AI Paper Atlas

Research map / Probabilistic and scientific modeling / Differential equations and PDEs

Iterative Refinement Neural Operators are Learned Fixed-Point Solvers: A Principled Approach to Spectral Bias Mitigation

Xiaotian Liu, Shuyuan Shang, Xiaopeng Wang, Pu Ren, Yaoqing Yang

ICML 2026 · Probabilistic and scientific modeling › Differential equations and PDEs · approach: Neural operators for general geometries · 0 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
add iterative refinement (IRNO) to a pre-trained FNOFNO · TR-2DVRMSE: 0.2394 -> 0.1309Improved
add iterative refinement (IRNO) to TFNOTFNO · Active MatterVRMSE: 0.1981 -> 0.0387Improved
add IRNO on top of a strong WDSR baseWDSR · ERA5 16x super-resolutionACC: 0.9091 -> 0.9104Improved
use progressive spectral loss (lambda 1->2) instead of fixed lambda (ablation)IRNO (TFNO base) · Active MatterVRMSE: 0.051-0.070 -> 0.039Improved
add deep supervision over every refinement step (ablation)IRNO (1-D synthetic) · 1-D synthetic Jacobian monotonicityfraction with m > 0: 3.9% -> 100%Improved
combine IRNO with HFS frequency-aware architectureHFS · Active MatterVRMSE: 0.0631 -> 0.0486Improved
use IRNO trained on a weaker base operator (TFNO) to refine a stronger one (FNO)FNO · TR-2DVRMSE improvement (%): 58.53% vs 45.32% for same-operator IRNOImproved
step size alpha 0.3-0.5 instead of 0.2 (ablation)IRNO (TFNO) · TR-2Dstability beyond training horizonHurt
step size alpha 0.6 (ablation)IRNO (TFNO) · TR-2DVRMSE: 0.12 -> 0.24Hurt
swap refinement backbone (ResNet, ConvNext, FNO) or normalization (BatchNorm, LayerNorm, GroupNorm) (ablation)IRNO · Active Matter / TR-2DVRMSE reduction: > 71% VRMSE reduction for all backbonesNo clear effect
apply IRNO to RIGNO on irregular meshes with autoregressive rolloutRIGNO · CE-Gaussrelative L2 error (%): 16.617 -> 13.080 at t = 7Improved
use iterative refinement instead of a capacity-matched monolithic residual modelFNO · ERA5ACC at 1100 GFLOPs: 0.79 -> 0.84Improved

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