Research map / Graph learning / GNN architectures
Equiformer: Equivariant Graph Attention Transformer for 3D Atomistic Graphs
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 |
|---|---|---|---|
| Add non-linear message passing to MLP attention (ablation) | Equiformer · QM9 | Δε MAE (meV): 32 -> 30 | Helped |
| Replace dot product attention with MLP attention (ablation) | Equiformer with linear message passing · QM9 | Δε MAE (meV): 32 -> 32 | No clear effect |
| Add non-linear message passing (ablation) | Equiformer with MLP attention · OC20 IS2RE validation | Energy MAE (eV), average: 0.5555 -> 0.5489 | Helped |
| Replace dot product attention with MLP attention (ablation) | Equiformer with linear message passing · OC20 IS2RE validation | Energy MAE (eV), average: 0.5689 -> 0.5555 | Helped |
| Increase L_max from 2 to 3 (ablation) | Equiformer · MD17 | MAE | Mixed |
| Add Noisy Nodes data augmentation | Equiformer with IS2RS training · OC20 IS2RE test | Energy MAE (eV), average: 0.4728 -> 0.4660 | Helped |
Closest papers by meaning
- E2Former: An Efficient and Equivariant Transformer with Linear-Scaling Tensor Products (NeurIPS 2025 · similarity 0.91)
- EquiformerV2: Improved Equivariant Transformer for Scaling to Higher-Degree Representations (ICLR 2024 · similarity 0.89)
- Equivariant Spherical Transformer for Efficient Molecular Modeling (NeurIPS 2026 · similarity 0.89)
- Generalist Equivariant Transformer Towards 3D Molecular Interaction Learning (ICML 2024 · similarity 0.88)
- Compact SO(3) Equivariant Atomistic Foundation Models via Structural Pruning (NeurIPS 2026 · similarity 0.88)
- On Explaining Equivariant Graph Networks via Improved Relevance Propagation (ICML 2025 · similarity 0.88)
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.