Reproduce “Randomized Ensembled Double Q-Learning: Learning Fast Without a Model” 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
Xinyue Chen, Che Wang, Zijian Zhou, Keith Ross
REDQ uses model-free reinforcement learning to improve sample efficiency in continuous-control tasks without a learned dynamics model or imagined transitions. It performs Q-function updates on replayed real experience, using an ensemble of critics and a minimum over a randomly selected subset for its target. In MuJoCo tests, it broadly matched or exceeded MBPO, but not on every task or budget.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.