Research map / Probabilistic and scientific modeling / Molecules and proteins
SchNet: A continuous-filter convolutional neural network for modeling quantum interactions
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 SchNet instead of DTNN | QM9 energy prediction · QM9, N=50,000 | energy MAE (kcal/mol): 0.94 -> 0.59 | Helped |
| Use SchNet instead of enn-s2s-ens5 | QM9 energy prediction · QM9, N=110,462 | energy MAE (kcal/mol): 0.33 -> 0.31 | Helped |
| Add force supervision to SchNet training (ablation) | SchNet · MD17, N=50,000 | force MAE (kcal/mol/Å): 1.23 -> 0.17 | Helped |
| Add force supervision to SchNet training (ablation) | SchNet · ISO17 known molecules, unknown conformation | energy MAE (kcal/mol): 0.52 -> 0.36 | Helped |
| Add force supervision to SchNet training (ablation) | SchNet · ISO17 unknown molecules, unknown conformation | force MAE (kcal/mol/Å): 5.71 -> 2.18 | Helped |
| Train SchNet on energies and forces rather than energies alone (ablation) | SchNet · MD17, N=1,000 | energy MAE (kcal/mol): 2.03 -> 0.13 | Helped |
| Train SchNet on energies and forces rather than energies alone (ablation) | SchNet · MD17, N=1,000 | force MAE (kcal/mol/Å): 6.56 -> 0.39 | Helped |
| Use SchNet rather than GDML | MD17 malonaldehyde energy prediction, N=1,000 · MD17 Malonaldehyde | energy MAE (kcal/mol): 0.16 -> 0.13 | Helped |
| Use SchNet rather than GDML | MD17 benzene force prediction, N=1,000 · MD17 Benzene | force MAE (kcal/mol/Å): 0.23 -> 0.31 | Hurt |
| Train across different molecules with shared elemental composition | SchNet · ISO17 unknown molecules, unknown conformation | energy MAE (kcal/mol) | Mixed |
Closest papers by meaning
- On the equivalence of molecular graph convolution and molecular wave function with poor basis set (NeurIPS 2020 · similarity 0.88)
- The Importance of Being Scalable: Improving the Speed and Accuracy of Neural Network Interatomic Potentials Across Chemical Domains (NeurIPS 2024 · similarity 0.85)
- Learning the Electronic Hamiltonian of Large Atomic Structures (ICML 2025 · similarity 0.84)
- All-atom Diffusion Transformers: Unified generative modelling of molecules and materials (ICML 2025 · similarity 0.84)
- From Density Matrices to Phase Transitions in Deep Learning: Spectral Early Warnings and Interpretability (NeurIPS 2026 · similarity 0.84)
- High-order Equivariant Flow Matching for Density Functional Theory Hamiltonian Prediction (NeurIPS 2025 · similarity 0.84)
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.