GFlowNet Training by Policy Gradients
This work addresses the efficiency of GFlowNet training for generating combinatorial objects, offering incremental improvements in gradient estimation and backward policy design.
The authors tackled the problem of training Generative Flow Networks (GFlowNets) by introducing a new policy-based training framework that bridges flow balance with reinforcement learning rewards, resulting in improved performance verified on simulated and real-world datasets.
Generative Flow Networks (GFlowNets) have been shown effective to generate combinatorial objects with desired properties. We here propose a new GFlowNet training framework, with policy-dependent rewards, that bridges keeping flow balance of GFlowNets to optimizing the expected accumulated reward in traditional Reinforcement-Learning (RL). This enables the derivation of new policy-based GFlowNet training methods, in contrast to existing ones resembling value-based RL. It is known that the design of backward policies in GFlowNet training affects efficiency. We further develop a coupled training strategy that jointly solves GFlowNet forward policy training and backward policy design. Performance analysis is provided with a theoretical guarantee of our policy-based GFlowNet training. Experiments on both simulated and real-world datasets verify that our policy-based strategies provide advanced RL perspectives for robust gradient estimation to improve GFlowNet performance.