3D Molecule Generation from Rigid Motifs via SE(3) Flows
cs.LG
Submitted: 2026-01-23
Updated: 2026-11-01
Comments: Accepted at NeurIPS 2026. 31 pages, 12 figures, 15 tables. Code: https://github.com/rpoletukhin/MotiFlow
License: http://creativecommons.org/licenses/by/4.0/
The gist: 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.
Terminology
Abstract
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.
Sources
- Building Normalizing Flows with Stochastic Interpolants
- SE(3)-Stochastic Flow Matching for Protein Backbone Generation
- Generative Flows on Discrete State-Spaces: Enabling Multimodal Flows with Applications to Protein Co-Design
- Scalable Autoregressive 3D Molecule Generation
- De Novo Molecular Generation via Connection-aware Motif Mining
- Bayesian Flow Networks
- Classifier-Free Diffusion Guidance
- SemlaFlow -- Efficient 3D Molecular Generation with Latent Attention and Equivariant Flow Matching
- Flow Matching for Generative Modeling
- Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
- MolDiff: Addressing the Atom-Bond Inconsistency Problem in 3D Molecule Diffusion Generation
- SigmaDock: Untwisting Molecular Docking With Fragment-Based SE(3) Diffusion
- Applications of Modular Co-Design for De Novo 3D Molecule Generation
- Simple Guidance Mechanisms for Discrete Diffusion Models
- Harmonic Self-Conditioned Flow Matching for Multi-Ligand Docking and Binding Site Design
- MiDi: Mixed Graph and 3D Denoising Diffusion for Molecule Generation
- SE(3) diffusion model with application to protein backbone generation
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis