Research map / Probabilistic and scientific modeling / Molecules and proteins
So3krates: Equivariant attention for interactions on arbitrary length-scales in molecular systems
Thorben Frank, Oliver Thorsten Unke, Klaus Robert Muller
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 geometric corrections localized in SPHC space (ablation) | So3krates · Cumulene rotor energy profiles | energy profile fidelity | Improved |
| Add spherical filters based on SPHC distances (ablation) | So3krates · MD17 | energy and force prediction accuracy | Improved |
| Increase maximum SPHC degree (ablation) | So3krates · QM7-X250, individual and joint models | force MAE | Improved |
| Increase maximum SPHC degree beyond 2 (ablation) | So3krates jointly trained model · QM7-X250 unknown molecules | force MAE | No clear effect |
| Increase feature dimension from F = 132 to F = 264 (ablation) | So3krates · QM7-X known and unknown molecules | Energy MAE: 16.815 -> 15.228; 21.733 -> 21.750 | Mixed |
| Use non-local corrections (ablation) | So3krates · QM7-X known and unknown molecules | Energy and force MAE: 16.815 / 20.422 -> 16.176 / 19.879; 21.733 / 25.211 -> 20.071 / 24.236 | No clear effect |
| Increase geometric order in SPHCs from lmax=0 to lmax≥2 (ablation) | So3krates · QM7-X250 individual structures | force prediction accuracy | Improved |
| Use non-local geometric corrections instead of local-only message passing (ablation) | So3krates · Cumulene molecules of different sizes | energy profile fidelity | Improved |
Closest papers by meaning
- Higher-Rank Irreducible Cartesian Tensors for Equivariant Message Passing (NeurIPS 2024 · similarity 0.87)
- Learning Equivariant Non-Local Electron Density Functionals (ICLR 2025 · similarity 0.87)
- Quadruple Attention in Many-body Systems for Accurate Molecular Property Predictions (ICML 2025 · similarity 0.86)
- Learning the Electronic Hamiltonian of Large Atomic Structures (ICML 2025 · similarity 0.86)
- Efficient and Scalable Density Functional Theory Hamiltonian Prediction through Adaptive Sparsity (ICML 2025 · similarity 0.85)
- The Importance of Being Scalable: Improving the Speed and Accuracy of Neural Network Interatomic Potentials Across Chemical Domains (NeurIPS 2024 · 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.
