EXG: Self-Evolving Agents with Experience Graphs
The paper introduces EXG, an experience graph framework designed for self-evolving agents. This framework allows agents to organize their experiences in a structured manner, facilitating both immediate and offline reuse of knowledge. Experimental results indicate that EXG improves performance and efficiency compared to existing methods.
- ▪EXG is the first experience graph specifically designed for self-evolving agents.
- ▪The framework supports real-time graph growth during execution for immediate experience reuse.
- ▪Experiments show that EXG achieves better performance-efficiency trade-offs than traditional reflection and memory-based approaches.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.17721 (cs) [Submitted on 18 May 2026] Title:EXG: Self-Evolving Agents with Experience Graphs Authors:Yuxin Jin, Siyuan Zhang, Hanchen Wang, Lu Qin, Ying Zhang, Wenjie Zhang View a PDF of the paper titled EXG: Self-Evolving Agents with Experience Graphs, by Yuxin Jin and 5 other authors View PDF HTML (experimental) Abstract:Large language model (LLM)-based agents have demonstrated strong capabilities in complex reasoning and problem solving through multi-step interactions, yet most deployed agents remain behaviorally static, with knowledge acquired during execution rarely translating into systematic improvement over time.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
