Efficient Molecular Conformer Generation with SO(3) Averaged Flow-Matching and Reflow

Zhonglin Cao, Mario Geiger, Allan Dos Santos Costa, Danny Reidenbach, Karsten Kreis, Tomas Geffner, Franco Pellegrini, Guoqing Zhou, Emine Kucukbenli

Track: Machine learning: computational method and/or computational results

Nature Biotechnology: No

Keywords: Flow-matching, few-shot generation, equivariance, small molecules

TL;DR: We present a novel flow-matching objective and combine it with reflow and distillation for highly efficient molecular conformer generation

Abstract: Fast and accurate generation of molecular conformers is desired for downstream computational chemistry and drug discovery tasks. In this work, we propose two mechanisms for accelerating the training and inference of flow-based generative model for 3D molecular conformer generation. For fast training, we introduce the SO(3)-*Averaged Flow* training objective, which we show to converge faster and generate better conformer ensembles compared to conditional optimal transport and Kabsch alignment-based optimal transport flow. For fast inference, we demonstrate that reflow methods and distillation of these models enable few-steps or even one-step molecular conformer generation with high quality. Using these two techniques, we demonstrate a model that can match the performance of strong transformer baselines with only a fraction of the number of parameters and generation steps.

Anonymization: This submission has been anonymized for double-blind review via the removal of identifying information such as names, affiliations, and identifying URLs.

Presenter: ~Zhonglin_Cao1

Format: Yes, the presenting author will attend in person if this work is accepted to the workshop.

Funding: No, the presenting author of this submission does *not* fall under ICLR’s funding aims, or has sufficient alternate funding.

Submission Number: 21