Trajectory balance: Improved credit assignment in GFlowNets
cs.LG, stat.ML
Submitted: 2022-01-31
Updated: 2026-08-28
Comments: NeurIPS 2022; see footnotes for code
DOI: 10.52202/068431-0431
Code: https://github.com/GFNOrg/gflownet
License: http://creativecommons.org/licenses/by/4.0/
The gist: Generative flow networks (GFlowNets) are a method for learning a stochastic policy for generating compositional objects, such as graphs or strings, from a given unnormalized density by sequences of
Terminology
Abstract
Generative flow networks (GFlowNets) are a method for learning a stochastic policy for generating compositional objects, such as graphs or strings, from a given unnormalized density by sequences of actions, where many possible action sequences may lead to the same object. We find previously proposed learning objectives for GFlowNets, flow matching and detailed balance, which are analogous to temporal difference learning, to be prone to inefficient credit propagation across long action sequences. We thus propose a new learning objective for GFlowNets, trajectory balance, as a more efficient alternative to previously used objectives. We prove that any global minimizer of the trajectory balance objective can define a policy that samples exactly from the target distribution. In experiments on four distinct domains, we empirically demonstrate the benefits of the trajectory balance objective for GFlowNet convergence, diversity of generated samples, and robustness to long action sequences and large action spaces.
Sources
- GFlowNet Foundations
- Soft Actor-Critic for Discrete Action Settings
- Learning GFlowNets from partial episodes for improved convergence and stability
- GFlowNets and variational inference
- Deep Reinforcement Learning and the Deadly Triad
- Unifying Generative Models with GFlowNets and Beyond
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks