跳到正文
Sakana AI:Blog· Sakana AI·· 2026-10-11AI 評分55

Sakana AI 提出 MASS,探討多智能體自我監督下的遞迴自我改進

Recursive Self-Improvement through Collective Intelligence

AI 導讀

Sakana AI 發表新論文,提出多智能體自我監督(MASS),讓共享模型透過虛擬子智能體團隊解題並搜尋更佳工作流程。研究使用 27B 開放權重模型,在合成開放式研究任務上進行兩輪訓練;第二輪後,四項研究基準上的每輸出 token 得分達基礎模型的 1.2-1.6x。研究亦指出,只以被優化模型的解題數據訓練可提升優化器與評估器能力,多智能體軌跡也比單智能體軌跡更有效作為 RSI 訓練資料。

正文

As AI systems tackle increasingly open-ended research problems, human experts may struggle to assess results and guide further progress. The model itself may be the best available optimizer and evaluator. The challenge is obtaining useful self-supervision within that homogeneous loop, where a model risks reinforcing its own blind spots rather than correcting them.

We ask whether a model can learn from the collective intelligence of its own team.

Introducing our new paper, Recursive Self-Improvement through Multi-Agent Self-Supervision:

We introduce Multi-Agent Self-Supervision (MASS). One shared model solves tasks through a team of virtual subagents, searches for better multi-agent workflows, then uses its own judgments to select the best ones. Training on the best team’s executions distills the team’s collective experience back into the shared model.

We ran two cycles using a 27B open-weights model on synthetic open-ended research tasks. After the second cycle, score per output token reached 1.2-1.6x the base model’s level across four research benchmarks.

Three key results:

  1. MASS enables role generalization within an RSI loop: training only on task-solving data from the optimizee naturally improves the capabilities of both the optimizer and the evaluator.
  2. multi-agent trajectories are more efficient training data than single-agent trajectories for RSI.
  3. optimizer capacity, rather than evaluator capacity, can be the bottleneck in the speed of improvement.

We believe the community needs to investigate homogeneous RSI carefully to support AI safety. When a model is also its own evaluator, shared blind spots could allow mistakes to be accepted as progress and reinforced through training. For AI to improve beyond human expertise safely, we need to understand these failures and preserve effective ways to detect them and intervene.

This work was done by Hyunin Lee, Jinglue Xu, Jeffrey Seely, Yujin Tang, and our collaborators from UC Berkeley (Somayeh Sojoudi, Matei Zaharia, Donghyun Lee).

來源:Sakana AI:Blog · sakana.ai