text-to-speech models trained on large-scale datasets have demonstrated impressive in-context learning capabilities and naturalness. However, control of speaker identity and style in these models typically requires conditioning on reference speech recordings, limiting creative applications. Alternatively, natural language prompting of speaker identity and style has demonstrated promising results and provides an intuitive method of control. However, reliance on human-labeled descriptions prevents scaling to large datasets. Our work bridges the gap between these two approaches. We propose a scalable method for labeling various aspects of speaker identity, style, and recording conditions. We then apply this method to a 45k hour dataset, which we use to train a speech language model. Furthermore, we propose simple methods for increasing audio fidelity, significantly outperforming recent work despite relying entirely on found data. Our results demonstrate high-fidelity speech generation in a diverse range of accents, prosodic styles, channel conditions, and acoustic conditions, all accomplished with a single model and intuitive natural language conditioning. Audio samples can be heard athttps://
text-description-to-speech.com/. |
2402.04825 | Generating long-form 44.1kHz stereo audio from text prompts can be computationally demanding. Further, most previous works do not tackle that music and sound effects naturally vary in their duration.
Our research focuses on the efficient generation of long-form, variable-length stereo music and sounds at 44.1kHz using text prompts with a generative model.
Stable Audio is based on latent diffusion, with its latent defined by a fully-convolutional variational autoencoder. It is conditioned on text prompts as well as timing embeddings, allowing for fine control over both the content and length of the generated music and sounds.
Stable Audio is capable of rendering stereo signals of up to 95 sec at 44.1kHz in 8 sec on an A100 GPU.
Despite its compute efficiency and fast inference, it is one of the best in two public text-to-music and -audio benchmarks and, differently from state-of-the-art models, can generate music with structure and stereo sounds. |
1912.07875 | We introduce a new collection of spoken English audio suitable for training speech recognition systems under limited or no supervision. It is derived from open-source audio books from the LibriVox project. It contains over 60K hours of audio, which is, to our knowledge, the largest freely-available corpus of speech. The audio has been segmented using voice activity detection and is tagged with SNR, speaker ID and genre descriptions. Additionally, we provide baseline systems and evaluation metrics working under three settings: (1) the zero resource/unsupervised setting (ABX) , (2) the semi-supervised setting (PER, CER) and (3) the distant supervision setting (WER) . Settings (2) and (3) use limited textual resources (10 minutes to 10 hours) aligned with the speech. Setting (3) uses large amounts of unaligned text. They are evaluated on the standard LibriSpeech dev and test sets for comparison with the supervised state-of-the-art. |
2212.11851 | Diffusion models have shown a great ability at bridging the performance gap between predictive and generative approaches for speech enhancement. We have shown that they may even outperform their predictive counterparts for non-additive corruption types or when they are evaluated on mismatched conditions. However, diffusion models suffer from a high computational burden, mainly as they require to run a neural network for each reverse diffusion step, whereas predictive approaches only require one pass. As diffusion models are generative approaches they may also produce vocalizing and breathing artifacts in adverse conditions. In comparison, in such difficult scenarios, predictive models typically do not produce such artifacts but tend to distort the target speech instead, thereby degrading the speech quality. In this work, we present astochastic regenerationapproach where an estimate given by a predictive model is provided as a guide for further diffusion. We show that the proposed approach uses the predictive model to remove the vocalizing and breathing artifacts while producing very high quality samples thanks to the diffusion model, even in adverse conditions. We further show that this approach enables to use lighter sampling schemes with fewer diffusion steps without sacrificing quality, thus lifting the computational burden by an order of magnitude. Source code and audio examples are available online111https://uhh.de/inf-sp-storm. |
2302.14748 | Recently, score-based generative models have been successfully employed for the task of speech enhancement. A stochastic differential equation is used to model the iterative forward process, where at each step environmental noise and white Gaussian noise are added to the clean speech signal. While in limit the mean of the forward process ends at the noisy mixture, in practice it stops earlier and thus only at an approximation of the noisy mixture. This results in a discrepancy between the terminating distribution of the forward process and the prior used for solving the reverse process at inference. In this paper, we address this discrepancy and propose a forward process based on a Brownian bridge. We show that such a process leads to a reduction of the mismatch compared to previous diffusion processes. More importantly, we show that our approach improves in objective metrics over the baseline process with only half of the iteration steps and having one hyperparameter less to tune. |
2402.01808 | This paper presents the speech restoration and enhancement system created by the 1024K team for the ICASSP 2024 Speech Signal Improvement (SSI) Challenge. Our system consists of a generative adversarial network (GAN) in complex-domain for speech restoration and a fine-grained multi-band fusion module for speech enhancement. In the blind test set of SSI, the proposed system achieves an overall mean opinion score (MOS) of 3.49 based on ITU-T P.804 and a Word Accuracy Rate (WAcc) of 0.78 for the real-time track, as well as an overall P.804 MOS of 3.43 and a WAcc of 0.78 for the non-real-time track, rankingin both tracks. |
2303.01664 | Speech restoration (SR) is a task of converting degraded speech signals into high-quality ones. In this study, we propose a robust SR model calledMiipher, and apply Miipher to a new SR application: increasing the amount of high-quality training data for speech generation by converting speech samples collected from the Web to studio-quality. To make our SR model robust against various degradation, we use (i) a speech representation extracted from w2v-BERT for the input feature, and (ii) a text representation extracted from transcripts via PnG-BERT as a linguistic conditioning feature. Experiments show that Miipher (i) is robust against various audio degradation and (ii) enable us to train a high-quality text-to-speech (TTS) model from restored speech samples collected from the Web.
Audio samples are available at our demo page:google.github.io/df-conformer/miipher/. |
2202.13669 | Structured document understanding has attracted considerable attention and made significant progress recently, owing to its crucial role in intelligent document processing. However, most existing related models can only deal with the document data of specific language(s) (typically English) included in the pre-training collection, which is extremely limited. To address this issue, we propose a simple yet effectiveLanguage-independentLayoutTransformer (LiLT) for structured
document understanding. LiLT can be pre-trained on the structured documents of a single language and then directly fine-tuned on other languages with the corresponding off-the-shelf monolingual/multilingual pre-trained textual models. Experimental results on eight languages have shown that LiLT can achieve competitive or even superior performance on diverse widely-used downstream benchmarks, which enables language-independent benefit from the pre-training of document layout structure. Code and model
are publicly available athttps://github.com/jpWang/LiLT. |
2406.17565 | Large language model (LLM) serving has transformed from stateless to stateful systems, utilizing techniques like context caching and disaggregated inference.
These optimizations extend the lifespan and domain of the KV cache, necessitating a new architectural approach.
We present MemServe, a unified system that integrates both inter-request and intra-request optimizations.
MemServe introduces MemPool, an elastic memory pool managing distributed memory and KV caches across serving instances.
Using MemPool APIs, MemServe combines context caching with disaggregated inference for the first time, supported by a global scheduler that enhances cache reuse through a global prompt tree-based locality-aware policy.
Tests show that MemServe significantly improves job completion time and time-to-first-token. |
2401.09670 | DistServe improves the performance of large language models (LLMs) serving by disaggregating the prefill and decoding computation. Existing LLM serving systems colocate the two phases and batch the computation of prefill and decoding across all users and requests.
We find that this strategy not only leads to strong prefill-decoding interferences but also couples the resource allocation and parallelism plans for both phases. LLM applications often emphasize individual latency for each phase: time to first token (TTFT) for the prefill phase and time per output token (TPOT) of each request for the decoding phase.
In the presence of stringent latency requirements, existing systems have to prioritize one latency over the other, or over-provision compute resources to meet both. DistServe assigns prefill and decoding computation to different GPUs, hence eliminating prefill-decoding interferences. Given the application’s TTFT and TPOT requirements, DistServe co-optimizes the resource allocation and parallelism strategytailoredfor each phase. DistServe also places the two phases according to the serving cluster’s bandwidth to minimize the communication caused by disaggregation. As a result, DistServe significantly improves LLM serving performance in terms of the maximum rate that can be served within both TTFT and TPOT constraints on each GPU.
Our evaluations show that on various popular LLMs, applications, and latency requirements, DistServe can serve 4.48more requests or 10.2tighter SLO, compared to state-of-the-art systems, while staying within latency constraints forof requests. |
2405.19715v2 | Speculative decoding reduces the inference latency of a target large language model via utilizing a smaller and faster draft model. Its performance depends on a hyperparameter— the candidate length, i.e., the number of candidate tokens for the target model to verify in each round.
However, previous methods often use simple heuristics to choose, which may result in sub-optimal performance.
We study the choice of the candidate lengthand formulate it as a Markov Decision Process. We theoretically show that the optimal policy of this Markov decision process takes the form of a threshold policy, i.e., the current speculation should stop and be verified when the probability of getting a rejection exceeds a threshold value.
Motivated by this theory, we proposeSpecDec++, an enhanced version of speculative decoding that adaptively determines the candidate length on the fly. We augment the draft model with a trained acceptance prediction head to predict the conditional acceptance probability of the candidate tokens.SpecDec++will stop the current speculation when the predicted probability thatat least one token gets rejectedexceeds a threshold.
We implementSpecDec++and apply it to the llama-2-chat 7B & 70B model pair.
Our adaptive method achieves a 2.04x speedup on the Alpaca dataset (an additional 7.2% improvement over the baseline speculative decoding) . On the GSM8K and HumanEval datasets, our method achieves a 2.26x speedup (9.4% improvement) and 2.23x speedup (11.1% improvement) , respectively. |
2407.01502v1 | AI agents are an exciting new research direction, and agent development is driven by benchmarks. Our analysis of current agent benchmarks and evaluation practices reveals several shortcomings that hinder their usefulness in real-world applications. First, there is a narrow focus on accuracy without attention to other metrics. As a result, SOTA agents are needlessly complex and costly, and the community has reached mistaken conclusions about the sources of accuracy gains. Our focus on cost in addition to accuracy motivates the new goal of jointly optimizing the two metrics. We design and implement one such optimization, showing its potential to greatly reduce cost while maintaining accuracy. Second, the benchmarking needs of model and downstream developers have been conflated, making it hard to identify which agent would be best suited for a particular application. Third, many agent benchmarks have inadequate holdout sets, and sometimes none at all. This has led to agents that are fragile because they take shortcuts and overfit to the benchmark in various ways. We prescribe a principled framework for avoiding overfitting. Finally, there is a lack of standardization in evaluation practices, leading to a pervasive lack of reproducibility. We hope that the steps we introduce for addressing these shortcomings will spur the development of agents that are useful in the real world and not just accurate on benchmarks. |
2409.18486 | This comprehensive study evaluates the performance of OpenAI’s o1-preview large language model across a diverse array of complex reasoning tasks, spanning multiple domains, including computer science, mathematics, natural sciences, medicine, linguistics, and social sciences. Through rigorous testing, o1-preview demonstrated remarkable capabilities, often achieving human-level or superior performance in areas ranging from coding challenges to scientific reasoning and from language processing to creative problem-solving. Key findings include: 83.3% success rate in solving complex competitive programming problems, surpassing many human experts. Superior ability in generating coherent and accurate radiology reports, outperforming other evaluated models. 100% accuracy in high school-level mathematical reasoning tasks, providing detailed step-by-step solutions. Advanced natural language inference capabilities across general and specialized domains like medicine. Impressive performance in chip design tasks, outperforming specialized models in areas such as EDA script generation and bug analysis. Remarkable proficiency in anthropology and geology, demonstrating deep understanding and reasoning in these specialized fields. Strong capabilities in quantitative investing. O1 has comprehensive financial knowledge and statistical modeling skills. Effective performance in social media analysis, including sentiment analysis and emotion recognition. The model excelled particularly in tasks requiring intricate reasoning and knowledge integration across various fields. While some limitations were observed, including occasional errors on simpler problems and challenges with certain highly specialized concepts, the overall results indicate significant progress towards artificial general intelligence. This evaluation not only highlights o1-preview’s current strengths and limitations but also identifies crucial areas for future development, including multi-modal integration, domain-specific validation, and ethical considerations for real-world applications. The findings provide valuable insights into the potential of large language models in numerous fields and pave the way for further advancements in AI research and application. |
2409.04701 | Many use cases require retrieving smaller portions of text, and dense vector-based retrieval systems often perform better with shorter text segments, as the semantics are less likely to be “over-compressed” in the embeddings.
Consequently, practitioners often split text documents into smaller chunks and encode them separately. However, chunk embeddings created in this way can lose contextual information from surrounding chunks, resulting in sub-optimal representations.
In this paper, we introduce a novel method called “late chunking”, which leverages long context embedding models to first embed all tokens of the long text, with chunking appliedafterthe transformer model and just before mean pooling - hence the term “late” in its naming.
The resulting chunk embeddings capture the full contextual information, leading to superior results across various retrieval tasks.
The method is generic enough to be applied to a wide range of long-context embedding models and works without additional training.
To further increase the effectiveness of late chunking, we propose a dedicated fine-tuning approach for embedding models. |
2406.13457 | Event-based vision has drawn increasing attention due to its unique characteristics, such as high temporal resolution and high dynamic range. It has been used in video super-resolution (VSR) recently to enhance the flow estimation and temporal alignment. Rather than for motion learning, we propose in this paper the first VSR method that utilizes event signals for texture enhancement. Our method, called EvTexture, leverages high-frequency details of events to better recover texture regions in VSR. In our EvTexture, a new texture enhancement branch is presented. We further introduce an iterative texture enhancement module to progressively explore the high-temporal-resolution event information for texture restoration. This allows for gradual refinement of texture regions across multiple iterations, leading to more accurate and rich high-resolution details. Experimental results show that our EvTexture achieves state-of-the-art performance on four datasets. For the Vid4 dataset with rich textures, our method can get up to 4.67dB gain compared with recent event-based methods. Code:https://github.com/DachunKai/EvTexture. |
2201.12288 | Video restoration (e.g., video super-resolution) aims to restore high-quality frames from low-quality frames. Different from single image restoration, video restoration generally requires to utilize temporal information from multiple adjacent but usually misaligned video frames. Existing deep methods generally tackle with this by exploiting a sliding window strategy or a recurrent architecture, which either is restricted by frame-by-frame restoration or lacks long-range modelling ability. In this paper, we propose a Video Restoration Transformer (VRT) with parallel frame prediction and long-range temporal dependency modelling abilities. More specifically, VRT is composed of multiple scales, each of which consists of two kinds of modules: temporal mutual self attention (TMSA) and parallel warping. TMSA divides the video into small clips, on which mutual attention is applied for joint motion estimation, feature alignment and feature fusion, while self attention is used for feature extraction. To enable cross-clip interactions, the video sequence is shifted for every other layer. Besides, parallel warping is used to further fuse information from neighboring frames by parallel feature warping. Experimental results on five tasks, including video super-resolution, video deblurring, video denoising, video frame interpolation and space-time video super-resolution, demonstrate that VRT outperforms the state-of-the-art methods by large margins (up to 2.16dB) on fourteen benchmark datasets. |
2201.12288v2 | Video restoration (e.g., video super-resolution) aims to restore high-quality frames from low-quality frames. Different from single image restoration, video restoration generally requires to utilize temporal information from multiple adjacent but usually misaligned video frames. Existing deep methods generally tackle with this by exploiting a sliding window strategy or a recurrent architecture, which either is restricted by frame-by-frame restoration or lacks long-range modelling ability. In this paper, we propose a Video Restoration Transformer (VRT) with parallel frame prediction and long-range temporal dependency modelling abilities. More specifically, VRT is composed of multiple scales, each of which consists of two kinds of modules: temporal mutual self attention (TMSA) and parallel warping. TMSA divides the video into small clips, on which mutual attention is applied for joint motion estimation, feature alignment and feature fusion, while self attention is used for feature extraction. To enable cross-clip interactions, the video sequence is shifted for every other layer. Besides, parallel warping is used to further fuse information from neighboring frames by parallel feature warping. Experimental results on five tasks, including video super-resolution, video deblurring, video denoising, video frame interpolation and space-time video super-resolution, demonstrate that VRT outperforms the state-of-the-art methods by large margins (up to 2.16dB) on fourteen benchmark datasets. |
2408.04820 | We propose usingnatural language outlinesas a novel modality and interaction surface for providing AI assistance to developers throughout the software development process. An NL outline for a code function comprises multiple statements written in concise prose, which partition the code and summarize its main ideas in the style of literate programming. Crucially, we find that modern LLMs can generate accurate and high-quality NL outlines in practice. Moreover, NL outlines enable a bidirectional sync between code and NL, allowing changes in one to be automatically reflected in the other. We discuss many use cases for NL outlines: they can accelerate understanding and navigation of code and diffs, simplify code maintenance, augment code search, steer code generation, and more. We then propose and compare multiple LLM prompting techniques for generating outlines and ask professional developers to judge outline quality. Finally, we present two case studies applying NL outlines toward code review and the difficult task of malware detection. |
2401.17686 | Recent advancements have significantly augmented the reasoning capabilities of Large Language Models (LLMs) through various methodologies, especially chain-of-thought (CoT) reasoning.
However, previous methods fail to address reasoning errors in intermediate steps, leading to accumulative errors.
In this paper, we proposeDeductive Beam Search(DBS) , which seamlessly integrates CoT and deductive reasoning with step-wise beam search for LLMs.
Our approach deploys a verifier, verifying the deducibility of a reasoning step and its premises, thus alleviating the error accumulation.
Furthermore, we introduce a scalable and labor-free data construction method to amplify our model’s verification capabilities.
Extensive experiments demonstrate that our approach significantly enhances the base performance of LLMs of various scales (7B, 13B, 70B, and ChatGPT) across 8 reasoning datasets from 3 diverse reasoning genres, including arithmetic, commonsense, and symbolic.
Moreover, our analysis proves DBS’s capability of detecting diverse and subtle reasoning errors and robustness on different model scales111Code is available athttps://github.com/OSU-NLP-Group/Deductive-Beam-Search.. |
2409.20325 | Deep learning optimizers are often motivated through a mix of convex and approximate second-order theory. We select three such methods—Adam, Shampoo and Prodigy—and argue that each method can instead be understood as a squarely first-order method without convexity assumptions. In fact, after switching off exponential moving averages, each method is equivalent tosteepest descentunder a particularnorm. By generalizing this observation, we chart a new design space for training algorithms. Different operator norms should be assigned to different tensors based on the role that the tensor plays within the network. For example, while linear and embedding layers may have the same weight space of, these layers play different roles and should be assigned different norms. We hope that this idea of carefully metrizing the neural architecture might lead to more stable, scalable and indeed faster training. |
2409.04057v1 | Chain-of-thought (CoT) prompting reveals that large language models are capable of performing complex reasoning via intermediate steps.
CoT prompting is mainly categorized into two paradigms.
The first approach utilizes straightforward prompts like “Let’s think step by step” to generate a sequential thought process before yielding an answer.
The second approach makes use of human-crafted, step-by-step demonstrations to guide the model’s reasoning process.
The latter method replaces manually written demonstrations with ones generated by large language models using “Let’s think step by step”.
This approach sometimes leads to reasoning errors, highlighting the need to diversify demonstrations to mitigate its misleading effects.
However, diverse demonstrations pose challenges for effective representations.
In this work, we propose ECHO, a self-harmonized chain-of-thought prompting method.
Consolidating diverse solution paths into a uniform and effective solution pattern,
ECHO demonstrates the best overall performance in three reasoning domains111Our code is available:https://github.com/Xalp/ECHO. |
2409.12517 | We train, for the first time, large language models using FP8 precision on datasets up to 2 trillion tokens — a 20-fold increase over previous limits. Through these extended training runs, we uncover critical instabilities in FP8 training that were not observable in earlier works with shorter durations. We trace these instabilities to outlier amplification by the SwiGLU activation function. Interestingly, we show, both analytically and empirically, that this amplification happens only over prolonged training periods, and link it to a SwiGLU weight alignment process. To address this newly identified issue, we introduce Smooth-SwiGLU, a novel modification that ensures stable FP8 training without altering function behavior. We also demonstrate, for the first time, FP8 quantization of both Adam optimizer moments. Combining these innovations, we successfully train a 7B parameter model using FP8 precision on 256 Intel Gaudi2 accelerators, achieving on-par results with the BF16 baseline while delivering up to athroughput improvement. |
2409.09916 | Retrieval Augmented Generation (RAG) , a paradigm that integrates external contextual information with large language models (LLMs) to enhance factual accuracy and relevance, has emerged as a pivotal area in generative AI. The LLMs used in RAG applications are required to faithfully and completely comprehend the provided context and users’ questions, avoid hallucination, handle unanswerable, counterfactual or otherwise low-quality and irrelevant contexts, perform complex multi-hop reasoning and produce reliable citations. In this paper, we introduce\ourmodel, a small LLM that is instruction-tuned with an emphasis on context-grounded generation and hallucination minimization. We also present ContextualBench, a new evaluation framework compiling multiple popular and diverse RAG benchmarks, such as HotpotQA and TriviaQA, with consistent RAG settings to ensure reproducibility and consistency in model assessments. Experimental results demonstrate that our\ourmodel-9B model outperforms leading baselines such as Command-R+ (104B) and GPT-4o, achieving state-of-the-art results in 3 out of 7 benchmarks in ContextualBench with significantly fewer parameters. The model is also shown to be resilient to alteration in the contextual information and behave appropriately when relevant context is removed. Additionally, the\ourmodelmodel maintains competitive performance in general instruction-following tasks and function-calling capabilities. |
2409.14160 | With the growing attention and investment in recent AI approaches such as large language models, the narrative that the larger the AI system the more valuable, powerful and interesting it is is increasingly seen as common sense. But what is this assumption based on, and how are we measuring value, power, and performance? And what are the collateral consequences of this race to ever-increasing scale? Here, we scrutinize the current scaling trends and trade-offs across multiple axes and refute two common assumptions underlying the ‘bigger-is-better’ AI paradigm: 1) that improved performance is a product of increased scale, and 2) that all interesting problems addressed by AI require large-scale models. Rather, we argue that this approach is not only fragile scientifically, but comes with undesirable consequences. First, it is not sustainable, as its compute demands increase faster than model performance, leading to unreasonable economic requirements and a disproportionate environmental footprint. Second, it implies focusing on certain problems at the expense of others, leaving aside important applications, e.g. health, education, or the climate. Finally, it exacerbates a concentration of power, which centralizes decision-making in the hands of a few actors while threatening to disempower others in the context of shaping both AI research and its applications throughout society. |
2409.18128 | Building on the success of diffusion models in visual generation, flow-based models reemerge as another prominent family of generative models that have achieved competitive or better performance in terms of both visual quality and inference speed. By learning the velocity field through flow-matching, flow-based models tend to produce a straighter sampling trajectory, which is advantageous during the sampling process. However, unlike diffusion models for which fast samplers are well-developed, efficient sampling of flow-based generative models has been rarely explored. In this paper, we propose a framework called FlowTurbo to accelerate the sampling of flow-based models while still enhancing the sampling quality. Our primary observation is that the velocity predictor’s outputs in the flow-based models will become stable during the sampling, enabling the estimation of velocity via a lightweight velocity refiner. Additionally, we introduce several techniques including a pseudo corrector and sample-aware compilation to further reduce inference time. Since FlowTurbo does not change the multi-step sampling paradigm, it can be effectively applied for various tasks such as image editing, inpainting,etc. By integrating FlowTurbo into different flow-based models, we obtain an acceleration ratio of 53.1%58.3% on class-conditional generation and 29.8%38.5% on text-to-image generation. Notably, FlowTurbo reaches an FID of 2.12 on ImageNet with 100 (ms / img) and FID of 3.93 with 38 (ms / img) , achieving the real-time image generation and establishing the new state-of-the-art. Code is available athttps://github.com/shiml20/FlowTurbo. |
2409.17422 | Large Language Models (LLMs) have demonstrated remarkable capabilities in handling long context inputs, but this comes at the cost of increased computational resources and latency. Our research introduces a novel approach for the long context bottleneck to accelerate LLM inference and reduce GPU memory consumption. Our research demonstrates that LLMs can identify relevant tokens in the early layers before generating answers to a query. Leveraging this insight, we propose an algorithm that uses early layers of an LLM as filters to select and compress input tokens, significantly reducing the context length for subsequent processing.
Our method, GemFilter, demonstrates substantial improvements in both speed and memory efficiency compared to existing techniques, such as standard attention and SnapKV/H2O. Notably, it achieves a 2.4speedup and 30% reduction in GPU memory usage compared to SOTA methods. Evaluation on the Needle in a Haystack task shows that GemFilter significantly outperforms standard attention, SnapKV and demonstrates comparable performance on the LongBench challenge.
GemFilter is simple, training-free, and broadly applicable across different LLMs. Crucially, it provides interpretability by allowing humans to inspect the selected input sequence. These findings not only offer practical benefits for LLM deployment, but also enhance our understanding of LLM internal mechanisms, paving the way for further optimizations in LLM design and inference.
Our code is available athttps://github.com/SalesforceAIResearch/GemFilter. |
2409.11321 | There is growing evidence of the effectiveness of Shampoo, a higher-order preconditioning method, over Adam in deep learning optimization tasks. However, Shampoo’s drawbacks include additional hyperparameters and computational overhead when compared to Adam, which only updates running averages of first- and second-moment quantities. This work establishes a formal connection between Shampoo (implemented with the 1/2 power) and Adafactor — a memory-efficient approximation of Adam — showing that Shampoo is equivalent to running Adafactor in the eigenbasis of Shampoo’s preconditioner. This insight leads to the design of a simpler and computationally efficient algorithm:ShampoOwithAdam in thePreconditioner’s eigenbasis (SOAP) . With regards to improving Shampoo’s computational efficiency, the most straightforward approach would be to simply compute Shampoo’s eigendecomposition less frequently.
Unfortunately, as our empirical results show, this leads to performance degradation that worsens with this frequency.
SOAP mitigates this degradation by continually updating the running average of the second moment, just as Adam does, but in the current (slowly changing) coordinate basis. Furthermore, since SOAP is equivalent to running Adam in a rotated space, it introduces only one additional hyperparameter (the preconditioning frequency) compared to Adam. We empirically evaluate SOAP on language model pre-training with 360m and 660m sized models. In the large batch regime, SOAP reduces the number of iterations by over 40% and wall clock time by over 35% compared to AdamW, with approximately 20% improvements in both metrics compared to Shampoo. An implementation of SOAP is available athttps://github.com/nikhilvyas/SOAP. |
2409.14254 | Instruction tuning commonly means finetuning a language model on instruction-response pairs.
We discover two forms of adaptation (tuning) that are deficient compared to instruction tuning, yet still yield instruction following; we call thisimplicit instruction tuning.
We first find that instruction-response pairs are not necessary: training solely onresponses, withoutanycorresponding instructions, yields instruction following.
This suggests pretrained models have an instruction-response mapping which is revealed by teaching the model the desired distribution of responses.
However, we then find it’s not necessary to teach the desired distribution of responses: instruction-response training on narrow-domain data like poetry still leads to broad instruction-following behavior like recipe generation.
In particular, when instructions are very different from those in the narrow finetuning domain, models’ responses do not adhere to the style of the finetuning domain.
To begin to explain implicit instruction tuning, we hypothesize that very simple changes to a language model’s distribution yield instruction following.
We support this byhand-writinga rule-based language model which yields instruction following in a product-of-experts with a pretrained model.
The rules are to slowly increase the probability of ending the sequence, penalize repetition, and uniformly change 15 words’ probabilities.
In summary, adaptations made without being designed to yield instruction following can do soimplicitly. |
2409.12640 | We introduce Michelangelo: a minimal, synthetic, and unleaked long-context reasoning evaluation for large language models which is also easy to automatically score. This evaluation is derived via a novel, unifying framework for evaluations over arbitrarily long contexts which measure the model’s ability to do more than retrieve a single piece of information from its context. The central idea of the Latent Structure Queries framework (LSQ) is to construct tasks which require a model to “chisel away” the irrelevant information in the context, revealing a latent structure in the context. To verify a model’s understanding of this latent structure, we query the model for details of the structure. Using LSQ, we produce three diagnostic long-context evaluations across code and natural-language domains intended to provide a stronger signal of long-context language model capabilities. We perform evaluations on several state-of-the-art models and demonstrate both that a) the proposed evaluations are high-signal and b) that there is significant room for improvement in synthesizing long-context information. |
2409.04599 | Language††*equal contributionmodels can largely benefit from efficient tokenization. However, they still mostly utilize the classical BPE algorithm, a simple and reliable method. This has been shown to cause such issues as under-trained tokens and sub-optimal compression that may affect the downstream performance. We introduce Picky BPE, a modified BPE algorithm that carries out vocabulary refinement during tokenizer training. Our method improves vocabulary efficiency, eliminates under-trained tokens, and does not compromise text compression. Our experiments show that our method does not reduce the downstream performance, and in several cases improves it. |
2409.11136 | Instruction-tuned language models (LM) are able to respond to imperative commands, providing a more natural user interface compared to their base counterparts.
In this work, we present Promptriever, the firstretrievalmodel able to be prompted like an LM.
To train Promptriever,
we curate and release a new instance-level instruction training set from MS MARCO, spanning nearly 500k instances. Promptriever not only achieves strong performance on standard retrieval tasks, but also follows instructions. We observe:
(1) large gains (reaching SoTA) on following detailed relevance instructions (+14.3 p-MRR / +3.1 nDCG on FollowIR) , (2) significantly increased robustness to lexical choices/phrasing in the query+instruction (+12.9 Robustness@10 on InstructIR) , and
(3) the ability to perform hyperparameter search via prompting to reliably improve retrieval performance (+1.4 average increase on BEIR) .
Promptriever demonstrates that retrieval models can be controlled with prompts on a per-query basis, setting the stage for future work aligning LLM prompting techniques with information retrieval.00footnotetext:∗Work performed during an internship at Samaya AI.111Code and data are available athttps://github.com/orionw/promptriever |
2409.07703 | Large Language Models (LLMs) and Large Vision-Language Models (LVLMs) have demonstrated impressive language/vision reasoning abilities, igniting the recent trend of building agents for targeted applications such as shopping assistants or AI software engineers. Recently, many data science benchmarks have been proposed to investigate their performance in the data science domain.
However, existing data science benchmarks still fall short when compared to real-world data science applications due to their simplified settings. To bridge this gap, we introduce\ours, a comprehensive benchmark designed to evaluate data science agents with realistic tasks. This benchmark includes 466 data analysis tasks and 74 data modeling tasks, sourced from Eloquence and Kaggle competitions.\oursoffers a realistic setting by encompassing long contexts, multimodal task backgrounds, reasoning with large data files and multi-table structures, and performing end-to-end data modeling tasks. Our evaluation of state-of-the-art LLMs, LVLMs, and agents shows that they struggle with most tasks, with the best agent solving only 34.12% of data analysis tasks and achieving a 34.74% Relative Performance Gap (RPG) . These findings underscore the need for further advancements in developing more practical, intelligent, and autonomous data science agents.111Our code and data are available at Githubhttps://github.com/LiqiangJing/DSBench.. |
2409.08239 | Large Language Models still struggle in challenging scenarios that leverage structured data, complex reasoning, or tool usage. In this paper, we proposeSource2Synth: a new method that can be used for teaching LLMs new skills without
relying on costly human annotations.Source2Synthtakes as input a custom data source and produces synthetic data points with intermediate reasoning steps grounded in real-world sources.Source2Synthimproves the dataset quality by discarding low-quality generations based on their answerability.
We demonstrate the generality of this approach by applying it to two challenging domains: we test reasoning abilities in multi-hop question answering (MHQA) , and tool usage in tabular question answering (TQA) .
Our method improves performance by 25.51% for TQA on WikiSQL and 22.57% for MHQA on HotPotQA
compared to the fine-tuned baselines. |
2409.07431 | Pretraining on large-scale, unstructured internet text enables language models to acquire a significant amount of world knowledge.
However, this knowledge acquisition isdata-inefficient—to learn a given fact, models must be trained on hundreds to thousands of diverse representations of it.
This poses a challenge when adapting a pretrained model to a small corpus of domain-specific documents, where each fact may appear rarely or only once.
We propose to bridge this gap withsynthetic continued pretraining: using the small domain-specific corpus to synthesize a large corpus more amenable to learning, and then performing continued pretraining on the synthesized corpus.
We instantiate this proposal with EntiGraph, a synthetic data augmentation algorithm that extracts salient entities from the source documents and then generates diverse text by drawing connections between the sampled entities.
Synthetic continued pretraining with EntiGraph enables a language model to answer questions and follow generic instructions related to the source documents without access to them.
If, instead, the source documents are available at inference time, we show that the knowledge acquired through our approach compounds with retrieval-augmented generation.
To better understand these results, we build a simple mathematical model of EntiGraph, and show how synthetic data augmentation can “rearrange” knowledge to enable more data-efficient learning. |
2409.01704 | Traditional OCR systems (OCR-1.0) are increasingly unable to meet people’s usage due to the growing demand for intelligent processing of man-made optical characters. In this paper, we collectively refer to all artificial optical signals (e.g., plain texts, math/molecular formulas, tables, charts, sheet music, and even geometric shapes) as "characters" and propose theGeneralOCRTheory along with an excellent model, namely GOT, to promote the arrival of OCR-2.0. The GOT, with 580M parameters, is a unified, elegant, and end-to-end model, consisting of a high-compression encoder and a long-contexts decoder. As an OCR-2.0 model, GOT can handle all the above "characters" under various OCR tasks. On the input side, the model supports commonly used scene- and document-style images in slice and whole-page styles. On the output side, GOT can generate plain or formatted results (markdown/tikz/smiles/kern) via an easy prompt. Besides, the model enjoys interactive OCR features, i.e., region-level recognition guided by coordinates or colors. Furthermore, we also adapt dynamic resolution and multi-page OCR technologies to GOT for better practicality. In experiments, we provide sufficient results to prove the superiority of our model. |
2104.08786v2 | When primed with only a handful of training samples, very large, pretrained language models such as GPT-3 have shown competitive results when compared to fully-supervised, fine-tuned, large, pretrained language models.
We demonstrate that the order in which the samples are provided can make the difference between near state-of-the-art and random guess performance: essentially some permutations are “fantastic” and some not.
We analyse this phenomenon in detail, establishing that: it is present across model sizes (even for the largest current models) , it is not related to a specific subset of samples, and that a given good permutation for one model is not transferable to another.
While one could use a development set to determine which permutations are performant, this would deviate from the true few-shot setting as it requires additional annotated data.
Instead, we use the generative nature of language models to construct an artificial development set and based on entropy statistics of the candidate permutations on this set, we identify performant prompts.
Our method yields a 13% relative improvement for GPT-family models across eleven different established text classification tasks. |
2408.04658v1 | This paper describes the winning solution of all 5 tasks for the Amazon KDD Cup 2024Multi Task Online
Shopping Challenge for LLMs. The challenge was to build a useful assistant, answering questions in the domain of online shopping. The competition contained 57 diverse tasks, covering 5 different task types (e.g. multiple choice) and across 4 different tracks (e.g. multi-lingual) . Our solution is a single model per track. We fine-tune Qwen2-72B-Instruct on our own training dataset. As the competition released only 96 example questions, we developed our own training dataset by processing multiple public datasets or using Large Language Models for data augmentation and synthetic data generation. We applywise-ftto account for distribution shifts and ensemble multiple LoRA adapters in one model. We employedLogits Processorsto constrain the model output on relevant tokens for the tasks.AWQ 4-bit QuantizationandvLLMare used during inference to predict the test dataset in the time constraints of 20 to 140 minutes depending on the track. Our solution achieved the first place in each individual track and is the first place overall of Amazon’s KDD Cup 2024. |
1711.03953 | We formulate language modeling as a matrix factorization problem, and show that the expressiveness of Softmax-based models (including the majority of neural language models) is limited by aSoftmax bottleneck. Given that natural language is highly context-dependent, this further implies that in practice Softmax with distributed word embeddings does not have enough capacity to model natural language. We propose a simple and effective method to address this issue, and improve the state-of-the-art perplexities on Penn Treebank and WikiText-2 to 47.69 and 40.68 respectively. The proposed method also excels on the large-scale 1B Word dataset, outperforming the baseline by over 5.6 points in perplexity.111Code is available athttps://github.com/zihangdai/mos. |
2402.01502 | Despite their remarkable effectiveness and broad application, the drivers of success underlying ensembles of trees (especially random forests and gradient boosting) are still not fully understood. In this paper, we highlight how interpreting tree ensembles as adaptive and self-regularizing smoothers can provide new intuition and deeper insight to this topic. We use this perspective to show that, when studied as smoothers – whose predictions can be understood as simple weighted averages of the training labels – randomized tree ensembles not only make predictions that are quantifiably more smooth than the predictions of the individual trees they consist of, but also further regulate their smoothness at test-time based on the dissimilarity between testing and training inputs. First, we use this insight to revisit, refine and reconcile two recent explanations of forest success by providing a new way of quantifying the conjectured behaviors of tree ensembles objectively by measuring the effective degree of smoothing they imply. Then, we move beyond existing explanations for the mechanisms by which tree ensembles improve upon individual trees and challenge the popular wisdom that the superior performance of forests should be understood as a consequence of variance reduction alone. We argue that the current high-level dichotomy into bias- and variance-reduction prevalent in statistics is insufficient to understand tree ensembles – because the prevailing definition of bias does not capture differences in the expressivity of the hypothesis classes formed by trees and forests. Instead, we show that forests can improve upon trees by three distinct mechanisms that are usually implicitly entangled. In particular, we demonstrate that the smoothing effect of ensembling can reduce variance in predictions due to noise in outcome generation, reduce variability in the quality of the learned function given fixed input dataandreduce potential bias in learnable functions by enriching the available hypothesis space. |
2401.09395 | Recent advancements in Large Language Models (LLMs) have showcased striking results on existing logical reasoning benchmarks, with some models even surpassing human performance. However, the true depth of their competencies and robustness, in mathematical reasoning tasks, remains an open question. In response, we develop (i) an ontology of perturbations of maths questions, (ii) a semi-automatic method of perturbation, and (iii) a dataset of perturbed maths questions to probe the limits of LLM capabilities in mathematical-reasoning tasks.
These controlled perturbations span across multiple fine dimensions of the structural and representational aspects of maths questions. Using GPT-4, we generated theMoredataset by perturbing randomly selected five seed questions from GSM8K. This process was guided by our ontology and involved a thorough automatic and manual filtering process, yielding a set of 216 maths problems.
We conducted comprehensive evaluation of both closed-source and open-source LLMs onMore. The results show a significant performance drop across all the models against the perturbed questions. This strongly suggests that current LLMs lack robust mathematical skills and deep reasoning abilities. This research not only identifies multiple gaps in the capabilities of current models, but also highlights multiple potential directions for future development. Our dataset will be made publicly available athttps://huggingface.co/datasets/declare-lab/GSM8k_MORE. |
2404.16710 | We present LayerSkip, an end-to-end solution to speed-up inference of large language models (LLMs) . First, during training we apply layer dropout, with low dropout rates for earlier layers and higher dropout rates for later layers, and an early exit loss where all transformer layers share the same exit. Second, during inference, we show that this training recipe increases the accuracy of early exit at earlier layers, without adding any auxiliary layers or modules to the model. Third, we present a novel self-speculative decoding solution where we exit at early layers and verify and correct with remaining layers of the model. Our proposed self-speculative decoding approach has less memory footprint than other speculative decoding approaches and benefits from shared compute and activations of the draft and verification stages. We run experiments on different Llama model sizes on different types of training: pretraining from scratch, continual pretraining, finetuning on specific data domain, and finetuning on specific task. We implement our inference solution and show speedups of up to 2.16on summarization for CNN/DM documents, 1.82on coding, and 2.0on TOPv2 semantic parsing task. |
2409.12737 | Current pre-trained cross-lingual sentence encoders approaches use sentence-level objectives only.
This can lead to loss of information, especially for tokens, which then degrades the sentence representation.
We propose MEXMA, a novel approach that integrates both sentence-level and token-level objectives.
The sentence representation in one language is used to predict masked tokens in another language, with both the sentence representation andall tokens directly updating the encoder.
We show that adding token-level objectives greatly improves the sentence representation quality across several tasks.
Our approach outperforms current pre-trained cross-lingual sentence encoders on bi-text mining as well as several downstream tasks.
We also analyse the information encoded in our tokens, and how the sentence representation is built from them. |
2407.20311 | Recent advances in language models have demonstrated their capability to solve mathematical reasoning problems, achieving near-perfect accuracy on grade-school level math benchmarks like GSM8K. In this paper, we formally study how language models solve these problems. We design a series of controlled experiments to address several fundamental questions: (1) Can language models truly develop reasoning skills, or do they simply memorize templates? (2) What is the model’s hidden (mental) reasoning process? (3) Do models solve math questions using skills similar to or different from humans? (4) Do models trained on GSM8K-like datasets develop reasoning skills beyond those necessary for solving GSM8K problems? (5) What mental process causes models to make reasoning mistakes? (6) How large or deep must a model be to effectively solve GSM8K-level math questions? Our study uncovers many hidden mechanisms by which language models solve mathematical questions, providing insights that extend beyond current understandings of LLMs. |
2308.06721 | Recent years have witnessed the strong power of large text-to-image diffusion models for the impressive generative capability to create high-fidelity images. However, it is very tricky to generate desired images using only text prompt as it often involves complex prompt engineering. An alternative to text prompt is image prompt, as the saying goes: "an image is worth a thousand words". Although existing methods of direct fine-tuning from pretrained models are effective, they require large computing resources and are not compatible with other base models, text prompt, and structural controls. In this paper, we present IP-Adapter, an effective and lightweight adapter to achieve image prompt capability for the pretrained text-to-image diffusion models. The key design of our IP-Adapter is decoupled cross-attention mechanism that separates cross-attention layers for text features and image features. Despite the simplicity of our method, an IP-Adapter with only 22M parameters can achieve comparable or even better performance to a fully fine-tuned image prompt model. As we freeze the pretrained diffusion model, the proposed IP-Adapter can be generalized not only to other custom models fine-tuned from the same base model, but also to controllable generation using existing controllable tools. With the benefit of the decoupled cross-attention strategy, the image prompt can also work well with the text prompt to achieve multimodal image generation. The project page is available athttps://ip-adapter.github.io. |
2405.00332 | Large language models (LLMs) have achieved impressive success on many benchmarks for mathematical reasoning.
However, there is growing concern that some of this performance actually reflects dataset contamination, where data closely resembling benchmark questions leaks into the training data, instead of true reasoning ability.
To investigate this claim rigorously, we commissionGrade School Math 1000(GSM1k) . GSM1k is designed to mirror the style and complexity of the established GSM8k benchmark,
the gold standard for measuring elementary mathematical reasoning. We ensure that the two benchmarks are comparable across important metrics such as human solve rates, number of steps in solution, answer magnitude, and more.
When evaluating leading open- and closed-source LLMs on GSM1k, we observe accuracy drops of up to 13%, with several families of models (e.g. Phi and Mistral) showing evidence of systematic overfitting across almost all model sizes.
At the same time, many models, especially those on the frontier, (e.g. Gemini/GPT/Claude) show minimal signs of overfitting.
Further analysis suggests a positive relationship (Spearman’s) between a model’s probability of generating an example from GSM8k and its performance gap between GSM8k and GSM1k, suggesting that many models may have partially memorized GSM8k. |
2111.15640 | Diffusion probabilistic models (DPMs) have achieved remarkable quality in image generation that rivals GANs’. But unlike GANs, DPMs use a set of latent variables that lack semantic meaning and cannot serve as a useful representation for other tasks.
This paper explores the possibility of using DPMs for representation learning and seeks to extract a meaningful and decodable representation of an input image via autoencoding. Our key idea is to use a learnable encoder for discovering the high-level semantics, and a DPM as the decoder for modeling the remaining stochastic variations.
Our method can encode any image into a two-part latent code where the first part is semantically meaningful and linear, and the second part captures stochastic details, allowing near-exact reconstruction. This capability enables challenging applications that currently foil GAN-based methods, such as attribute manipulation onrealimages. We also show that this two-level encoding improves denoising efficiency and naturally facilitates various downstream tasks including few-shot conditional sampling. Please visit our page:https://Diff-AE.github.io/ |
2308.05037 | Language-queried audio source separation (LASS) is a new paradigm for computational auditory scene analysis (CASA) . LASS aims to separate a target sound from an audio mixture given a natural language query, which provides a natural and scalable interface for digital audio applications. Recent works on LASS, despite attaining promising separation performance on specific sources (e.g., musical instruments, limited classes of audio events) , are unable to separate audio concepts in the open domain. In this work, we introduce AudioSep, a foundation model for open-domain audio source separation with natural language queries. We train AudioSep on large-scale multimodal datasets and extensively evaluate its capabilities on numerous tasks including audio event separation, musical instrument separation, and speech enhancement. AudioSep demonstrates strong separation performance and impressive zero-shot generalization ability using audio captions or text labels as queries, substantially outperforming previous audio-queried and language-queried sound separation models. For reproducibility of this work, we released the source code, evaluation benchmark and pre-trained model at:https://github.com/Audio-AGI/AudioSep. |
2409.02529 | For learned image representations, basic autoencoders often produce blurry results. Reconstruction quality can be improved by incorporating additional penalties such as adversarial (GAN) and perceptual losses. Arguably, these approaches lack a principled interpretation. Concurrently, in generative settings diffusion has demonstrated a remarkable ability to create crisp, high quality results and has solid theoretical underpinnings (from variational inference to direct study as the Fisher Divergence) . Our work combines autoencoder representation learning with diffusion and is, to our knowledge, the first to demonstrate the efficacy ofjointly learning a continuous encoder and decoder under a diffusion-based loss.
We demonstrate that this approach yields better reconstruction quality as compared to GAN-based
autoencoders while being easier to tune.
We also show that the resulting representation is easier to model
with a latent diffusion model as compared to the representation obtained from a state-of-the-art GAN-based loss.
Since our decoder is stochastic, it can generate details not encoded in the otherwise deterministic latent representation; we therefore name our approach “Sample what you can’t compress”, or SWYCC for short. |
2201.03533 | NLP benchmarks have largely focused onshorttexts, such as sentences and paragraphs, even thoughlongtexts comprise a considerable amount of natural language in the wild.
We introduceScrolls, a suite of tasks that require reasoning over long texts.
We examine existing long-text datasets, and handpick ones where the text is naturally long, while prioritizing tasks that involve synthesizing information across the input.Scrollscontains summarization, question answering, and natural language inference tasks, covering multiple domains, including literature, science, business, and entertainment.
Initial baselines, including Longformer Encoder-Decoder, indicate that there is ample room for improvement onScrolls.
We make all datasets available in a unified text-to-text format and host a live leaderboard to facilitate research on model architecture and pretraining methods.111https://www.scrolls-benchmark.com |
2405.16406 | Post-training quantization (PTQ) techniques applied to weights, activations, and the KV cache greatly reduce memory usage, latency, and power consumption of Large Language Models (LLMs) , but may lead to large quantization errors when outliers are present. Recent findings suggest that rotating activation or weight matrices helps remove outliers and benefits quantization. In this work, we identify a collection of applicable rotation parameterizations that lead to identical outputs in full-precision Transformer architectures, and find that some random rotations lead to much better quantization than others, with an up to13 pointsdifference in downstream zero-shot reasoning performance. As a result, we proposeSpinQuantthatoptimizes(orlearns) the rotation matrices withCayleyoptimization on a small validation set. With 4-bit quantization of weight, activation, and KV-cache,SpinQuantnarrows the accuracy gap on zero-shot reasoning tasks with full precision to merely 2.9 points on the LLaMA-2 7B model, surpassing LLM-QAT by 19.1 points and SmoothQuant by 25.0 points.SpinQuantalso outperforms concurrent work QuaRot, which applies random rotations to remove outliers. In particular, for LLaMA-2 7B/LLaMA-3 8B models that are hard to quantize,SpinQuantreduces the gap to full precision by 30.2%/34.1% relative to QuaRot. |