強化學習訓練常要研究人員盯著進度、找問題,再手動調整做法。
- 論文提出 AutoRL,讓 AI 助理檢視訓練曲線、測試過程和評估結果,找出問題並提出修改,再由另一個 AI 助理核准套用。文中提到,AutoRL 讓 Qwen3 在 AIME 2024 的成績從 46.67% 提升到 66.67%。
- 用 LongCat-Flash-Lite 做程式任務時,pass@1 從 52.29% 升到 65.13%。作者也指出,每次修改仍需要大量運算資源來驗證。
AI 伴隨問答
針對這篇報導有疑問?可直接點選快捷問題或向 AI 追問背景脈絡與技術細節:
Self-improving AI is one of the most discussed ideas right now.
This is one of the most exciting papers I've read recently on the topic.
(bookmark it)
It proposes AutoRL, which lifts Qwen3 on AIME 2024 from 46.67% to 66.67% by letting agents supervise a live RL run instead of following a fixed recipe.
AutoRL works like an RL researcher on call.
Perceptor agents watch training curves, rollouts, and evaluations and flag problems. Three Proposer agents diagnose the issue and suggest changes. An Orchestrator approves them and applies them with a hot update or a restart from a checkpoint.
On math, AIME 2025 and HMMT 2025 each improve by 11.67 points. On agentic coding with LongCat-Flash-Lite, pass@1 rises from 52.29% to 65.13% over a long unattended run with interruptions.
Given a deliberately broken recipe, AutoRL found and fixed the problems.
The authors say agents can already handle routine RL supervision. Validating each change still takes a lot of compute.
Chat with Paper: https://t.co/szm81Negyj