Thompson sampling for improved exploration in GFlowNets

Jarrid Rector-Brooks*, Kanika Madan, Moksh Jain, Maksym Korablyov, Cheng-Hao Liu, Sarath Chandar, Nikolay Malkin, Yoshua Bengio

*Corresponding author for this work

Research output: Contribution to conferencePosterpeer-review

Abstract / Description of output

Generative flow networks (GFlowNets) are amortized variational inference algorithms that treat sampling from a distribution over compositional objects as a sequential decision-making problem with a learnable action policy. Unlike other algorithms for hierarchical sampling that optimize a variational bound, GFlowNet algorithms can stably run off-policy, which can be advantageous for discovering modes of the target distribution. Despite this flexibility in the choice of behaviour policy, the optimal way of efficiently selecting trajectories for training has not yet been systematically explored. In this paper, we view the choice of trajectories for training as an active learning problem and approach it using Bayesian techniques inspired by methods for multi-armed bandits. The proposed algorithm, Thompson sampling GFlowNets (TS-GFN), maintains an approximate posterior distribution over policies and samples trajectories from this posterior for training. We show in two domains that TS-GFN yields improved exploration and thus faster convergence to the target distribution than the off-policy exploration strategies used in past work.
Original languageEnglish
Pages1-8
Number of pages8
Publication statusPublished - 28 Jul 2023
EventStructured Probabilistic Inference and Generative Modeling Workshop -
Duration: 28 Jul 202328 Jul 2023

Workshop

WorkshopStructured Probabilistic Inference and Generative Modeling Workshop
Abbreviated titleSPIGM@ICML 2023
Period28/07/2328/07/23

Fingerprint

Dive into the research topics of 'Thompson sampling for improved exploration in GFlowNets'. Together they form a unique fingerprint.

Cite this