Research map / Architectures and efficiency / RNNs and normalization
Efficient Equivariant High-Order Crystal Tensor Prediction via Cartesian Local-Environment Many-Body Coupling
Dian Jin, Yancheng Yuan, Xiaoming Tao
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 channel interaction matrix (vs identity) (ablation) | CEITNet · JARVIS dielectric (order-2) | Fnorm: 3.05 -> 2.87 | Improved |
| add interaction (coupling) term to separable path (ablation) | CEITNet · JARVIS dielectric (order-2) | Fnorm: 3.21 -> 2.87 | Improved |
| add separable path (ablation) | CEITNet · JARVIS dielectric (order-2) | Fnorm: 2.91 -> 2.87 | No clear effect |
| add separable path (ablation) | CEITNet · JARVIS dielectric (order-2) | EwT 5%: 35.2% -> 39.3% | Improved |
| add learnable degree normalization (ablation) | CEITNet · JARVIS dielectric (order-2) | Fnorm: 3.23 -> 2.87 | Improved |
| add learnable degree normalization (ablation) | CEITNet · JARVIS dielectric (order-2) | EwT 10%: 52.9% -> 63.8% | Improved |
| reduce channel width to K=4 from K=16 (ablation) | CEITNet · JARVIS dielectric (order-2) | Fnorm: 2.87 -> 3.16 | Hurt |
| increase channel width to K=64 from K=16 (ablation) | CEITNet · JARVIS dielectric (order-2) | Fnorm: 2.87 -> 3.22 | Hurt |
| increase channel width to K=32 from K=16 (ablation) | CEITNet · JARVIS dielectric (order-2) | Fnorm: 2.87 -> 3.22 | Hurt |
| use CEITNet instead of GeoCTP | tensor prediction · JARVIS piezoelectric (order-3) | Fnorm: 0.778 -> 0.517 | Improved |
| use CEITNet instead of GMTNet | tensor prediction · JARVIS elastic (order-4) | Fnorm: 67.38 -> 70.11 | Hurt |
| use CEITNet instead of GMTNet | tensor prediction · JARVIS elastic (order-4) | EwT 5%: 7.7% -> 14.4% | Improved |
Closest papers by meaning
- A Space Group Symmetry Informed Network for O(3) Equivariant Crystal Tensor Prediction (ICML 2024 · similarity 0.90)
- Higher-Rank Irreducible Cartesian Tensors for Equivariant Message Passing (NeurIPS 2024 · similarity 0.88)
- Learning the Electronic Hamiltonian of Large Atomic Structures (ICML 2025 · similarity 0.87)
- ELECTRA: A Cartesian Network for 3D Charge Density Prediction with Floating Orbitals (NeurIPS 2025 · similarity 0.87)
- Spin-Weighted Spherical Harmonics Enable Complete and Scalable E(3)-Equivariant Networks (NeurIPS 2026 · similarity 0.87)
- Efficient and Scalable Density Functional Theory Hamiltonian Prediction through Adaptive Sparsity (ICML 2025 · 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.
