ASI进化：人工智能驱动人工智能加速发展

ASI-Evolve: AI Accelerates AI

March 31, 2026

作者: Weixian Xu, Tiantian Mi, Yixiu Liu, Yang Nan, Zhimeng Zhou, Lyumanshan Ye, Lin Zhang, Yu Qiao, Pengfei Liu

cs.AI

摘要

人工智能能否加速其自身发展？尽管近期智能体系统已在反馈迅速的明确任务中展现出强大性能，但其能否应对推动真实AI进步所需的高成本、长周期、弱监督的研究闭环仍存疑问。我们提出ASI-Evolve——一个面向AI自我研究的智能体框架，通过"学习-设计-实验-分析"循环实现闭环研究。该框架通过两大核心组件增强标准进化智能体：注入人类先验知识的认知基库，以及将复杂实验结果提炼为可复用见解的专用分析器。据我们所知，ASI-Evolve是首个在AI开发三大核心维度（数据、架构、学习算法）均实现AI自主发现的统一框架。在神经架构搜索中，该系统发现105个线性注意力SOTA架构，最优模型较DeltaNet提升0.97个点，增益达近期人工设计改进的3倍；在预训练数据构建方面，进化流程使基准任务平均提升3.96个点，MMLU任务增益超18点；在强化学习算法设计中，新算法在AMC32、AIME24和OlympiadBench上分别较GRPO提升12.5、11.67和5.04个点。我们进一步通过数学与生物医学实验证明，这种AI自我研究范式可迁移至非AI领域。这些结果表明ASI-Evolve为实现AI在基础研发阶段自我加速迈出重要一步，为闭环AI研究可行性提供了早期实证。

English

Can AI accelerate the development of AI itself? While recent agentic systems have shown strong performance on well-scoped tasks with rapid feedback, it remains unclear whether they can tackle the costly, long-horizon, and weakly supervised research loops that drive real AI progress. We present ASI-Evolve, an agentic framework for AI-for-AI research that closes this loop through a learn-design-experiment-analyze cycle. ASI-Evolve augments standard evolutionary agents with two key components: a cognition base that injects accumulated human priors into each round of exploration, and a dedicated analyzer that distills complex experimental outcomes into reusable insights for future iterations. To our knowledge, ASI-Evolve is the first unified framework to demonstrate AI-driven discovery across three central components of AI development: data, architectures, and learning algorithms. In neural architecture design, it discovered 105 SOTA linear attention architectures, with the best discovered model surpassing DeltaNet by +0.97 points, nearly 3x the gain of recent human-designed improvements. In pretraining data curation, the evolved pipeline improves average benchmark performance by +3.96 points, with gains exceeding 18 points on MMLU. In reinforcement learning algorithm design, discovered algorithms outperform GRPO by up to +12.5 points on AMC32, +11.67 points on AIME24, and +5.04 points on OlympiadBench. We further provide initial evidence that this AI-for-AI paradigm can transfer beyond the AI stack through experiments in mathematics and biomedicine. Together, these results suggest that ASI-Evolve represents a promising step toward enabling AI to accelerate AI across the foundational stages of development, offering early evidence for the feasibility of closed-loop AI research.