AI RESEARCH
From Static Constraints to Dynamic Adaptation: Sample-Level Constraint Release for Offline-to-Online Reinforcement Learning
arXiv CS.LG
•
ArXi:2511.03828v2 Announce Type: replace Offline-to-online reinforcement learning (O2O RL) faces a central challenge between retaining offline conservatism and adapting to online feedback under distribution shift. This challenge arises because data behavior evolves during fine-tuning, rendering data origin a misleading basis for constraint handling and thereby leading to objective-data mismatch. We therefore propose Dynamic Alignment for RElease (DARE), a distribution-aware framework for sample-level constraint release based on the behavioral consistency with a behavior model.