arXiv Machine Learning

UNIQ: Conformal Calibration for Adaptive Conservatism in Offline Reinforcement Learning

arXiv:2606. 07592v1 Announce Type: new Abstract: Offline reinforcement learning requires careful conservatism to mitigate distribution shift, yet most existing methods apply a fixed penalty uniformly across all states regardless of local data coverage.

Hugging Face Trending Papers
Jul 29

Do You Really Need to Pretrain Q-Functions for Online RL Fine-Tuning?

Pre-training followed by fine-tuning has become the dominant recipe for learning performant policies, and in value-based reinforcement learning (RL) this raises a natural question: given a pretrained policy, should the Q-function be pretrained on offline data too? Conventional wisdom suggests it should, but recent results show that online RL with a randomly-initialized Q-function can result in highly performant and reliable policies without needing to pretrain the Q-function.

arXiv Machine Learning
Jul 30

Do You Really Need to Pretrain Q-Functions for Online RL Fine-Tuning?

arXiv:2607. 27203v1 Announce Type: new Abstract: Pre-training followed by fine-tuning has become the dominant recipe for learning performant policies, and in value-based reinforcement learning (RL) this raises a natural question: given a pretrained policy, should the Q-function be pretrained on offline data too?

By Perry Dong, Ron Polonsky, Dorsa Sadigh, Chelsea Fin
arXiv AI
Jun 29

Uncertainty-Aware Reward Discounting for Mitigating Reward Hacking

arXiv:2604. 26360v2 Announce Type: replace-cross Abstract: Reinforcement learning from human feedback (RLHF) systems face a compounding alignment challenge: not only are learned reward models uncertain about unseen state-action pairs, but the human preference annotations they are trained on are themselves inconsistent, context-dependent, and noisy.

By Disha Singha
arXiv Machine Learning
Aug 4

Diffusion Policy with Behavioral Advantage Correction for Offline Reinforcement Learning

arXiv:2608. 02332v1 Announce Type: new Abstract: In offline reinforcement learning (RL), the distribution shift between behavioral data and the learned policy can lead to erroneous \emph{Q}-value estimation, thereby misguiding the direction of policy optimization.

By Botao Dong, Longyang Huang, Ning Pang, Hongtian Chen