# Consistent Training via Energy-Based GFlowNets for Modeling Discrete Joint Distributions

arXiv: https://arxiv.org/abs/2211.00568 · HTML: https://mlsanity.com/paper/2211.00568/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 30 references that are in the map.
- First posted on arXiv 1 Nov 2022.

## Builds on (most cited first, 18 of 18)

- [Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation](https://mlsanity.com/paper/2106.04399/) (2021)
- [GFlowNet Foundations](https://mlsanity.com/paper/2111.09266/) (2021)
- [How to Train Your Energy-Based Models](https://mlsanity.com/paper/2101.03288/) (2021)
- [Trajectory balance: Improved credit assignment in GFlowNets](https://mlsanity.com/paper/2201.13259/) (2022)
- [Biological Sequence Design with GFlowNets](https://mlsanity.com/paper/2203.04115/) (2022)
- [Generative Flow Networks for Discrete Probabilistic Modeling](https://mlsanity.com/paper/2202.01361/) (2022)
- [Design-Bench: Benchmarks for Data-Driven Offline Model-Based Optimization](https://mlsanity.com/paper/2202.08450/) (2022)
- [Your GAN is Secretly an Energy-based Model and You Should use Discriminator Driven Latent Sampling](https://mlsanity.com/paper/2003.06060/) (2020)
- [Controllable and Compositional Generation with Latent-Space Energy-Based Models](https://mlsanity.com/paper/2110.10873/) (2021)
- [BOSS: Bayesian Optimization over String Spaces](https://mlsanity.com/paper/2010.00979/) (2020)
- [Active Learning for Domain Adaptation: An Energy-Based Approach](https://mlsanity.com/paper/2112.01406/) (2021)
- [Conditional molecular design with deep generative models](https://mlsanity.com/paper/1805.00108/) (2018)
- [Energy-Based Models for Continual Learning](https://mlsanity.com/paper/2011.12216/) (2020)
- [Joint Energy-based Model Training for Better Calibrated Natural Language Understanding Models](https://mlsanity.com/paper/2101.06829/) (2021)
- [Goal-directed Generation of Discrete Structures with Conditional Generative Models](https://mlsanity.com/paper/2010.02311/) (2020)
- [JointGAN: Multi-Domain Joint Distribution Learning with Generative Adversarial Nets](https://mlsanity.com/paper/1806.02978/) (2018)
- [Multivariate-Information Adversarial Ensemble for Scalable Joint Distribution Matching](https://mlsanity.com/paper/1907.03426/) (2019)
- [Directly Training Joint Energy-Based Models for Conditional Synthesis and Calibrated Prediction of Multi-Attribute Data](https://mlsanity.com/paper/2108.04227/) (2021)

## Built on by (most cited first, 4 of 4)

- [Bootstrapped Training of Score-Conditioned Generator for Offline Design of Biological Sequences](https://mlsanity.com/paper/2306.03111/) (2023)
- [Multi-Agent Continuous Control with Generative Flow Networks](https://mlsanity.com/paper/2408.06920/) (2024)
- [Dynamic Backtracking in GFlowNets: Enhancing Decision Steps with Reward-Dependent Adjustment Mechanisms](https://mlsanity.com/paper/2404.05576/) (2024)
- [Enhancing Solution Efficiency in Reinforcement Learning: Leveraging Sub-GFlowNet and Entropy Integration](https://mlsanity.com/paper/2410.00461/) (2024)

## Abstract (arXiv)

Generative Flow Networks (GFlowNets) have demonstrated significant performance improvements for generating diverse discrete objects x given a reward function R(x), indicating the utility of the object and trained independently from the GFlowNet by supervised learning to predict a desirable property y given x. We hypothesize that this can lead to incompatibility between the inductive optimization biases in training R and in training the GFlowNet, potentially leading to worse samples and slow adaptation to changes in the distribution. In this work, we build upon recent work on jointly learning energy-based models with GFlowNets and extend it to learn the joint over multiple variables, which we call Joint Energy-Based GFlowNets (JEBGFNs), such as peptide sequences and their antimicrobial activity. Joint learning of the energy-based model, used as a reward for the GFlowNet, can resolve the issues of incompatibility since both the reward function R and the GFlowNet sampler are trained jointly. We find that this joint training or joint energy-based formulation leads to significant improvements in generating anti-microbial peptides. As the training sequences arose out of evolutionary or artificial selection for high antibiotic activity, there is presumably some structure in the distribution of sequences that reveals information about the antibiotic activity. This results in an advantage to modeling their joint generatively vs. pure discriminative modeling. We also evaluate JEBGFN in an active learning setting for discovering anti-microbial peptides.