{"id":"arxiv:2512.18746","title":"MemEvolve: Meta-Evolution of Agent Memory Systems","date":"2025-12-21","section":"Agents, open-endedness, AGI","quality":0.399186,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/","source_url":"https://arxiv.org/abs/2512.18746","text_available":true,"authors":["Guibin Zhang","Haotian Ren","Chong Zhan","Zhenhong Zhou"],"outgoing":[{"source":"arxiv:2512.18746","target":"arxiv:2505.22954","evidence":"title","id":"arxiv:2505.22954","title":"Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents","url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c29f716871f80314868b5647692de387f26e12166b93bd362915fd272b6e081c/"}],"incoming":[{"source":"arxiv:2602.07755","target":"arxiv:2512.18746","evidence":"id","id":"arxiv:2602.07755","title":"Learning to Continually Learn via Meta-learning Agentic Memory Designs","url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/4d0a0ea160495ea50583e0e24027b6085e2f430a44195e3053ba69cda8cc82ea/"},{"source":"arxiv:2603.28052","target":"arxiv:2512.18746","evidence":"id","id":"arxiv:2603.28052","title":"Meta-Harness: End-to-End Optimization of Model Harnesses","url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/8bb0b87384dfdd42c419cf26cc7e52f1f0d89f63afa5d5968e246b73a44c99cb/"},{"source":"arxiv:2604.08224","target":"arxiv:2512.18746","evidence":"id","id":"arxiv:2604.08224","title":"Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering","url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/392317a8b3bedec8258ff8288d69ac72bc9e60299c5f7bec8962838bb448c2ce/"},{"source":"arxiv:2605.13821","target":"arxiv:2512.18746","evidence":"id","id":"arxiv:2605.13821","title":"Harnessing Agentic Evolution","url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/1f1060a507e2bf439d2934d610827d1b3177d85c90c1a2cab483608a56de5138/"}],"text_parts":[{"start":0,"end":15974,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-1.html"},{"start":15974,"end":31214,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-2.html"},{"start":31214,"end":40185,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-3.html"}],"contents":[{"title":"MemEvolve: Meta-Evolution of Agent Memory Systems","offset":59,"level":1,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-1.html#text"},{"title":"Abstract","offset":112,"level":6,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-1.html#text"},{"title":"1 Introduction","offset":1722,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-1.html#text"},{"title":"2 Related Work","offset":8961,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-1.html#text"},{"title":"3 EvolveLab: A Unified Codebase for Self-Evolving Memory","offset":11532,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-1.html#text"},{"title":"3.1 Preliminary","offset":11900,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-1.html#text"},{"title":"3.2 Modular Design Space of Memory Systems","offset":13722,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-1.html#text"},{"title":"3.3 EvolveLab Codebase","offset":15946,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-1.html#text"},{"title":"4 MemEvolve: A Meta-Evolving Memory Framework","offset":17569,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-2.html#text"},{"title":"4.1 Dual-Evolution Process","offset":17619,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-2.html#text"},{"title":"4.2 Diagnose-and-Design Evolution","offset":22991,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-2.html#text"},{"title":"5 Experiments","offset":27407,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-2.html#text"},{"title":"5.1 Experiment Setup","offset":27425,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-2.html#text"},{"title":"5.2 Main Results","offset":29677,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-2.html#text"},{"title":"5.3 Self-Evolving Memory Comparison","offset":32546,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-3.html#text"},{"title":"5.4 Meta-Evolving Dynamics","offset":35787,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-3.html#text"},{"title":"6 Conclusion","offset":38749,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-3.html#text"},{"title":"7 Contributions","offset":39816,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/part-3.html#text"}],"text_code_points":40185,"source_markdown":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c2161237fe9d9768afaa2880a40019455779801af5fba05be5781c229c18ea86/source.md"}