AI Paper Atlas

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

Learning Functional Transduction

Mathieu Chalvidal, Thomas Serre, Rufin VanRullen

NeurIPS 2023 · Probabilistic and scientific modeling › Differential equations and PDEs · approach: Physics-informed operator learning · 1 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 meta-trained Transducer instead of per-instance FNOTransducer · ADR PDE regressionRMSE: 2.96e^{-4} -> 2.39e^{-4}Improved
use Transducer instead of DeepONetTransducer · ADR PDE regressionRMSE: 2.02e^{-2} -> 2.39e^{-4}Improved
single feedforward pass of pre-trained Transducer instead of fitting FNO from scratchTransducer · ADR PDE regressionregression time (s): 1.72e^{2} -> 3.10e^{-3}Improved
Transducer instead of FNO for computeTransducer · ADR PDE regressionGFLOPs: 1.68e^{2} -> 1.06e^{-1}Improved
increase number of examples beyond 60FNO vs Transducer · ADR PDE regressionRMSE vs dataset sizeMixed
increase Transducer depth with untied weights (ablation)Transducer · ADR PDE regressiontraining lossImproved
replace exp dot product kernel with RBF kernel (ablation)Transducer · ADR PDE regression, s=10RMSE: 2.71e-3 -> 8.71e-3Hurt
replace exp dot product kernel with RBF kernel (ablation)Transducer · ADR PDE regression, s=500RMSE: 1.79e-4 -> 3.22e-4Hurt
replace exp dot product kernel with l2-norm kernel (ablation)Transducer · ADR PDE regression, s=100RMSE: 2.39e-4 -> 6.98e-4Hurt
replace exp dot product kernel with l2-norm kernel, test at s=500 beyond meta-training sizes (ablation)Transducer · ADR PDE regression, s=500RMSE: 1.79e-4 -> 7.33e5Hurt
condition on previous-year window (P.Y) instead of ViT baselineTransducer · ERA5 pressureLWMSE (hPa): 32.826 -> 25.293Improved
condition on 15-day window lagging one week (P.W) instead of previous year window (P.Y)Transducer · ERA5 pressureLWMSE (hPa): 25.293 -> 22.718Improved

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