AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Transform Once: Efficient Operator Learning in Frequency Domain

Michael Poli, Stefano Massaroli, Federico Berto, Jinkyoo Park, Tri Dao, Christopher Re, Stefano Ermon

NeurIPS 2022 · Architectures and efficiency › RNNs and normalization · approach: Scalable approximation and equivariant compute · 3 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
replace FNO with T1+vpT1 · Navier-Stokes high viscosityN-MSE: 0.379 -> 0.257Improved
replace FFNO with T1+vpT1 · Navier-Stokes high viscosityN-MSE: 0.997 -> 0.257Improved
replace FNO with T1+vpT1 · Navier-Stokesstep time (ms): 31 -> 19Improved
add variance preserving init to FNOFNO · Navier-Stokes high viscosityN-MSE: 0.379 -> 0.351Improved
add variance preserving init to FNOFNO · Navier-Stokes low viscosityN-MSE: 0.328 -> 0.315Improved
add vp init to T1 (ablation)T1 · Navier-Stokes high viscosityN-MSE: 0.491 -> 0.304Improved
replace k-space model with UNet (ablation)T1 · Navier-Stokes high viscosityN-MSE: 0.491 -> 0.295Improved
combine UNet k-space model with vp init (ablation)T1 · Navier-Stokes low viscosityN-MSE: 0.449 -> 0.240Improved
increase training dataset sizeT1 vs FNO · Navier-Stokes nu=1e-4N-MSEMixed
use T1+vp instead of FNOT1 · Flow around airfoilstraining time (hrs): 6.0 -> 1.3Improved
use T1+vp instead of FNOT1 · Flow around airfoilsN-MSE: 0.020 -> 0.024No clear effect
use T1+vp instead of FNOT1 · ScalarFlowtraining time (hrs): 32.4 -> 4.7Improved

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