Geometric Representation Condition Improves Equivariant Molecule Generation

Z. Li, C. Zhou, X. Wang, X. Peng,
May 2025

Abstract

This work introduces geometric representation conditions that significantly improve equivariant molecule generation, achieving state-of-the-art performance in molecular design tasks.

Type
Conference paper
Publication
Proc. International Conference on Machine Learning (ICML-25), spotlight presentation (2.6% acceptance rate)
Molecule Generation Equivariant Networks Geometric Deep Learning Drug Discovery