AlignIQL: Policy Alignment in Implicit Q-Learning through Constrained Optimization

Longxiang He, Li Shen, Junbo Tan, Xueqian Wang

Keywords: Offline reinforcement learning, optimization, Implict Q learning, diffusion model

TL;DR: We introduce a new method (AlignIQL) to extract the policy from the IQL-style value function and explain when IQL can utilize weighted regression for policy extraction.

Abstract: Implicit Q-learning (IQL) serves as a strong baseline for offline RL, which never needs to evaluate actions outside of the dataset through quantile regression. However, it is unclear how to recover the implicit policy from the learned implicit Q-function and whether IQL can utilize weighted regression for policy extraction. IDQL reinterprets IQL as an actor-critic method and gets weights of implicit policy, however, this weight only holds for the optimal value function under certain critic loss functions. In this work, we introduce a different way to solve the $\textit{implicit policy-finding problem}$ (IPF) by formulating this problem as an optimization problem. Based on this optimization problem, we further propose two practical algorithms AlignIQL and AlignIQL-hard, which inherit the advantages of decoupling actor from critic in IQL and provide insights into why IQL can use weighted regression for policy extraction. Compared with IQL and IDQL, we find that our method keeps the simplicity of IQL and solves the implicit policy-finding problem. Experimental results on D4RL datasets show that our method achieves competitive or superior results compared with other SOTA offline RL methods. Especially in complex sparse reward tasks like AntMaze and Adroit, our method outperforms IQL and IDQL by a significant margin.

Primary Area: reinforcement learning

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 1875