Model-Based Reinforcement Learning Under Confounding
It addresses a fundamental inconsistency in model-based RL for confounded settings, enabling principled learning and planning when contextual information is missing, which is incremental as it builds on existing frameworks like MaxCausalEnt.
The paper tackles model-based reinforcement learning in confounded environments where unobserved context biases offline data, by adapting a proximal off-policy evaluation method to identify reward expectations and integrating it with a behavior-averaged transition model to create a consistent surrogate MDP.
We investigate model-based reinforcement learning in contextual Markov decision processes (C-MDPs) in which the context is unobserved and induces confounding in the offline dataset. In such settings, conventional model-learning methods are fundamentally inconsistent, as the transition and reward mechanisms generated under a behavioral policy do not correspond to the interventional quantities required for evaluating a state-based policy. To address this issue, we adapt a proximal off-policy evaluation approach that identifies the confounded reward expectation using only observable state-action-reward trajectories under mild invertibility conditions on proxy variables. When combined with a behavior-averaged transition model, this construction yields a surrogate MDP whose Bellman operator is well defined and consistent for state-based policies, and which integrates seamlessly with the maximum causal entropy (MaxCausalEnt) model-learning framework. The proposed formulation enables principled model learning and planning in confounded environments where contextual information is unobserved, unavailable, or impractical to collect.