Graph Forward Distribution Matching for Molecular Inverse Design

2026-09-25

AI Summary

This study addresses the challenge of simultaneously achieving precise multi-attribute control and chemical validity in inverse molecular design by proposing the GraphFDM framework. This method introduces a novel online reinforcement learning paradigm based on forward distribution matching, which integrates reinforcement signals into the supervised training of graph diffusion models. By eliminating the need to store reverse trajectories, it enables joint optimization of molecular size and structure. Experimental results demonstrate that GraphFDM reduces the mean absolute error across target attributes by 53% while maintaining a chemical validity exceeding 0.99. Furthermore, the framework exhibits strong generalization capabilities. Overall, this work provides an efficient new approach for controllable molecular generation.

Abstract

Achieving precise control over multiple properties without sacrificing chemical validity remains a central challenge in molecular inverse design. Existing reinforcement learning (RL) methods fine-tune graph diffusion models by treating **reverse** sampling as a sequential policy, using a single terminal reward to optimize hundreds of coupled decisions. They often suffer from instability, validity collapse, and limited property gains. We introduce GraphFDM (Graph Forward Distribution Matching), a new online RL paradigm for graph diffusion that performs optimization through the **forward** process. GraphFDM uses valid generations to define a reward-tilted target distribution jointly optimized over graph size and molecular structure for each property condition, incorporating reinforcement signals into supervised learning without storing reverse trajectories. We derive the unique optimal target, prove a condition-wise improvement guarantee, and show that the fixed graph-size prior of standard graph diffusion leaves an irreducible matching gap. In multi-conditional polymer and small-molecule generation, GraphFDM achieves the lowest MAE on every target property, with reductions of up to 53.0\% relative to the strongest baselines and chemical validity above 0.99. It further generalizes to out-of-distribution property combinations.

Problem

• Molecular Inverse Design
• Multi-property Control
• Chemical Validity
• Graph Diffusion Models
• Reinforcement Learning

Innovation

• Graph Forward Distribution Matching
• Molecular Inverse Design
• Graph Diffusion Models
• Reinforcement Learning
• Multi-property Optimization