← 研踪 PaperTrailarXiv:2601.16955 · v2
3D Molecule Generation from Rigid Motifs via SE(3) Flows
首次提交 2026-01-24 02:24
版本更新 2026-11-02 05:26
雷达观测 2026-11-04 23:42
摘要
Three-dimensional molecular structure generation is typically performed at the level of individual atoms, yet molecular graph generation techniques often consider fragments as their structural units. Building on the advances in frame-based protein structure generation, we extend these fragmentation ideas to 3D, treating general molecules as sets of rigid-body motifs. Utilising this representation, we employ SE(3)-equivariant generative modelling for de novo 3D molecule generation from rigid motifs. In our evaluations, we observe results competitive with state-of-the-art bond-implicit methods, surpassing them in atom stability on GEOM-Drugs, while yielding a 2x to 10x reduction in generation steps and 3.5x compression in molecular representations compared to standard atom-based methods. We release our code at https://github.com/rpoletukhin/MotiFlow.
作者备注：Accepted at NeurIPS 2026. 31 pages, 12 figures, 15 tables. Code: https://github.com/rpoletukhin/MotiFlow
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v2 · 2026-11-02 05:26 · 最新观测版本
3D Molecule Generation from Rigid Motifs via SE(3) Flows
Three-dimensional molecular structure generation is typically performed at the level of individual atoms, yet molecular graph generation techniques often consider fragments as their structural units. Building on the advances in frame-based protein structure generation, we extend these fragmentation ideas to 3D, treating general molecules as sets of rigid-body motifs. Utilising this representation, we employ SE(3)-equivariant generative modelling for de novo 3D molecule generation from rigid motifs. In our evaluations, we observe results competitive with state-of-the-art bond-implicit methods, surpassing them in atom stability on GEOM-Drugs, while yielding a 2x to 10x reduction in generation steps and 3.5x compression in molecular representations compared to standard atom-based methods. We release our code at https://github.com/rpoletukhin/MotiFlow.
查看此版本解读 →查看此版本原文 ↗