arXiv Robotics — research abstracts· Anshul Nayak, Shahil Shaik, Jonathan Smereka, Yue Wang·· 16 hours agoEditorial score16
Reason-Guided Preference-Based Reinforcement Learning for Human-Aligned Policies
Reason-guided Human Reward Modeling for Preference-based Reinforcement Learning
Summary
This paper introduces a reason-guided PbRL framework that learns human reward models from trajectory feedback. It improves interpretability and aligns policies with human preferences and rationale.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
Source：arXiv Robotics — research abstracts · arxiv.org