Zhipeng Chen, Yunsen Liang, Yanxi Chen, Hongkang Zhao, Lingyun Feng, Ruiyang Ren, Guoxin Chen, Jia Deng · 2026 · arXiv
DOI: 10.48550/arXiv.2610.29784 · open-access PDF
ほとんどの現代の人工知能(AI)システムは凍結されたパラメータでリリースされ、外部のアップデートを通じて改善される一方で、理想的なシステムは独自の経験から改善し続けるだろう、これは我々がリクルシブな自己改善(RSI)と呼ぶ能力である。しかし、RSIを研究するには、継続的に変化し、客観的に検証可能で、自然に基づく環境が必要であり、エージェントの進化にとって根本的な課題となる。
Machine translation (Meta M2M100); the English original follows.
Most contemporary artificial intelligence (AI) systems are released with frozen parameters and improve through external updates, whereas an ideal system would keep improving from its own experience, a capacity we call recursive self-improvement (RSI). Yet studying RSI requires an environment that is continuously changing, objectively verifiable, and naturally grounded, posing a fundamental challenge for agent evolution. Here we introduce DeepEvolve, a sustainable evolution agent system grounded in future event prediction, in which agents continuously interact with a dynamically evolving enviro
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.29784, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange