DROMO: Distributionally Robust Offline Model-based Policy Optimization
This addresses offline RL for safer policy learning, but it is incremental as it builds on existing regularization ideas.
The paper tackles offline reinforcement learning by proposing DROMO, a method that uses distributionally robust optimization to penalize out-of-distribution state-action pairs without uncertainty quantification, and theoretically shows it optimizes a lower bound on policy evaluation.
We consider the problem of offline reinforcement learning with model-based control, whose goal is to learn a dynamics model from the experience replay and obtain a pessimism-oriented agent under the learned model. Current model-based constraint includes explicit uncertainty penalty and implicit conservative regularization that pushes Q-values of out-of-distribution state-action pairs down and the in-distribution up. While the uncertainty estimation, on which the former relies on, can be loosely calibrated for complex dynamics, the latter performs slightly better. To extend the basic idea of regularization without uncertainty quantification, we propose distributionally robust offline model-based policy optimization (DROMO), which leverages the ideas in distributionally robust optimization to penalize a broader range of out-of-distribution state-action pairs beyond the standard empirical out-of-distribution Q-value minimization. We theoretically show that our method optimizes a lower bound on the ground-truth policy evaluation, and it can be incorporated into any existing policy gradient algorithms. We also analyze the theoretical properties of DROMO's linear and non-linear instantiations.