Tag
The paper introduces the Generative Value Classifier (GVC), which reframes value learning by modeling a return-conditioned action policy and return prior to derive Q(s,a) via Bayes' rule, showing gains over discriminative and distributional RL baselines on offline and online tasks.
This paper introduces Deep-BQRL, a model-free distributional reinforcement learning framework that extends buffered-quantile learning to neural function approximation for risk-sensitive decision-making, showing improved performance over existing methods in specific experiments.
This paper introduces RQIQN, a robust quantile-based method for distributional reinforcement learning that uses Wasserstein geometry regularization to prevent distribution degeneration and improve performance in risk-sensitive tasks.
This paper introduces C-DSAC, a new distributional reinforcement learning algorithm that uses the Cramér distance to improve performance and stability in robotic benchmarks compared to standard SAC.
This paper presents a finite-iteration theory for asynchronous categorical distributional temporal-difference learning, bridging the gap between existing theoretical frameworks and practical online implementations.