method · first seen 11 Oct 2026
one-step generative policy
A reinforcement learning policy that produces actions in a single network evaluation instead of iterative diffusion sampling.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Drift, Then Spread: Learning One-Step Generative Policies for Online Reinforcement LearningPapers · 11 Oct 2026 · Chunyu Luo, Hongyu Zhao, Guangyi Ni, Yue Song, Lingkai Kong
Latest sources
- Drift, Then Spread: Learning One-Step Generative Policies for Online Reinforcement LearningPapers · 11 Oct 2026 · Chunyu Luo, Hongyu Zhao, Guangyi Ni, Yue Song, Lingkai Kong