mlsanityData to 7 Oct 2026

Coarse-to-Fine: a Hierarchical Diffusion Model for Molecule Generation in 3D

arXiv 2305.13266 ↗First posted 5 May 2023q-bio.BM cs.AI cs.LGcited by 24

On the map: Diffusion models → Flow matching, time series and science

Abstract (arXiv)

Generating desirable molecular structures in 3D is a fundamental problem for drug discovery. Despite the considerable progress we have achieved, existing methods usually generate molecules in atom resolution and ignore intrinsic local structures such as rings, which leads to poor quality in generated structures, especially when generating large molecules. Fragment-based molecule generation is a promising strategy, however, it is nontrivial to be adapted for 3D non-autoregressive generations because of the combinational optimization problems. In this paper, we utilize a coarse-to-fine strategy to tackle this problem, in which a Hierarchical Diffusion-based model (i.e.~HierDiff) is proposed to preserve the validity of local segments without relying on autoregressive modeling. Specifically, HierDiff first generates coarse-grained molecule geometries via an equivariant diffusion process, where each coarse-grained node reflects a fragment in a molecule. Then the coarse-grained nodes are decoded into fine-grained fragments by a message-passing process and a newly designed iterative refined sampling module. Lastly, the fine-grained fragments are then assembled to derive a complete atomic molecular structure. Extensive experiments demonstrate that HierDiff consistently improves the quality of molecule generation over existing methods

Key facts. Cited by 24 papers in the map (data through 7 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 44 of its 65 references that are in the map. First posted on arXiv 5 May 2023.

Builds on

44 of 65 references.

Closest ancestors · 1

Same direction, newest first

  1. 2022GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation6 Mar

Other references · 43

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  3. 2021E(n) Equivariant Graph Neural Networkscited by 951
  4. 2018Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point cloudscited by 735
  5. 2018PointConv: Deep Convolutional Networks on 3D Point Cloudscited by 656
  6. 2018Junction Tree Variational Autoencoder for Molecular Graph Generationcited by 647
  7. 2021Score-based Generative Modeling in Latent Spacecited by 634
  8. 2017SchNet: A continuous-filter convolutional neural network for modeling quantum interactionscited by 569
  9. 2021Argmax Flows and Multinomial Diffusion: Learning Categorical Distributionscited by 566
  10. 2018MolGAN: An implicit generative model for small molecular graphscited by 392
  11. 2018GraphVAE: Towards Generation of Small Graphs Using Variational Autoencoderscited by 375
  12. 2018GuacaMol: Benchmarking Models for De Novo Molecular Designcited by 294
31 more
  1. 2017Grammar Variational Autoencodercited by 283
  2. 2022Torsional Diffusion for Molecular Conformer Generationcited by 262
  3. 2020Equivariant Flows: Exact Likelihood Generative Learning for Symmetric Densitiescited by 248
  4. 2020GEOM: Energy-annotated molecular conformations for property prediction and molecular generationcited by 238
  5. 2022Diffusion Probabilistic Modeling for Video Generationcited by 237
  6. 2018Learning Deep Generative Models of Graphscited by 227
  7. 2021Geometric and Physical Quantities Improve E(3) Equivariant Message Passingcited by 225
  8. 2020Hierarchical Generation of Molecular Graphs using Structural Motifscited by 204
  9. 2018Constrained Graph Variational Autoencoders for Molecule Designcited by 182
  10. 2022Pocket2Mol: Efficient Molecular Sampling Based on 3D Protein Pocketscited by 182
  11. 2019Symmetry-adapted generation of 3d point sets for the targeted discovery of moleculescited by 171
  12. 2021E(n) Equivariant Normalizing Flowscited by 165
  13. 2022A 3D Generative Model for Structure-Based Drug Designcited by 148
  14. 2021MARS: Markov Molecular Sampling for Multi-objective Drug Discoverycited by 125
  15. 2019Generalization in Generation: A closer look at Exposure Biascited by 123
  16. 2018Syntax-Directed Variational Autoencoder for Structured Datacited by 103
  17. 2018Multi-Objective De Novo Drug Design with Conditional Graph Generative Modelcited by 101
  18. 2018Language GANs Falling Shortcited by 97
  19. 2018Learning Multimodal Graph-to-Graph Translation for Molecular Optimizationcited by 87
  20. 2019MolecularRNN: Generating realistic molecular graphs with optimized propertiescited by 80
  21. 2021Hit and Lead Discovery with Explorative RL and Fragment-based Molecule Generationcited by 68
  22. 2020MIMOSA: Multi-constraint Molecule Sampling for Molecule Optimizationcited by 48
  23. 2023De Novo Molecular Generation via Connection-aware Motif Miningcited by 45
  24. 2022Relational Attention: Generalizing Transformers for Graph-Structured Taskscited by 41
  25. 2021HDMapGen: A Hierarchical Graph Generative Model of High Definition Mapscited by 40
  26. 2021Learning to design drug-like molecules in three-dimensional space using deep generative modelscited by 37
  27. 2021Top-N: Equivariant set and graph generation without exchangeabilitycited by 36
  28. 2019Exposure Bias versus Self-Recovery: Are Distortions Really Incremental for Autoregressive Text Generation?cited by 33
  29. 2021MolGrow: A Graph Normalizing Flow for Hierarchical Molecular Generationcited by 22
  30. 2021A Graph VAE and Graph Transformer Approach to Generating Molecular Graphscited by 16
  31. 2018Multiscale Planar Graph Generationcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 1; 2024: 8; 2025: 10; 2026: 5.

24 papers.

Sort:

By area:

  1. 2024A Survey of Geometric Graph Neural Networks: Data Structures, Models and Applicationscited by 28
  2. 2025Equivariant Neural Diffusion for Molecule Generationcited by 13
  3. 2024Geometry Informed Tokenization of Molecules for Language Model Generationcited by 10
  4. 2025Unified Guidance for Geometry-Conditioned Molecular Generationcited by 7
  5. 2025HOG-Diff: Higher-Order Guided Diffusion for Graph Generationcited by 6
  6. 2025FragFM: Hierarchical Framework for Efficient Molecule Generation via Fragment-Level Discrete Flow Matchingcited by 5
  7. 2023Leveraging Hyperbolic Embeddings for Coarse-to-Fine Robot Designcited by 4
  8. 2024Diffusion Twigs with Loop Guidance for Conditional Graph Generationcited by 4
  9. 2026SynHAT: A Two-stage Coarse-to-Fine Diffusion Framework for Synthesizing Human Activity Tracescited by 4
  10. 2024Structure Preserving Diffusion Modelscited by 3
  11. 2024A Review on Fragment-based De Novo 2D Molecule Generationcited by 3
  12. 2025Unraveling the Potential of Diffusion Models in Small Molecule Generationcited by 3
12 more
  1. 20263D Molecule Generation from Rigid Motifs via SE(3) Flowscited by 3
  2. 2026Elign: Equivariant Diffusion Model Alignment from Foundational Machine Learning Force Fieldscited by 3
  3. 2024Diffusion-based Graph Generative Methodscited by 2
  4. 2024Constraint Decoupled Latent Diffusion for Protein Backmappingcited by 2
  5. 2024Equivariant Blurring Diffusion for Hierarchical Molecular Conformer Generationcited by 2
  6. 2025Molecule Generation for Target Protein Binding with Hierarchical Consistency Diffusion Modelcited by 1
  7. 2025Predictive Feature Caching for Training-free Acceleration of Molecular Geometry Generationcited by 1
  8. 2025Concept-Driven Deep Learning for Enhanced Protein-Specific Molecular Generationcited by 0
  9. 2025Personalized Federated Training of Diffusion Models with Privacy Guaranteescited by 0
  10. 2025Feature-Aware (Hyper)graph Generation via Next-Scale Predictioncited by 0
  11. 2026Latent Diffusion-Based 3D Molecular Recovery from Vibrational Spectracited by 0
  12. 2026Coarse-to-Fine Multi-Resolution Diffusion Models for Trajectory Generation in Urban Systemscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Diffusion Models in textitDe Novo Drug Design
  2. 2022MDM: Molecular Diffusion Model for 3D Molecule Generation
  3. 2022Equivariant 3D-Conditional Diffusion Models for Molecular Linker Design
  4. 2026Equivariant Asynchronous Diffusion: An Adaptive Denoising Schedule for Accelerated Molecular Conformation Generation
  5. 2023Navigating the Design Space of Equivariant Diffusion-Based Generative Models for De Novo 3D Molecule Generation
  6. 2025D3MES: Diffusion Transformer with multihead equivariant self-attention for 3D molecule generation
  7. 2025DMol: A Highly Efficient and Chemical Motif-Preserving Molecule Generation Platform
  8. 2022DiffBP: Generative Diffusion of 3D Molecules for Target Protein Binding