AI Paper Atlas

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

Heavy-tailed Physics-Informed Neural Networks

Jephte Abijuru, Mayank Nagda, Jan Tauberschmidt, Phil Ostheimer, Sebastian Josef Vollmer, Stephan Mandt, Marius Kloft, Sophie Fellenz

ICML 2026 · Probabilistic and scientific modeling › Differential equations and PDEs · approach: Physics-informed neural network training

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 Gaussian (MSE) residual model with Student-t EM-weighted lossPINN with MLP 3x32 · Failure-mode benchmarks (Wave, Reaction, Convection, Allen-Cahn)rRMSE: Allen-Cahn best baseline 0.0545 -> 0.0097; Wave 0.0081 -> 0.0019Improved
apply t-PINN loss to a standard MLPMLP PINN · Waverelative l2 error: 9.0 × 10−3 -> 1.6 × 10−3Improved
apply t-PINN loss to a standard MLPMLP PINN · Allen-Cahnrelative l2 error: 2.1 × 10−2 -> 2.7 × 10−3Improved
apply t-PINN loss to a standard MLPMLP PINN · Korteweg-de Vriesrelative l2 error: 1.6 × 10−1 -> 2.9 × 10−2Improved
apply t-PINN loss to modified MLPModified MLP PINN · Ginzburg-Landaurelative l2 error: 3.1 × 10−2 -> 9.9 × 10−3Improved
apply t-PINN loss to PirateNetPirateNet PINN · Allen-Cahnrelative l2 error: 1.4 × 10−5 -> 1.9 × 10−5Hurt
apply t-PINN loss to PirateNetPirateNet PINN · Korteweg-de Vriesrelative l2 error: 4.3 × 10−4 -> 5.1 × 10−4Hurt
apply t-PINN loss to PirateNetPirateNet PINN · Waverelative l2 error: 7.1 × 10−5 -> 6.9 × 10−5No clear effect
add Student-t likelihood with EM residual weights (fixed nu) (ablation)Gaussian PINN · 1D convection (MLP)relative l2 error: 0.0304 -> 0.0109Improved
learn degrees of freedom nu instead of fixing it (ablation)Student-t PINN · 1D convection (MLP)relative l2 error: 0.0109 -> 0.0028Improved

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