Self-retiring distillation improves reinforcement learning for multi-turn agents

RetireOPD uses a skill-conditioned teacher for dense supervision, then automatically retires it when the student is ready to continue with reinforcement learning alone.

Chinese Tech
Yan Yu · Zhengxi Lu · Yizhou Liu · Yichen Pan · Aozhe Wang · Qipeng Chen · +5 more

Zhejiang University · Alibaba Group

Research Digest··2 min read
The authors train a privileged teacher using environment rewards, then jointly train a skill-free student through reinforcement learning and token-level on-policy distillation.

The authors developed RetireOPD for multi-turn agent training, where reinforcement learning ordinarily provides only one scalar reward for an entire trajectory.

Why this paper

From Alibaba Group and Zhejiang University · Released code · Part of Credit Assignment in Agentic RL, now 17 papers

In one line

RetireOPD adaptively removes teacher supervision during agentic RL when it no longer benefits the student.

What it released

Code

What we could check

  • ✓Code link in the paper (github.com)
  • ·No weights link found
  • ·No dataset link found
  • ·No compute details found
  • ·No stated limitations found
  • ·No benchmark numbers found

Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.

§
newspaper

Research Digest

Articles published under the Zotpaper byline are synthesized from multiple source publications by our AI editor and reviewed by our editorial process. Each story combines reporting from credible outlets to give readers a balanced, comprehensive view.