multi-agent systems
13 mentions across all digests
Multi-agent systems are computational frameworks in which multiple autonomous AI agents collaborate or interact to accomplish tasks such as research discovery, paper writing, or complex knowledge work, with emerging research into how value misalignment shapes their collective behaviors.
From Skills to Talent: Organising Heterogeneous Agents as a Real-World Company
Heterogeneous AI agents can be effectively coordinated by organizing them with company-like role structures and skill-to-position mappings.
Human Values Matter: Investigating How Misalignment Shapes Collective Behaviors in LLM Agent Communities
Misaligned LLM agents in multi-agent systems develop emergent collective behaviors that diverge from human values, revealing new coordination-based safety risks.
Paper Circle: An Open-source Multi-agent Research Discovery and Analysis Framework
Open-source framework uses multiple AI agents to automate research paper discovery, retrieval, and analysis, reducing manual literature review overhead.
PaperOrchestra: A Multi-Agent Framework for Automated AI Research Paper Writing
Multi-agent orchestration framework automates research paper generation by coordinating specialized agents through synthesis, drafting, and revision stages—demonstrating that coordinated AI can handle complex end-to-end knowledge work.
MMORF: A Multi-agent Framework for Designing Multi-objective Retrosynthesis Planning Systems
Multi-agent framework optimizes competing objectives in drug synthesis routes—balancing yield, cost, and feasibility through coordinated AI reasoning.