This is a collection of recent papers submitted to ICLR’24, which are focusing on building autonomous agent. These papers have been integrated to this github repo, which is in active maintainance. Feel free to star/follow the repo.

Here is how Wikipedia defines Agent:

In artificial intelligence, an intelligent agent is an agent acting in an intelligent manner; It perceives its environment, takes actions autonomously in order to achieve goals, and may improve its performance with learning or acquiring knowledge. An intelligent agent may be simple or complex: A thermostator other control systemis considered an example of an intelligent agent, as is a human being, as is any system that meets the definition, such as a firm, a state, or a biome.

Awesome-Papers-Autonomous-Agent

The key of an agent is that it can achieve goals, acquire knowledge and continually improveme.

Specifically, this repo is interested in two types of agent: RL-based agent and LLM-based agent.

RL-based agent

[RL-based、world model] Learning to Model the World with Language

[RL-based、world model] MAMBA: an Effective World Model Approach for Meta-Reinforcement Learning

[RL-based、language knowledge、continual learning] Learning with Language Inference and Tips for Continual Reinforcement Learning

[RL-based、language knowledge] Informing Reinforcement Learning Agents by Grounding Natural Language to Markov Decision Processes

[RL-based、language knowledge] Language Reward Modulation for Pretraining Reinforcement Learning

[RL-based、instruction following] Compositional Instruction Following with Language Models and Reinforcement Learning

[RL-based、algorithm] ADAPTER-RL: Adaptation of Any Agent using Reinforcement Learning

[RL-based、Other agent] Online Continual Learning for Interactive Instruction Following Agents

[RL-based、LLM as tool] Leveraging Large Language Models for Optimised Coordination in Textual Multi-Agent Reinforcement Learning

[RL-based、LLM as tool] Text2Reward: Dense Reward Generation with Language Models for Reinforcement Learning

[RL-based agent、generalization] STARLING: Self-supervised Training of Text-based Reinforcement Learning Agent with Large Language Models

[RL-based agent、generalization] AMAGO: Scalable In-Context Reinforcement Learning for Adaptive Agents

[RL-based agent] Understanding Your Agent: Leveraging Large Language Models for Behavior Explanation

[RL-based agent] Enhancing Human Experience in Human-Agent Collaboration: A Human-Centered Modeling Approach Based on Positive Human Gain

[RL-based agent] A Competition Winning Deep Reinforcement Learning Agent in microRTS

[RL-based agent] Aligning Agents like Large Language Models

[RL-based + LLM-based、robotics] RoboGPT : An intelligent agent of making embodied long-term decisions for daily instruction tasks

[RL-based + LLM-based] Can Language Agents Approach the Performance of RL? An Empirical Study On OpenAI Gym

[RL-based + LLM-based] RLAdapter: Bridging Large Language Models to Reinforcement Learning in Open Worlds

[RL-based] Multi-agent Trajectory Prediction with Scalable Diffusion Transformer

LLM-based agent

[LLM-based、vision] Steve-Eye: Equipping LLM-based Embodied Agents with Visual Perception in Open Worlds

[LLM-based、vision] Multimodal Web Navigation with Instruction-Finetuned Foundation Models

[LLM-based、vision] We propose Auto-UI, a multimodal model with a novel chain-of-action technique. Auto-UI can directly interact with the user interface, bypassing the need for environment parsing or reliance on application-dependent APIs.

[LLM-based、vision] Learning Embodied Vision-Language Programming From Instruction, Exploration, and Environmental Feedback

[LLM-based、vision] An Embodied Generalist Agent in 3D World

[LLM-based、training、task-specific] Language Agents with Reinforcement Learning for Strategic Play in the Werewolf Game

[LLM-based、training] FireAct: Toward Language Agent Finetuning

[LLM-based、training] Adapting LLM Agents Through Communication

[LLM-based、training] AgentTuning: Enabling Generalized Agent Abilities for LLMs

[LLM-based、training] Retroformer: Retrospective Large Language Agents with Policy Gradient Optimization

[LLM-based、task-specific] Rethinking the Buyer’s Inspection Paradox in Information Markets with Language Agents

[LLM-based、task-specific] A Language-Agent Approach to Formal Theorem-Proving

[LLM-based、task-specific] Agent Instructs Large Language Models to be General Zero-Shot Reasoners

[LLM-based、protocol、code] LUMOS: Towards Language Agents that are Unified, Modular, and Open Source

[LLM-based、multi-agent] Building Cooperative Embodied Agents Modularly with Large Language Models

[LLM-based、multi-agent] OKR-Agent: An Object and Key Results Driven Agent System with Hierarchical Self-Collaboration and Self-Evaluation

[LLM-based、multi-agent] MetaGPT: Meta Programming for Multi-Agent Collaborative Framework

[LLM-based、multi-agent] AutoAgents: A Framework for Automatic Agent Generation

[LLM-based、multi-agent] Dynamic LLM-Agent Network: An LLM-agent Collaboration Framework with Agent Team Optimization

[LLM-based、multi-agent] AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviors

[LLM-based、multi-agent] Exploring Collaboration Mechanisms for LLM Agents: A Social Psychology View

[LLM-based、learning] REX: Rapid Exploration and eXploitation for AI agents

[LLM-based、framework] AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation

[LLM-based、experiments、safety] Identifying the Risks of LM Agents with an LM-Emulated Sandbox

[LLM-based、experiments] Evaluating Multi-Agent Coordination Abilities in Large Language Models

[LLM-based、experiments] Large Language Models as Gaming Agents

[LLM-based、experiments] Benchmarking Large Language Models as AI Research Agents

[LLM-based、dynamic environment] Adaptive Environmental Modeling for Task-Oriented Language Agents

[LLM-based、continual learning] CLIN: A Continually Learning Language Agent for Rapid Task Adaptation and Generalization

[LLM-based、benchmark] SmartPlay : A Benchmark for LLMs as Intelligent Agents

[LLM-based、benchmark] AgentBench: Evaluating LLMs as Agents

[LLM-based、benchmark] Put Your Money Where Your Mouth Is: Evaluating Strategic Planning and Execution of LLM Agents in an Auction Arena

[LLM-based、benchmark] SOTOPIA: Interactive Evaluation for Social Intelligence in Language Agents

[LLM-based、benchmark] SocioDojo: Building Lifelong Analytical Agents with Real-world Text and Time Series

[LLM-based、benchmark] WebArena: A Realistic Web Environment for Building Autonomous Agents

[LLM-based、benchmark] LLM-Deliberation: Evaluating LLMs with Interactive Multi-Agent Negotiation Game

[LLM-based、benchmark] Evaluating Large Language Models at Evaluating Instruction Following

[LLM-based、benchmark] CivRealm: A Learning and Reasoning Odyssey for Decision-Making Agents

[LLM-based、application] Lyfe Agents: generative agents for low-cost real-time social interactions

[LLM-based、algorithm] Prospector: Improving LLM Agents with Self-Asking and Trajectory Ranking

[LLM-based、algorithm] Formally Specifying the High-Level Behavior of LLM-Based Agents

[LLM-based、RL] Teaching LLMs to Teach Themselves Better Instructions via Reinforcement Learning

[LLM-based] PaperQA: Retrieval-Augmented Generative Agent for Scientific Research

[LLM-based] Language Agents for Detecting Implicit Stereotypes in Text-to-image Models at Scale

[LLM-based] Ghost in the Minecraft: Hierarchical Agents for Minecraft via Large Language Models with Text-based Knowledge and Memory

[LLM-based] Lemur: Harmonizing Natural Language and Code for Language Agents

[LLM-based] Language Agent Tree Search Unifies Reasoning Acting and Planning in Language Models