AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Learning Generative Dynamics for 3D Molecule Generation via Sequential Neuro-Symbolic Constraints

Zhuoran Li, Xu Sun, Chang Chen, Wanyu Lin

NeurIPS 2026 · Image generation › Autoregressive and latent generation · approach: Equivariant diffusion for 3D molecules · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes sequential neuro-symbolic constraints integrated into diffusion dynamics to make generated 3D molecules chemically valid, not just statistically plausible.

Problem

Diffusion models for 3D molecule generation learn atomic distributions implicitly and can produce statistically plausible but chemically invalid molecules.

Approach

Explicitly encodes physicochemical rules as sequential neuro-symbolic constraints into the generative dynamics of a 3D molecular diffusion model, rather than learning them implicitly from data.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Enforce all four constraints simultaneously instead of sequentially (ablation)SensDiff · QM9molecule stability: 88.8% with one constraint -> 78.8% with all fourHurt
Apply constraints in the coarse-to-fine order (global geometry first, valence last) instead of other permutations (ablation)SensDiff · QM9validity: 74.6% to 90.9% across permutationsImproved
Activate fine-grained local rules early (ablation)SensDiff · QM9validityHurt
Use hierarchical weights (global geometry over local bonds/valence) rather than flat or inverted (ablation)SensDiff · QM9generation qualityImproved
Add linear ramp-up of newly introduced constraint weights (ablation)SensDiff · QM9training stabilityImproved
Add energy-adaptive weighting (ablation)SensDiff · QM9generation qualityImproved
Apply constraints during training rather than sampling only (ablation)SensDiff · QM9generation qualityImproved
Apply constraints more frequently during sampling (ablation)SensDiff · QM9stability and constraint satisfactionImproved
Add symbolic constraints to GEOLDM diffusion trainingGEOLDM · QM9molecule stability: 86.2 -> 95.6Improved
Add symbolic constraints to GEOLDM diffusion trainingGEOLDM · GEOM-Drugsatom stability: 80.7 -> 87.9Improved
Add symbolic constraints to GEOLDM diffusion trainingGEOLDM · QM9 conditionalalpha MAE: 2.46 -> 1.49Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy