KL DIVERGENCE OPTIMIZATION WITH ENTROPY- RATIO ESTIMATION FOR STOCHASTIC GFLOWNETS
Keywords: Stochastic environments, MDP, GFlowNets, KL divergence, Molecule Generation
TL;DR: This paper introduces a novel approach for optimizing Generative Flow Networks (GFlowNets) in stochastic environments by incorporating KL divergence objectives with entropy-ratio estimation.
Abstract: This paper introduces a novel approach for optimizing Generative Flow Networks (GFlowNets) in stochastic environments by incorporating KL divergence objectives with entropy-ratio estimation. We leverage the relationship between high and low entropy states, as defined in entropy-regularized Markov Decision Processes (MDPs), to dynamically adjust exploration and exploitation. Detailed proofs and analysis demonstrate the efficacy of this methodology in enhancing mode discovery, state coverage, and policy robustness in complex environments.
Primary Area: probabilistic methods (Bayesian methods, variational inference, sampling, UQ, etc.)
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.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
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: 8291
Loading