**摘要**
Preference-based RL provides an approach to learning reward models from pairwise comparisons of behaviors, bypassing the need for explicit reward design. However, existing methods typically rely on passive data collection and suffer from poor sample efficiency, especially during the early stages of learning. We introduce a model-based approach that actively directs exploration by jointly reasoning
👤 作者: Mohamed Nabail, Leo Cheng, Jingmin Wang, Nicholas Rhinehart

---
🔗 **[UBP2: Uncertainty-Balanced Preference Planning for Efficient Preference-based Reinforcement Learning](https://arxiv.org/abs/2606.19328v1)**

> UBP2: Uncertainty-Balanced Preference Planning for Efficient Preference-based Reinforcement Learning
🏷️ 来源: ArXiv cs.AI
⏱️ 2026-06-18 14:00