Research map / Probabilistic and scientific modeling / Molecules and proteins
Learning from Protein Structure with Geometric Vector Perceptrons
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 GVP-GNN instead of Structured GNN | protein sequence design · CATH 4.2 all | perplexity: 6.55 -> 5.29 | Helped |
| Use GVP-GNN instead of Structured GNN | protein sequence design · CATH 4.2 all | recovery (%): 37.3 -> 40.2 | Helped |
| Use GVP-GNN instead of GraphQA | model quality assessment · CASP 11 stage 2 | global correlation: 0.82 -> 0.87 | Helped |
| Use GVP-GNN instead of GraphQA | model quality assessment · CASP 12 stage 2 | global correlation: 0.81 -> 0.82 | Helped |
| Use GVP-GNN instead of ProQ3D | model quality assessment · CASP 13 stage 2 | global correlation: 0.847 -> 0.888 | Helped |
| Use GVP-GNN instead of ProQ4 | model quality assessment · CASP 13 stage 2 | per-target correlation: 0.691 -> 0.671 | Hurt |
| Add direct scalar-and-vector geometric propagation instead of an MLP layer (ablation) | GVP-GNN · CATH 4.2 all | perplexity: 7.76 -> 5.29 | Helped |
| Add both scalar and vector feature propagation instead of only scalar propagation (ablation) | GVP-GNN · CATH 4.2 all | recovery (%): 32.4 -> 40.2 | Helped |
| Add scalar feature propagation instead of propagating only vectors (ablation) | GVP-GNN · CASP 11 stage 2 | global correlation: 0.56 -> 0.87 | Helped |
| Add the second vector transformation W_mu (ablation) | GVP-GNN · CATH 4.2 all | perplexity: 5.85 -> 5.29 | Helped |
| Add the pairwise loss term | MQA training · model quality assessment | global correlation | Helped |
Closest papers by meaning
- Evaluating Representation Learning on the Protein Structure Universe (ICLR 2024 · similarity 0.90)
- Towards Multiscale Graph-based Protein Learning with Geometric Secondary Structural Motifs (NeurIPS 2025 · similarity 0.88)
- Clustering for Protein Representation Learning (CVPR 2024 · similarity 0.88)
- Protein 3D Graph Structure Learning for Robust Structure-Based Protein Property Prediction (AAAI 2024 · similarity 0.87)
- DualEqui: A Dual-Space Hierarchical Equivariant Network for Large Biomolecules (NeurIPS 2025 · similarity 0.87)
- Pushing the Limits of All-Atom Geometric Graph Neural Networks: Pre-Training, Scaling, and Zero-Shot Transfer (ICLR 2025 · similarity 0.87)
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.