Research map / Graph learning / GNN architectures
GemNet: Universal Directional Graph Neural Networks for Molecules
Johannes Gasteiger, Florian Becker, Stephan Günnemann
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 |
|---|---|---|---|
| GemNet-Q vs DimeNet++ | DimeNet++ · COLL | force MAE: 40 -> 26.4 | Improved |
| add symmetric message passing (ablation) | GemNet-Q · COLL | force MAE: 28.5 -> 27.0 | Improved |
| use bilinear layer instead of Hadamard product (ablation) | GemNet-Q · COLL | force MAE: 29.3 -> 27.0 | Improved |
| add atom embedding updates (ablation) | GemNet-Q · COLL | force MAE: 28.3 -> 27.0 | Improved |
| add one-hop message passing (ablation) | GemNet-Q · COLL | force MAE: 31.3 -> 27.0 | Improved |
| add two-hop message passing (ablation) | GemNet-Q · COLL | force MAE: 32.4 -> 27.0 | Improved |
| add two-hop message passing (GemNet-Q vs GemNet-T) (ablation) | GemNet-T · MD17 | force MAE: Aspirin 9.5 -> 9.4 | No clear effect |
| add precomputed variance scaling factors (ablation) | GemNet-Q · COLL | force MAE: 29.1 -> 27.0 | Improved |
| use layer normalization instead of scaling factors (ablation) | GemNet-Q · COLL | force MAE: 27.0 -> 33.3 | Hurt |
| add biases (ablation) | GemNet-Q · COLL | force MAE: 27.0 -> 27.2 | No clear effect |
| predict forces directly instead of via energy gradient (ablation) | GemNet-Q · COLL | force MAE: 26.4 -> 38.1 | Hurt |
| predict forces directly instead of via energy gradient (ablation) | GemNet · OC20 S2EF | force MAE | Improved |
Closest papers by meaning
- Neural Atoms: Propagating Long-range Interaction in Molecular Graphs through Efficient Communication Channel (ICLR 2024 · similarity 0.89)
- Long-Short-Range Message-Passing: A Physics-Informed Framework to Capture Non-Local Interaction for Scalable Molecular Dynamics Simulation (ICLR 2024 · similarity 0.88)
- On the Scalability of GNNs for Molecular Graphs (NeurIPS 2024 · similarity 0.88)
- Neural P$^3$M: A Long-Range Interaction Modeling Enhancer for Geometric GNNs (NeurIPS 2024 · similarity 0.88)
- Expressivity and Generalization: Fragment-Biases for Molecular GNNs (ICML 2024 · similarity 0.88)
- 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.
