Pessimistic Causal Reinforcement Learning with Mediators for Confounded Offline Data

arXiv:2403.11841v1 Announce Type: cross
Abstract: In real-world scenarios, datasets collected from randomized experiments are often constrained by size, due to limitations in time and budget. As a result, leveraging large observational datasets becomes a more attractive option for achieving high-quality policy learning. However, most existing offline reinforcement learning (RL) methods depend on two key assumptions--unconfoundedness and positivity--which frequently do not hold in observational data contexts. Recognizing these challenges, we propose a novel policy learning algorithm, PESsimistic CAusal Learning (PESCAL). We …

abstract arxiv budget causal cs.ai cs.lg data datasets however limitations offline policy quality reinforcement reinforcement learning stat.ml type world