Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
Job has been terminated due to a temporary spike in resource usage and may be restarted later.
Error code: JobManagerCrashedError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
title string | url string | conference string | year int64 | label string | abstract string | keywords string | primary area string | main string | approach string | rating float64 | contribute float64 | soundness float64 | novelty string | analysis dict | embedding list |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
AgentTuning: Enabling Generalized Agent Abilities for LLMs | https://openreview.net/forum?id=OqlmgmS4Wr | ICLR | 2,024 | reject | Open large language models (LLMs) with great performance in various tasks have significantly advanced the development of LLMs. However, they are far inferior to commercial models such as ChatGPT and GPT-4 when acting as agents to tackle complex tasks in the real world. These agent tasks employ LLMs as the central contr... | Large language models, Autonomous agents, Instruction tuning, Reasoning | general machine learning (i.e., none of the above) | To enhance the agent capabilities of open-source large language models (LLMs) while maintaining their general abilities | Construct a lightweight, high-quality instruction-tuning dataset (AgentDataset) focused on agent interaction trajectories; Employ a hybrid instruction-tuning strategy by combining AgentDataset with open-domain instruction data; Apply the AgentTuning method to instruction-tune open LLMs (specifically Llama 2 series) to ... | 3.084779 | 3.0864 | 2.735635 | 0 | {
"main_target": "This research introduces a new method, AgentTuning, to enhance the agent capabilities of large language models while maintaining their general abilities.",
"core_breakthrough": "AgentTuning enhances the agent capabilities of large language models while maintaining their general abilities, making t... | [
0.013565224595367908,
0.02365085668861866,
-0.019709408283233643,
0.0005152739468030632,
-0.009773542173206806,
-0.036873746663331985,
0.0012512614484876394,
-0.02724866382777691,
-0.03088361583650112,
0.017608968541026115,
0.013539995066821575,
0.009395353496074677,
0.017610765993595123,
... |
AutoAgents: A Framework for Automatic Agent Generation | https://openreview.net/forum?id=PhJUd3mbhP | ICLR | 2,024 | reject | Large language models (LLMs) have enabled remarkable advances in automated task-solving with multi-agent systems. However, most existing LLM-based multi-agent approaches rely on predefined agents to handle simple tasks, limiting the adaptability of multi-agent collaboration to different scenarios. Therefore, we introdu... | Multi-Agent Framework, Large Language Models, Agent Generation | applications to robotics, autonomy, planning | To propose a novel framework (AutoAgents) for adaptive multi-agent collaboration in LLM-based task-solving | Dynamically generating specialized agents according to task content; Coordinating expert agents to collaboratively plan and solve tasks; Coupling task requirements with agent roles for flexible team assembly; Incorporating an observer role to reflect on plans and agent responses for iterative improvement; Demonstrating... | 2.959779 | 2.8364 | 2.068968 | 0 | {
"main_target": "This research introduces AutoAgents, an innovative framework that adaptively generates and coordinates multiple specialized agents to build an AI team according to different tasks.",
"core_breakthrough": "AutoAgents introduces a framework that dynamically generates and coordinates multiple special... | [
0.02637295238673687,
0.037663839757442474,
-0.036163173615932465,
0.023295307531952858,
-0.02167409472167492,
-0.034308988600969315,
-0.026479285210371017,
0.01503751054406166,
0.018995657563209534,
0.0020226582419127226,
0.027256088331341743,
-0.00814026314765215,
-0.020357372239232063,
-... |
Agent Instructs Large Language Models to be General Zero-Shot Reasoners | https://openreview.net/forum?id=zIJFG7wW2d | ICLR | 2,024 | reject | We introduce a method to improve the zero-shot reasoning abilities of large language models on general language understanding tasks. Unlike existing zero-shot reasoning approaches that are often suboptimal for general tasks, we build an autonomous agent to generate task-specific instructions to optimize the reasoning p... | Language Models, Large Language Models, Reasoning, Agents, Chain of Thought | transfer learning, meta learning, and lifelong learning | To improve the zero-shot reasoning abilities of large language models on general language understanding tasks | Develop an autonomous agent to generate task-specific instructions for optimizing LLM reasoning; Evaluate the method across diverse datasets spanning generation, classification, and reasoning tasks; Compare performance against existing zero-shot and zero-shot chain of thought approaches; Demonstrate state-of-the-art re... | 2.418112 | 2.0864 | 2.402302 | 0 | {
"main_target": "This research introduces a method to enhance the zero-shot reasoning abilities of large language models across various language understanding tasks.",
"core_breakthrough": "This paper introduces an autonomous agent that generates task-specific instructions to enhance the zero-shot reasoning abilit... | [
0.03526664152741432,
0.030228329822421074,
-0.03082815557718277,
-0.0219961479306221,
-0.01423436775803566,
0.0006881923764012754,
0.002248915610834956,
-0.004690598696470261,
-0.03760180249810219,
0.05377530679106712,
0.02593342401087284,
0.022728627547621727,
0.019901875406503677,
-0.035... |
OKR-Agent: An Object and Key Results Driven Agent System with Hierarchical Self-Collaboration and Self-Evaluation | https://openreview.net/forum?id=Mngdhgi711 | ICLR | 2,024 | reject | In this study, we introduce the concept of OKR-Agent designed to enhance the capabilities of Large Language Models (LLMs) in task-solving. Our approach utilizes both self-collaboration and self-correction mechanism, facilitated by hierarchical agents, to address the inherent complexities in task-solving. Our key observ... | Large Language Model, Agent | generative models | To introduce OKR-Agent, a hierarchical agent framework that enhances Large Language Models (LLMs) for complex task-solving | Incorporation of self-collaboration and self-correction mechanisms; Deployment of hierarchical agents specialized for sub-tasks; Hierarchical Objects and Key Results (OKR) generation for recursive task decomposition and agent assignment; Multi-level evaluation module leveraging agent feedback for iterative solution ref... | 2.459779 | 2.3364 | 2.568968 | 0 | {
"main_target": "This research introduces a new framework called OKR-Agent for enhancing the task-solving capabilities of Large Language Models through hierarchical agents and a self-collaboration and self-correction mechanism.",
"core_breakthrough": "This paper introduces OKR-Agent, a hierarchical and self-collab... | [
0.052445266395807266,
-0.00008304835500894114,
-0.04119718447327614,
-0.016154926270246506,
0.006549533922225237,
-0.037995774298906326,
0.010530336759984493,
-0.023251788690686226,
-0.003224062267690897,
0.03407598286867142,
0.005068144295364618,
0.003293013898655772,
0.002805394120514393,
... |
MADiff: Offline Multi-agent Learning with Diffusion Models | https://openreview.net/forum?id=mHF35XVjmm | ICLR | 2,024 | reject | Diffusion model (DM), as a powerful generative model, recently achieved huge success in various scenarios including offline reinforcement learning, where the policy learns to conduct planning by generating trajectory in the online evaluation. However, despite the effectiveness shown for single-agent learning, it remain... | Multi-agent RL, Diffusion Models, Offline RL | reinforcement learning | To propose MADiff, a novel diffusion-based generative framework for multi-agent offline reinforcement learning | Utilize an attention-based diffusion model to capture complex coordination among multiple agents; Enable MADiff to function as both a decentralized policy and a centralized controller; Incorporate opponent modeling during decentralized execution; Apply the centralized controller for multi-agent trajectory prediction; D... | 2.784779 | 2.6864 | 2.468968 | 0 | {
"main_target": "This research proposes MADIFF, a novel generative multiagent learning framework using an attention-based diffusion model for complex coordination in multi-agent problems.",
"core_breakthrough": "MADIFF introduces the first diffusion-based multi-agent offline RL framework that effectively models co... | [
0.015091747045516968,
0.04312135651707649,
-0.0314951166510582,
0.011631409637629986,
0.008131753653287888,
-0.06044472008943558,
-0.061530742794275284,
0.02114110440015793,
-0.0026674780528992414,
0.03436799719929695,
0.0020112013444304466,
0.013020221143960953,
0.010376178659498692,
-0.0... |
FireAct: Toward Language Agent Finetuning | https://openreview.net/forum?id=RqUMWdDg52 | ICLR | 2,024 | reject | Recent efforts have augmented language models (LMs) with external tools or environments, leading to the development of language agents that can reason and act. However, most of these agents rely on few-shot prompting techniques, which can result in a lack of robustness in agent performance due to the limited learning s... | language agent, language model, large language model, finetuning, agent, tool use | representation learning for computer vision, audio, language, and other modalities | To systematically investigate the effectiveness and robustness of fine-tuning language models to create language agents for tool-augmented question answering | Employ a controlled setup with a Google search API for QA; Explore a variety of base LMs and agent methods; Fine-tune LMs using trajectory data from different tasks and agent methods; Evaluate the impact of scaling base LMs and fine-tuning data; Assess robustness to different types of data perturbations; Derive empiric... | 2.209779 | 2.0864 | 2.568968 | 0 | {
"main_target": "This research investigates and demonstrates the benefits of fine-tuning language models to create more effective language agents for question answering tasks.",
"core_breakthrough": "This paper demonstrates that fine-tuning language models with diverse agent trajectories significantly improves the... | [
0.04541543871164322,
0.02443549782037735,
-0.05156244710087776,
-0.020065447315573692,
-0.016932541504502296,
0.010173840448260307,
-0.026613011956214905,
-0.008920695632696152,
-0.011834369041025639,
0.008855806663632393,
0.025816205888986588,
0.02350369095802307,
-0.0022721849381923676,
... |
Detecting Influence Structures in Multi-Agent Reinforcement Learning | https://openreview.net/forum?id=DQTxr8JtPX | ICLR | 2,024 | reject | We consider the problem of quantifying the amount of influence one agent can exert on another in the setting of multi-agent reinforcement learning (MARL). As a step towards a unified approach to express agents' interdependencies, we introduce the total and state influence measurement functions. Both of these are valid ... | multi-agent reinforcement learning, multi-agent interdependencies, stochastic approximation, decentralized algorithms | reinforcement learning | To introduce novel influence measurement functions and efficient approximation algorithms for quantifying agent influence in multi-agent reinforcement learning (MARL) | Define total and state influence measurement functions applicable to common MARL settings; Propose total impact measurement (TIM) and state impact measurement (SIM) for characterizing agent influence in the average reward setting; Develop decentralized approximation algorithms for TIM and SIM that only require observin... | 2.209779 | 2.5864 | 2.568968 | 0 | {
"main_target": "This research introduces new functions and measurements to quantify and analyze the influence of one agent on another in multi-agent reinforcement learning systems.",
"core_breakthrough": "This paper introduces the first unified approach to measure and quantify the influence among agents in multi-... | [
0.021040575578808784,
0.06133851781487465,
-0.016677433624863625,
0.01490064151585102,
-0.012809191830456257,
-0.045105401426553726,
-0.05622291937470436,
0.013033701106905937,
-0.0017173135420307517,
0.02540556713938713,
-0.007138100452721119,
0.03214054927229881,
0.0076018087565898895,
0... |
Generalizing Poincaré Policy Representations in Multi-agent Reinforcement Learning | https://openreview.net/forum?id=2DJUXmHZ2O | ICLR | 2,024 | reject | "Learning policy representations is essential for comprehending the intricacies of agent interaction(...TRUNCATED) | policy representation, reinforcement learning, multi-agent | reinforcement learning | "To introduce a new geometric framework for learning policy representations in multi-agent reinforce(...TRUNCATED) | "Project multi-agent trajectory growth as tree structures onto the Poincaré ball; Design a hyperbol(...TRUNCATED) | 2.459779 | 2.3364 | 2.568968 | 0 | {"main_target":"This research introduces a new framework for learning Poincaré Policy Representatio(...TRUNCATED) | [0.0031482491176575422,0.041641589254140854,-0.02465793490409851,-0.008213028311729431,0.03765483200(...TRUNCATED) |
MACCA: Offline Multi-agent Reinforcement Learning with Causal Credit Assignment | https://openreview.net/forum?id=mFBR2ksIwY | ICLR | 2,024 | reject | "Offline Multi-agent Reinforcement Learning (MARL) is valuable in scenarios where online interaction(...TRUNCATED) | Offline Multi-Agent Reinforcement Learning, Credit Assignment, Causal Inference | causal reasoning | "To develop a novel credit assignment method for offline Multi-agent Reinforcement Learning (MARL) t(...TRUNCATED) | "Modeling the generative process as a Dynamic Bayesian Network to capture relationships among variab(...TRUNCATED) | 2.584779 | 2.5864 | 2.818968 | 0 | {"main_target":"This research proposes a new method, MACCA, for accurate credit assignment in offlin(...TRUNCATED) | [0.018677091225981712,0.05792172625660896,-0.02195468544960022,0.011274329386651516,-0.0038842833600(...TRUNCATED) |
"Translating cognitive models into neural and statistical descriptions of real-world multi-agent for(...TRUNCATED) | https://openreview.net/forum?id=2qLSkTuqrb | ICLR | 2,024 | reject | "Foraging is a multi-agent social behavior that has been studied from many perspectives, including c(...TRUNCATED) | "multi-agent systems, animal behavior, reinforcement learning, probabilistic methods, decision makin(...TRUNCATED) | applications to neuroscience & cognitive science | "To develop a biologically plausible and interpretable framework for analyzing multi-agent foraging (...TRUNCATED) | "Implement cognitive agent models as neural networks with value functions; Formulate an equivalent s(...TRUNCATED) | 2.459779 | 2.5864 | 2.568968 | 0 | {"main_target":"This research develops a biologically plausible neural network and an equivalent sta(...TRUNCATED) | [0.0008098215912468731,0.05448922514915466,-0.012290899641811848,-0.014710215851664543,0.00821905490(...TRUNCATED) |
End of preview.
No dataset card yet
- Downloads last month
- 40