AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Scaling Laws and Symmetry, Evidence from Neural Force Fields

Khang Ngo, Siamak Ravanbakhsh

ICLR 2026 · Architectures and efficiency › CNN design · approach: Group symmetry equivariant networks · 6 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
use equivariant architecture (eSEN) instead of unconstrained MPNNMPNN · force-field data scalingdata-scaling exponent beta: 0.311 -> 0.753Improved
use equivariant architecture (eSEN) instead of unconstrained MPNNMPNN · force-field parameter scalingparameter-scaling exponent alpha: 0.276 -> 0.817Improved
use higher-order equivariance (GemNet-OC) instead of EGNNEGNN · force-field scalingalpha: 0.387 -> 0.524Improved
use eSEN instead of unconstrained MPNNMPNN · compute-optimal frontiergamma_c: 0.142 -> 0.403Improved
add symmetry-regularization loss (ablation)unconstrained MPNN · force-field data scalingdata exponent beta: slightly increases betaImproved
add symmetry-regularization loss (ablation)unconstrained MPNN · force-field parameter scalingparameter exponent alpha: decreases alphaHurt
add symmetry-regularization loss (ablation)unconstrained MPNN · compute-optimal slopegamma_c: unchanged exponent; frontier shifted by 0.14 log10(6)No clear effect
add data augmentation under 1% data for 100 epochs (ablation)unconstrained MPNN · multi-epoch regimecompute-optimal power law: recovers gamma_c 0.14, F_c 0.96Improved
add test-time group averaging (M=32) (ablation)unconstrained model · force-field parameter scalingloss: minor change in constant ANo clear effect
scale equivariant vector channels in EGNN (ablation)EGNN · validation lossvalidation lossImproved

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