GFlowNets for Learning Better Drug-Drug Interaction Representations

ICLR 2025 Workshop LMRL Submission112 Authors

Track: Tiny Paper Track

Keywords: DDI Prediction, Class Imbalance, GFlowNet, Variational Autoencoder, Synthetic Data Generation

TL;DR: We propose a GFlowNet-VAE framework to address class imbalance in DDI prediction by generating synthetic samples for rare interaction types, improving model performance and clinical reliability.

Abstract: Drug–drug interactions pose a significant challenge in clinical pharmacology, with severe class imbalance among interaction types limiting the effectiveness of predictive models. Common interactions dominate datasets, while rare but critical interactions remain underrepresented, leading to poor model performance on infrequent cases. Existing methods often treat DDI prediction as a binary problem, ignoring class-specific nuances and exacerbating bias toward frequent interactions. To address this, we propose a framework combining Generative Flow Networks (GFlowNet) with Variational Graph Autoencoders (VGAE) to generate synthetic samples for rare classes, improving model balance and generate effective and novel DDI pairs. Our approach enhances predictive performance across interaction types, ensuring better clinical reliability.

Submission Number: 112