Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ASI-Evolve: AI Accelerates AI

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

Abstract

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.

ASI-Evolve: AI Accelerates AI

Abstract

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.

Paper Structure

This paper contains 56 sections, 1 equation, 7 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Preliminary
  3. Scientific question answering.
  4. Structured task execution.
  5. Lightweight scientific discovery.
  6. Large-scale scientific exploration.
  7. ASI-Evolve
  8. Researcher
  9. Engineer
  10. Analyzer
  11. Cognition
  12. Database
  13. Main Tasks
  14. Scenario 1: Model Architecture Design
  15. Task Formulation
  16. ...and 41 more sections

Figures (7)

  • Figure 1: Representative scientific automation settings in the $L_\text{task}=\langle C_\text{exec}, S_\text{space}, D_\text{feedback}\rangle$ space.
  • Figure 2: Asi-Evolve pipeline: in each round, the system samples context nodes from database, retrieves relevant cognition items via embedding search, generates a new candidate program, runs an experiment-specific evaluation script under timeouts, and summarizes results into an analysis report that is stored back into the database for future rounds.
  • Figure 4: Sampling algorithm comparison. Evolution curves for ASI-Evolve with MAP-Elites, UCB1, and Random sampling on Qwen3-32B. Shaded regions show run-to-run variability.
  • Figure 5: Ablation on circle packing. Evolution curves for the full method and ablated variants. Shaded regions indicate variability across repeated runs.
  • Figure : (a) ASI-Evolve vs. GEPA vs. OpenEvolve.
  • ...and 2 more figures