Research map / Probabilistic and scientific modeling / Molecules and proteins
Molecular Hypergraph Grammar with Its Application to Molecular Optimization
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use MHG parse sequences with a single VAE instead of a fragment-tree representation requiring auxiliary neural networks | Molecular optimization on ZINC · ZINC test | reconstruction rate: 76.7% -> 94.8% | Helped |
| Use MHG-VAE rather than JT-VAE | Global molecular optimization · Limited oracle | Top 50 average: JT-VAE: -1.33; Ours: 4.53 | Helped |
| Use MHG-VAE rather than GCPN | Global molecular optimization · Limited oracle | Top 50 average: 1.36 -> 4.53 | Helped |
| Use MHG-VAE rather than JT-VAE | Local molecular optimization · δ=0.2 | improvement: 1.68 ± 1.85 -> 2.40 ± 2.16 | Helped |
| Use MHG-VAE rather than JT-VAE | Local molecular optimization · δ=0.4 | improvement: 0.84 ± 1.45 -> 1.00 ± 1.87 | Helped |
| Use MHG-VAE rather than JT-VAE | Local molecular optimization · δ=0.6 | improvement: 0.21 ± 0.71 -> 0.61 ± 1.20 | Helped |
| Use MHG-VAE rather than GCPN | Local molecular optimization · δ=0.0 | improvement: 4.20 ± 1.28 -> 3.28 ± 2.19 | Hurt |
Closest papers by meaning
- Constrained Graph Variational Autoencoders for Molecule Design (NeurIPS 2018 · similarity 0.90)
- Reinforced Molecular Optimization with Neighborhood-Controlled Grammars (NeurIPS 2020 · similarity 0.90)
- Foundation Molecular Grammar: Multi-Modal Foundation Models Induce Interpretable Molecular Graph Languages (ICML 2025 · similarity 0.87)
- Molecule Design by Latent Prompt Transformer (NeurIPS 2024 · similarity 0.86)
- Domain-Agnostic Molecular Generation with Chemical Feedback (ICLR 2024 · similarity 0.86)
- MIPT: Multilevel Informed Prompt Tuning for Robust Molecular Property Prediction (ICML 2025 · similarity 0.85)
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.