AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviors
2023/08/21 by Chen, Weize, Su, Yusheng, Zuo, Jingwei +13 · 95 citations
#Computation and Language (cs.CL) #FOS: Computer and information sciences
paper · doi:10.48550/arxiv.2308.10848
Abstract
Autonomous agents empowered by Large Language Models (LLMs) have undergone significant improvements, enabling them to generalize across a broad spectrum of tasks. However, in real-world scenarios, cooperation among individuals is often required to enhance the efficiency and effectiveness of task accomplishment. Hence, inspired by human group dynamics, we propose a multi-agent framework \framework that can collaboratively and dynamically adjust its composition as a greater-than-the-sum-of-its-parts system. Our experiments demonstrate that \framework framework can effectively deploy multi-agent groups that outperform a single agent. Furthermore, we delve into the emergence of social behaviors among individual agents within a group during collaborative task accomplishment. In view of these behaviors, we discuss some possible strategies to leverage positive ones and mitigate negative ones for improving the collaborative potential of multi-agent groups. Our codes for \framework will soon be released at \urlhttps://github.com/OpenBMB/AgentVerse.
Cited by
- Matryoshka Agent: Unfolding Sub-Agents for Long-Horizon Machine Learning Engineering
- Agent Team Work Zone: An Automated, Persistent Workspace for Long-Lived Claude Code Agent Teams
- Evolving from Lessons: Skill-Augmented Table Graph Reasoning for Operation-wise Table Question Answering
- An Information Theoretic Perspective on Agentic System Design
- Multi-Agent LLM Committees for Autonomous Software Beta Testing
- Agent Tools Orchestration Leaks More: Dataset, Benchmark, and Mitigation
- Towards a Science of Scaling Agent Systems
- LoopBench: Discovering Emergent Symmetry Breaking Strategies with LLM Swarms
- MSME: A Multi-Stage Multi-Expert Framework for Zero-Shot Stance Detection
- Reason-Plan-ReAct: A Reasoner-Planner Supervising a ReAct Executor for Complex Enterprise Tasks
- RoCo: Role-Based LLMs Collaboration for Automatic Heuristic Design
- Tipping the Dominos: Topology-Aware Multi-Hop Attacks on LLM-Based Multi-Agent Systems
- Beyond Single-Agent Safety: A Taxonomy of Risks in LLM-to-LLM Interactions
- Agent-Kernel: A MicroKernel Multi-Agent System Framework for Adaptive Social Simulation Powered by LLMs
- CLIMATEAGENT: Multi-Agent Orchestration for Complex Climate Data Science Workflows
- From Natural Language to Certified H-infinity Controllers: Integrating LLM Agents with LMI-Based Synthesis
- Maestro: Learning to Collaborate via Conditional Listwise Policy Optimization for Multi-Agent LLMs
- PEFA-AI: Advancing Open-source LLMs for RTL generation using Progressive Error Feedback Agentic-AI
- PublicAgent: Multi-Agent Design Principles From an LLM-Based Open Data Analysis Framework
- Optimal-Agent-Selection: State-Aware Routing Framework for Efficient Multi-Agent Collaboration
- The Collaboration Gap
- SIGMA: Search-Augmented On-Demand Knowledge Integration for Agentic Mathematical Reasoning
- Generalizing Test-time Compute-optimal Scaling as an Optimizable Graph
- Enhancing XAI Narratives through Multi-Narrative Refinement and Knowledge Distillation
- How memory can affect collective and cooperative behaviors in an LLM-Based Social Particle Swarm
- Ripple Effect Protocol: Coordinating Agent Populations
- Does Socialization Emerge in AI Agent Society? A Case Study of Moltbook
- How Well Can LLM Agents Simulate End-User Security and Privacy Attitudes and Behaviors?
- ProMediate: A Socio-cognitive framework for evaluating proactive agents in multi-party negotiation
- DEBATE: A Large-Scale Benchmark for Role-Playing LLM Agents in Multi-Agent, Long-Form Debates
- MIC-BEV: Multi-Infrastructure Camera Bird's-Eye-View Transformer with Relation-Aware Fusion for 3D Object Detection
- Integrating LLM and Diffusion-Based Agents for Social Simulation
- Scalable Oversight via Partitioned Human Supervision
- Monitoring LLM-based Multi-Agent Systems Against Corruptions via Node Evaluation
- LLMartini: Seamless and Interactive Leveraging of Multiple LLMs through Comparison and Composition
- Communication to Completion: Modeling Collaborative Workflows with Intelligent Multi-Agent Communication
- Formalizing the Safety, Security, and Functional Properties of Agentic AI Systems
- A2FM: An Adaptive Agent Foundation Model for Tool-Aware Hybrid Reasoning
- HyperAgent: Leveraging Hypergraphs for Topology Optimization in Multi-Agent Communication
- Failure-Driven Workflow Refinement
- Dynamic Generation of Multi-LLM Agents Communication Topologies with Graph Diffusion Models
- Multimodal Safety Evaluation in Generative Agent Social Simulations
- Simulating Teams with LLM Agents: Interactive 2D Environments for Studying Human-AI Dynamics
- Traceability and Accountability in Role-Specialized Multi-Agent LLM Pipelines
- Measuring and Mitigating Identity Bias in Multi-Agent Debate via Anonymization
- ProSEA: Problem Solving via Exploration Agents
- AMAS: Adaptively Determining Communication Topology for LLM-based Multi-Agent System
- ARM: Discovering Agentic Reasoning Modules for Generalizable Multi-Agent Systems
- Alignment Tipping Process: How Self-Evolution Pushes LLM Agents Off the Rails
- LEGOMem: Modular Procedural Memory for Multi-agent LLM Systems for Workflow Automation
- LLM Chemistry Estimation for Multi-LLM Recommendation
- AutoMaAS: Self-Evolving Multi-Agent Architecture Search for Large Language Models
- JoyAgent-JDGenie: Technical Report on the GAIA
- Stochastic Self-Organization in Multi-Agent Systems
- CORTEX: Collaborative LLM Agents for High-Stakes Alert Triage
- ACT: Agentic Classification Tree
- The Hunger Game Debate: On the Emergence of Over-Competition in Multi-Agent Systems
- Flash-Searcher: Fast and Effective Web Agents via DAG-Based Parallel Execution
- MCPMark: A Benchmark for Stress-Testing Realistic and Comprehensive MCP Use
- RADAR: A Risk-Aware Dynamic Multi-Agent Framework for LLM Safety Evaluation via Role-Specialized Collaboration
- Diagnose, Localize, Align: A Full-Stack Framework for Reliable LLM Multi-Agent Systems under Instruction Conflicts
- Peacemaker or Troublemaker: How Sycophancy Shapes Multi-Agent Debate
- RobustFlow: Towards Robust Agentic Workflow Generation
- Eigen-1: Adaptive Multi-Agent Refinement with Monitor-Based RAG for Scientific Reasoning
- What Do LLM Agents Do When Left Alone? Evidence of Spontaneous Meta-Cognitive Patterns
- Embodied AI: From LLMs to World Models
- FaithEyes: Towards Faithful Tool Use via Multi-Agent Process-Image Verification
- Scaling LLM-Driven Multi-Agent Systems: Design Principles and Architectural Scalability Analysis
- Escaping the Self-Confirmation Trap: An Execute-Distill-Verify Paradigm for Agentic Experience Learning
- RepoTransAgent: Multi-Agent LLM Framework for Repository-Aware Code Translation
- Through the Lens of Human-Human Collaboration: A Configurable Research Platform for Exploring Human-Agent Collaboration
- Towards Transparent and Incentive-Compatible Collaboration in Decentralized LLM Multi-Agent Systems: A Blockchain-Driven Approach
- Debate or Vote: Which Yields Better Decisions in Multi-Agent Large Language Models?
- A Knowledge-driven Adaptive Collaboration of LLMs for Enhancing Medical Decision-making
- Aegis: Automated Error Generation and Attribution for Multi-Agent Systems
- AI Agents with Human-Like Collaborative Tools: Adaptive Strategies for Enhanced Problem-Solving
- From Language to Action: A Review of Large Language Models as Autonomous Agents and Tool Users
- Abduct, Act, Predict: Scaffolding Causal Inference for Automated Failure Attribution in Multi-Agent Systems
- Learning from Diverse Reasoning Paths with Routing and Collaboration
- PillagerBench: Benchmarking LLM-Based Agents in Competitive Minecraft Team Environments
- DRF: LLM-AGENT Dynamic Reputation Filtering Framework
- Orchestrator: Active Inference for Multi-Agent Systems in Long-Horizon Tasks
- LLM-Assisted Iterative Evolution with Swarm Intelligence Toward SuperBrain
- Scene-Aware Vectorized Memory Multi-Agent Framework with Cross-Modal Differentiated Quantization VLMs for Visually Impaired Assistance
- Building and Measuring Trust between Large Language Models
- FROGENT: An End-to-End Full-process Drug Design Agent
- Exploring Large Language Model Agents for Piloting Social Experiments
- DevNous: An LLM-Based Multi-Agent System for Grounding IT Project Management in Unstructured Conversation
- VirtLab: An AI-Powered System for Flexible, Customizable, and Large-scale Team Simulations
- DRAMA: A Dynamic and Robust Allocation-based Multi-Agent System for Changing Environments
- StackPilot: Autonomous Function Agents for Scalable and Environment-Free Code Execution
- Everyone Contributes! Incentivizing Strategic Cooperation in Multi-LLM Systems via Sequential Public Goods Games
- A Survey on Agent Workflow -- Status and Future
- DynaSwarm: Dynamically Graph Structure Selection for LLM-based Multi-agent System
- Multi-Agent-as-Judge: Aligning LLM-Agent-Based Automated Evaluation with Multi-Dimensional Human Evaluation
Related