method · first seen 20 Oct 2026
preference-conditioned reinforcement learning
Reinforcement learning where a single policy is conditioned on user preference weights to adjust behavior style at runtime.
Mentions per week
PapersCodeDiscussion
Earliest sources
- One Policy, Many Styles: Where to Aggregate Preferences for Steerable Humanoid Control?Papers · 20 Oct 2026 · Ziqian Huang, Yuzhuo Chen, Shimin Zhang, Junyang He, Xinyue Annie Yang, Animesh…
Latest sources
- One Policy, Many Styles: Where to Aggregate Preferences for Steerable Humanoid Control?Papers · 20 Oct 2026 · Ziqian Huang, Yuzhuo Chen, Shimin Zhang, Junyang He, Xinyue Annie Yang, Animesh…