Revisiting Non-Acyclic GFlowNets in Discrete Environments

TL;DR

This paper extends the theory of GFlowNets to non-acyclic discrete environments, providing a simpler framework, new theoretical insights, and experimental validation of loss stability.

Contribution

It introduces a simplified theoretical framework for non-acyclic GFlowNets and explores their connections to entropy-regularized RL and flow functions.

Findings

Theoretical insights on training with fixed backward policies

Connections between entropy-regularized RL and non-acyclic GFlowNets

Experimental validation of loss stability in training

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.