AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Overtone: Cyclic Patch Modulation for Cleaner, Faster Physics Emulators

Payel Mukhopadhyay, Michael McCabe, Ruben Ohana, Miles Cranmer

ICLR 2026 · Architectures and efficiency › RNNs and normalization · approach: Scalable approximation and equivariant compute

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
cycle stride/patch size at inference (CSM) instead of fixed patchAxial ViT (50M) · 10-step rollout, Shear FlowVRMSE: 0.0793 -> 0.0565Improved
cycle stride/patch size at inference (CSM) instead of fixed patchVanilla ViT (100M) · 10-step rollout, SupernovaVRMSE: 1.75 -> 1.20Improved
cycle stride/patch size at inference (CSM) instead of fixed patchVanilla ViT (100M) · 10-step rollout, Active MatterVRMSE: 0.370 -> 0.359Improved
use 4->8->16 stride order instead of random order (ablation)CSM Vanilla ViT · 10-step rollout, Shear FlowVRMSE: 0.0433 -> 0.0375Improved
use 4->8->16 stride order instead of 8->4->16 (ablation)CSM Vanilla ViT · 10-step rollout, Shear FlowVRMSE: 0.0442 -> 0.0375Improved
use 4->8->16 stride order instead of 16->4->8 (ablation)CSM Vanilla ViT · 10-step rollout, SupernovaVRMSE: 1.325 -> 1.204Improved
reduce patch size from 16 to 4 at inferenceCSM/CKM · Active Matter 256^2VRMSEImproved
train one flexible CSM model instead of separate fixed patch modelsVanilla ViT (100M) · Next-step Shear Flow, 2048 tokensVRMSE: 0.00677 -> 0.00546Improved
use flexible CKM CViT encoder instead of fixed patchCViT · Shear Flow 128/16VRMSE: 0.319 -> 0.127Improved
use flexible CKM CViT encoder instead of fixed patchCViT · Turbulent Radiative Layer 2D 3072/4VRMSE: 0.299 -> 0.302No clear effect
omit patch size 8 from training (ablation)CSM/CKM · excluded patch sizeVRMSEHurt
use warm-up then coarse schedule or two-step dwell instead of 4->8->16CSM Vanilla ViT · rollout schedulesVRMSENo clear effect

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