Research map / Multi-agent systems and games / LLM agents
MAGE: Multi-Agent Self-Evolution with Co-Evolutionary Knowledge Graphs
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes MAGE, a multi-agent self-evolution framework that stores cross-iteration knowledge in a co-evolving knowledge graph and uses graph/bandit guidance to improve a frozen weak backbone at inference time.
Problem
Self-evolving LM agents must decide what to learn next and how to preserve learned knowledge across iterations; existing approaches use natural-language feedback, flat episodic memory, or implicit RL signals, which do not cleanly support a frozen weak backbone at inference.
Approach
Carries cross-iteration knowledge in a co-evolutionary knowledge graph with append-only protected memory, and uses a graph-guided bandit to choose what to learn next, keeping the backbone frozen.
Key results
- Abstract-level claim of strong performance across seven benchmarks; specific numbers not present in the provided excerpt.
Limitations
- Small gains on open-domain factoid QA
- Reliance on a strong guidance tier
- Uneven cross-backbone behavior
Experimental setup
Benchmarks: seven benchmarks (names not given in provided text)
Metrics: accuracy
Closest papers by meaning
- MemRL: Self-Evolving Agents via Runtime Reinforcement Learning on Episodic Memory (NeurIPS 2026 · similarity 0.90)
- ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory (ICLR 2026 · similarity 0.89)
- EvoMemBench: Benchmarking Agent Memory from a Self-Evolving Perspective (NeurIPS 2026 · similarity 0.89)
- MemEvolve: Meta-Evolution of Agent Memory Systems (ICML 2026 · similarity 0.89)
- CoMAS: Co-Evolving Multi-Agent Systems via Interaction Rewards (ICLR 2026 · similarity 0.89)
- PACE: Two-Timescale Self-Evolution for Small Language Model Agents (NeurIPS 2026 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.