Reproduce “CLARE: Conservative Model-Based Reward Learning for Offline Inverse Reinforcement Learning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Sheng Yue, Guanbo Wang, Wei Shao, Zhaofeng Zhang, Sen Lin, Ju Ren, Junshan Zhang
CLARE presents offline inverse reinforcement learning using joint reward learning and policy improvement with estimated dynamics. It weights expert and diverse offline data and penalizes model-generated rollouts. On D4RL Walker2d with 5,000 expert and 5,000 random tuples, its seven-seed average return was 2,873.8, versus IQ-Learn’s 256.9; this benchmark comparison does not directly measure reward extrapolation error.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.