Model-based Offline Reinforcement Learning with Count-based Conservatism
This work addresses the challenge of data-efficient and stable learning in offline RL for applications like robotics and autonomous systems, representing an incremental improvement with a novel integration of count-based techniques.
The paper tackles the problem of offline reinforcement learning by proposing a model-based method with count-based conservatism, named Count-MORL, which uses state-action pair frequencies to quantify model error and achieves near-optimal performance guarantees, significantly outperforming existing algorithms on D4RL benchmarks.
In this paper, we propose a model-based offline reinforcement learning method that integrates count-based conservatism, named $\texttt{Count-MORL}$. Our method utilizes the count estimates of state-action pairs to quantify model estimation error, marking the first algorithm of demonstrating the efficacy of count-based conservatism in model-based offline deep RL to the best of our knowledge. For our proposed method, we first show that the estimation error is inversely proportional to the frequency of state-action pairs. Secondly, we demonstrate that the learned policy under the count-based conservative model offers near-optimality performance guarantees. Through extensive numerical experiments, we validate that $\texttt{Count-MORL}$ with hash code implementation significantly outperforms existing offline RL algorithms on the D4RL benchmark datasets. The code is accessible at $\href{https://github.com/oh-lab/Count-MORL}{https://github.com/oh-lab/Count-MORL}$.