Cognitive Architectures for Language Agents
2023/09/05 by Theodore R. Sumers, Sumers, Theodore R., Shunyu Yao +5 · 4 voices · 68 citations
Computer Science · Social Sciences · #Topic Modeling #Natural Language Processing Techniques #Language and cultural evolution
paper · pdf · doi:10.48550/arxiv.2309.02427
Abstract
Recent efforts have augmented large language models (LLMs) with external resources (e.g., the Internet) or internal control flows (e.g., prompt chaining) for tasks requiring grounding or reasoning, leading to a new class of language agents. While these agents have achieved substantial empirical success, we lack a systematic framework to organize existing agents and plan future developments. In this paper, we draw on the rich history of cognitive science and symbolic artificial intelligence to propose Cognitive Architectures for Language Agents (CoALA). CoALA describes a language agent with modular memory components, a structured action space to interact with internal memory and external environments, and a generalized decision-making process to choose actions. We use CoALA to retrospectively survey and organize a large body of recent work, and prospectively identify actionable directions towards more capable agents. Taken together, CoALA contextualizes today's language agents within the broader history of AI and outlines a path towards language-based general intelligence.
Cited by
- PhoenixRepair: Rethinking Repair Strategy Exploration in Software Agents
- Supra Cognitive Modes: A Routed Architecture for Agent Memory
- Memory in the Loop: In-Process Retrieval as Extended Working Memory for Language Agents
- Beyond Memory Leaderboards: Evaluating Scientific Memory as Budgeted Context Restoration
- Digital Pantheon: Simulating and Auditing Coalition Formation with LLM Agents
- Binding Drift in Multi-Step Tool-Augmented Agents
- Self-Aware Recursively Self-Improving Agents for Personal Singularity: A Goal-, Scope-, Tool-, and Benchmark-Driven Multi-Agent Architecture
- Trajectory-Aware Retrieval Agents for Temporal Decision- Making
- Simulating Human Memory with Language Models
- Is Grep All You Need? How Agent Harnesses Reshape Agentic Search
- Reason to Play: Behavioral and Brain Alignment Between Frontier LRMs and Human Game Learners
- MAPS: Modeling Co-Existing Subjective Perspectives and Shared Meaning in Multi-Agent Cognitive Dialogue
- FlowEvo: Self-Evolving Agents through the Co-Evolution of Workflows and Executable Skills
- Ads in AI Chatbots? An Analysis of How Large Language Models Navigate Conflicts of Interest
- How Open Must Language Models be to Enable Reliable Scientific Inference?
- Position: Modular Memory is the Key to Continual Learning Agents
- SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks
- Future of Work with AI Agents: Auditing Automation and Augmentation Potential across the U.S. Workforce
- El Agente: An autonomous agent for quantum chemistry
- From Cognitive Architectures to Language Agents: A Mechanism-Level Review of Lineage, Convergence, and Migration Gaps
- SymStep: Symbolic Step Verification for Logical Reasoning
- Measuring and Improving Behavioral Consistency in Large Language Models through Fact-Heuristic-Emotion State Enforcement
- From Controlled to the Wild: Evaluation of Pentesting Agents for the Real-World
- El Agente Estructural: An Artificially Intelligent Molecular Editor
- Unifying Dynamic Tool Creation and Cross-Task Experience Sharing through Cognitive Memory Architecture
- Towards a Science of Scaling Agent Systems
- Learning Evolving Latent Strategies for Multi-Agent Language Systems without Model Fine-Tuning
- EWE: An Agentic Framework for Extreme Weather Analysis
- Episodic Memory in Agentic Frameworks: Suggesting Next Tasks
- Smarter Together: Creating Agentic Communities of Practice through Shared Experiential Learning
- AUTO-Explorer: Automated Data Collection for GUI Agent
- Completion ≠ Collaboration: Scaling Collaborative Effort with Agents
- Agentic AI: A Comprehensive Survey of Architectures, Applications, and Future Directions
- Cognitive Convergence: Deep Similarities Between Large Language Models and Human Cognition
- IFCMemoryBench: Evaluating Long-Term Memory of LLM-Based Agents in BIM Information Retrieval
- The Moltbook Files: A Harmless Slopocalypse or Humanity's Last Experiment
- Compiler.next: A Search-Based Compiler to Power the AI-Native Future of Software Engineering
- A Survey of Data Agents: Emerging Paradigm or Overstated Hype?
- Agentic Meta-Orchestrator for Multi-task Copilots
- Energy-Efficient Domain-Specific Artificial Intelligence Models and Agents: Pathways and Paradigms
- Co-Designing Quantum Codes with Transversal Diagonal Gates via Multi-Agent Systems
- Code-enabled language models can outperform reasoning models on diverse tasks
- Learning from Supervision with Semantic and Episodic Memory: A Reflective Approach to Agent Adaptation
- From Craft to Constitution: A Governance-First Paradigm for Principled Agent Engineering
- Sample-Efficient Online Learning in LM Agents via Hindsight Trajectory Rewriting
- How can we assess human-agent interactions? Case studies in software agent design
- Agentic Systems in Radiology: Design, Applications, Evaluation, and Challenges
- NL2GenSym: Natural Language to Generative Symbolic Rules for SOAR Cognitive Architecture via Large Language Models
- Q-Router: Agentic Video Quality Assessment with Expert Model Routing and Artifact Localization
- Agent Learning via Early Experience
- Opponent Shaping in LLM Agents
- A Qualitative Comparative Evaluation of Cognitive and Generative Theories
- AutoMaAS: Self-Evolving Multi-Agent Architecture Search for Large Language Models
- Dual-Scale World Models for LLM Agents Towards Hard-Exploration Problems
- LABELING COPILOT: A Deep Research Agent for Automated Data Curation in Computer Vision
- AutoMem: Automated Learning of Memory as a Cognitive Skill
- An LLM-based multi-agent framework for agile effort estimation
- Towards Fully Automated Molecular Simulations: Multi-Agent Framework for Simulation Setup and Force Field Extraction
- MCP-AgentBench: Evaluating Real-World Language Agent Performance with MCP-Mediated Tools
- Memento: Fine-tuning LLM Agents without Fine-tuning LLMs
- CyberSleuth: Autonomous Blue-Team LLM Agent for Web Attack Forensics
- Guidelines for Empirical Studies in Software Engineering involving Large Language Models
- Cognitive Agents Powered by Large Language Models for Agile Software Project Management
- Virtual Community: An Open World for Humans, Robots, and Society
- Designing Memory-Augmented AR Agents for Spatiotemporal Reasoning in Personalized Task Assistance
- Cognitive Duality for Adaptive Web Agents
- Sari Sandbox: A Virtual Retail Store Environment for Embodied AI Agents
- PilotRL: Training Language Model Agents via Global Planning-Guided Progressive Reinforcement Learning
Discussions
Related