Research map / Graph learning / Graph self-supervised learning
Molecule Generation by Principal Subgraph Mining and Assembling
Xiangzhe Kong, Wenbing Huang, Zhixing Tan, Yang Liu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add principal-subgraph vocabulary (vs single atoms) with two-step decoder (ablation) | PS-VAE · Property optimization, Penalized logP | top-1 score: 6.91 -> 13.95 | Improved |
| add two-step generation on top of atom-level vocabulary (ablation) | PS-VAE without principal subgraphs · Property optimization, Penalized logP | top-1 score: 3.54 -> 6.91 | Improved |
| add two-step generation on top of atom-level vocabulary (ablation) | PS-VAE without principal subgraphs · Constrained PlogP optimization delta=0.6 | success rate: 45.1% -> 27.0% | Hurt |
| add principal subgraphs (vs single atoms) with two-step decoder (ablation) | PS-VAE · Constrained PlogP optimization delta=0.2 | improvement: 2.33 -> 6.42 | Improved |
| use principal-subgraph vocabulary instead of HierVAE/FREED vocabularies in same two-step decoder | PS-VAE · GuacaMol goal-directed ZINC250K | Celecoxib rediscovery: 0.310 -> 0.451 | Improved |
| use principal-subgraph vocabulary instead of FREED vocabulary | two-step VAE · Constrained PlogP delta=0.2 | improvement: 4.82 -> 6.42 | Improved |
| use two-step assembling with HierVAE vocabulary (H-VAE*) instead of HierVAE sequential assembling | HierVAE · ZINC250K distribution learning | Uniqueness: 0.131 -> 0.991 | Improved |
| PS-VAE vs best atom-level baselines | Constrained optimization success rate · Constrained PlogP delta=0.6 | success rate: 90.3% vs 98.4% (GraphAF) | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GraphDF — GraphDF: A Discrete Flow Model for Molecular Graph Generation (ICML 2021) · better on 2 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Hierarchical Generation of Molecular Graphs using Structural Motifs (ICML 2020 · similarity 0.89)
- DMol: A Highly Efficient and Chemical Motif-Preserving Molecule Generation Platform (NeurIPS 2025 · similarity 0.88)
- SubgDiff: A Subgraph Diffusion Model to Improve Molecular Representation Learning (NeurIPS 2024 · similarity 0.87)
- Lift Your Molecules: Molecular Graph Generation in Latent Euclidean Space (ICLR 2025 · similarity 0.87)
- Pre-Training Graph Neural Networks on Molecules by Using Subgraph-Conditioned Graph Information Bottleneck (AAAI 2025 · similarity 0.86)
- Graph Generation with $K^2$-trees (ICLR 2024 · similarity 0.86)
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.
