AIで化学反応を高精度予測する新システムを開発（An MIT-developed AI system provides realistic predictions for a wide variety of chemical reactions）

2025-09-03 マサチューセッツ工科大学（MIT）

The FlowER (Flow matching for Electron Redistribution) system allows a researcher to explicitly keep track of all the electrons in a reaction to ensure that none are spuriously added or deleted in the process of predicting the outcome of a chemical reaction.
Credits:Image courtesy of the researchers.

＜関連情報＞

• https://news.mit.edu/2025/generative-ai-approach-to-predicting-chemical-reactions-0903
• https://www.nature.com/articles/s41586-025-09426-9

生成反応機構予測のための電子流動マッチング Electron flow matching for generative reaction mechanism prediction

Joonyoung F. Joung,Mun Hong Fong,Nicholas Casetti,Jordan P. Liles,Ne S. Dassanayake & Connor W. Coley
Nature  Published:20 August 2025
DOI:https://doi.org/10.1038/s41586-025-09426-9

Abstract

Central to our understanding of chemical reactivity is the principle of mass conservation¹, which is fundamental for ensuring physical consistency, balancing equations and guiding reaction design. However, data-driven computational models^{2,3,4,5,6,7,8,9} for tasks such as reaction product prediction rarely abide by this most basic constraint^10,11,12,13. Here we recast the problem of reaction prediction as a problem of electron redistribution using the modern deep generative framework of flow matching^14,15,16, explicitly conserving both mass and electrons through the bond-electron (BE) matrix representation^17,18. Our model, FlowER, overcomes limitations inherent in previous approaches by enforcing exact mass conservation, resolving hallucinatory failure modes, recovering mechanistic reaction sequences for unseen substrate scaffolds and generalizing effectively to out-of-domain reaction classes with extremely data-efficient fine-tuning. FlowER also enables downstream estimation of thermodynamic or kinetic feasibility and manifests a degree of chemical intuition in reaction prediction tasks. This inherently interpretable framework represents an important step in bridging the gap between predictive accuracy and mechanistic understanding in data-driven reaction outcome prediction.