Revisiting Non-Acyclic GFlowNets in Discrete Environments

Nikita Morozov, Ian Maksimov, Daniil Tiapkin, Sergey Samsonov
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:44887-44910, 2025.

Abstract

Generative Flow Networks (GFlowNets) are a family of generative models that learn to sample objects from a given probability distribution, potentially known up to a normalizing constant. Instead of working in the object space, GFlowNets proceed by sampling trajectories in an appropriately constructed directed acyclic graph environment, greatly relying on the acyclicity of the graph. In our paper, we revisit the theory that relaxes the acyclicity assumption and present a simpler theoretical framework for non-acyclic GFlowNets in discrete environments. Moreover, we provide various novel theoretical insights related to training with fixed backward policies, the nature of flow functions, and connections between entropy-regularized RL and non-acyclic GFlowNets, which naturally generalize the respective concepts and theoretical results from the acyclic setting. In addition, we experimentally re-examine the concept of loss stability in non-acyclic GFlowNet training, as well as validate our own theoretical findings.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-morozov25a, title = {Revisiting Non-Acyclic {GF}low{N}ets in Discrete Environments}, author = {Morozov, Nikita and Maksimov, Ian and Tiapkin, Daniil and Samsonov, Sergey}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {44887--44910}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/morozov25a/morozov25a.pdf}, url = {https://proceedings.mlr.press/v267/morozov25a.html}, abstract = {Generative Flow Networks (GFlowNets) are a family of generative models that learn to sample objects from a given probability distribution, potentially known up to a normalizing constant. Instead of working in the object space, GFlowNets proceed by sampling trajectories in an appropriately constructed directed acyclic graph environment, greatly relying on the acyclicity of the graph. In our paper, we revisit the theory that relaxes the acyclicity assumption and present a simpler theoretical framework for non-acyclic GFlowNets in discrete environments. Moreover, we provide various novel theoretical insights related to training with fixed backward policies, the nature of flow functions, and connections between entropy-regularized RL and non-acyclic GFlowNets, which naturally generalize the respective concepts and theoretical results from the acyclic setting. In addition, we experimentally re-examine the concept of loss stability in non-acyclic GFlowNet training, as well as validate our own theoretical findings.} }
Endnote
%0 Conference Paper %T Revisiting Non-Acyclic GFlowNets in Discrete Environments %A Nikita Morozov %A Ian Maksimov %A Daniil Tiapkin %A Sergey Samsonov %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-morozov25a %I PMLR %P 44887--44910 %U https://proceedings.mlr.press/v267/morozov25a.html %V 267 %X Generative Flow Networks (GFlowNets) are a family of generative models that learn to sample objects from a given probability distribution, potentially known up to a normalizing constant. Instead of working in the object space, GFlowNets proceed by sampling trajectories in an appropriately constructed directed acyclic graph environment, greatly relying on the acyclicity of the graph. In our paper, we revisit the theory that relaxes the acyclicity assumption and present a simpler theoretical framework for non-acyclic GFlowNets in discrete environments. Moreover, we provide various novel theoretical insights related to training with fixed backward policies, the nature of flow functions, and connections between entropy-regularized RL and non-acyclic GFlowNets, which naturally generalize the respective concepts and theoretical results from the acyclic setting. In addition, we experimentally re-examine the concept of loss stability in non-acyclic GFlowNet training, as well as validate our own theoretical findings.
APA
Morozov, N., Maksimov, I., Tiapkin, D. & Samsonov, S.. (2025). Revisiting Non-Acyclic GFlowNets in Discrete Environments. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:44887-44910 Available from https://proceedings.mlr.press/v267/morozov25a.html.

Related Material