AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Geometric Transformer with Interatomic Positional Encoding

Yusong Wang, Shaoning Li, Tong Wang, Bin Shao, Nanning Zheng, Tie-Yan Liu

NeurIPS 2023 · Architectures and efficiency › Attention and transformers · approach: Transformer expressivity and learnability · 1 citations (OpenAlex)

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 changedWhereMeasuredEffect
update IPE across layers via residual connection (vs non-updated IPE) (ablation)Geoformer · QM9 U0MAE (meV): 5.63 -> 4.43Improved
update IPE across layers via residual connection (vs non-updated IPE) (ablation)Geoformer · QM9 UMAE (meV): 5.87 -> 4.41Improved
update IPE across layers via residual connection (vs non-updated IPE) (ablation)Geoformer · QM9 HMAE (meV): 6.21 -> 4.39Improved
update IPE across layers via residual connection (vs non-updated IPE) (ablation)Geoformer · QM9 GMAE (meV): 7.24 -> 6.13Improved
multiply IPE with attention scores instead of adding as attention bias (ablation)Geoformer · QM9 U0MAE (meV): 7.62 -> 4.43Improved
multiply IPE with attention scores instead of adding as attention bias (ablation)Geoformer · QM9 UMAE (meV): 7.66 -> 4.41Improved
multiply IPE with attention scores instead of adding as attention bias (ablation)Geoformer · QM9 HMAE (meV): 7.93 -> 4.39Improved
multiply IPE with attention scores instead of adding as attention bias (ablation)Geoformer · QM9 GMAE (meV): 9.01 -> 6.13Improved
Geoformer vs Transformer-M (adds directional info beyond distances, no pretraining)Geoformer · QM9 U0MAE (meV): 9.37 -> 4.43Improved
Geoformer vs Transformer-MGeoformer · QM9 eps_LUMOMAE (meV): 16.2 -> 15.4Improved
Geoformer vs Transformer-MGeoformer · QM9 eps_HOMOMAE (meV): 17.5 -> 18.4Hurt
Geoformer vs Transformer-MGeoformer · QM9 delta epsMAE (meV): 27.4 -> 33.8Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy