Newest papers

Evolving with Open Worlds: Towards Recursive Self-Improvement for Future Forecasting

Zhipeng Chen, Yongli Liang, Yanxi Chen, Hongkang Zhao, Lingyun Feng, Ruiyang Ren, Guoxin Chen, Jia Deng, Feng Chang, Chuan Hao, Ran Tao, Jian Yang, Ji-Rong Wen, Bryan Dai, Wayne Xin Zhao · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.29784 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

La mayoría de los sistemas de inteligencia artificial (AI) de agentes vivos fuertes son liberados con parámetros fijos y mejorados a través de actualizaciones externas, mientras que un sistema ideal seguiría mejorando a partir de su propia experiencia, una capacidad que llamamos capacitación de auto mejora recursiva (RSI). No obstante, el estudio de RSI requiere un entorno que cambie continuamente, sea objetivamente verificable y esté naturalmente fundamentado, lo que representa un desafío fundamental para la evolución de los agentes. Aquí introducimos EvoForesight, un sistema de agentes de evolución sostenible basado en la predicción de eventos futuros, en el que los agentes interactúan continuamente con un entorno en evolución dinámica y mejoran internalizando y externalizando la experiencia adquirida a través de estas interacciones. Un entorno viviente sintetiza tareas

Machine translation (Meta M2M100); the English original follows.

Abstract

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 EvoForesight, a sustainable evolution agent system grounded in future event prediction, in which agents continuously interact with a dynamically evolving environment and improve by internalizing and externalizing the experience gained through these interactions. A living environment synthesizes verifiable tasks, and a recursive-evolution engine turns the reasoning trajectory behind each prediction back into capability: decoupling search from prediction adapts parameter-level evolution to the contamination-prone live environment, while distilling reusable skills lets the agent evolve from the limited live trajectories such an environment affords. On the external FutureX leaderboard, which evaluates predictions against real-world outcomes and is updated weekly, our self-improved forecaster ranked first across all systems over three consecutive weeks, achieving up to a 195% improvement in rating. On our synthesized evaluation, it ranked among the top three across four base LLMs, outperforming the strongest baseline by up to 10.7 Bradley--Terry points. More broadly, these results suggest that a system built for self-improvement is best understood as an agent that co-evolves with a genuinely living, open-ended environment, pointing to a shift in AI from training static models on fixed data toward cultivating agents that grow stronger from their own experience, with such an open world as the indispensable basis of that growth.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Artificial intelligence, Machine learning, Baseline (sea), Event (particle physics), Verifiable secret sharing

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange