AI Paper Atlas

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

Hypernetwork-based Meta-Learning for Low-Rank Physics-Informed Neural Networks

Woojin Cho, Kookjin Lee, Donsub Rim, Noseong Park

NeurIPS 2023 · Probabilistic and scientific modeling › Differential equations and PDEs · approach: Physics-informed neural network training · 7 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 hypernetwork-based meta-learning instead of MAMLNaive-LR-PINN · Convection beta=30relative error: 0.6294 -> 0.0379Improved
use hypernetwork-based meta-learning instead of ReptileNaive-LR-PINN · Convection beta=40relative error: 0.5632 -> 0.0722Improved
add curriculum learning pre-trainingNaive-LR-PINN · Convection beta=35relative error: 0.5357 -> 0.5465No clear effect
use adaptive rank instead of full rankHyper-LR-PINN · Convection beta=30relative error: 0.0379 -> 0.0389No clear effect
use adaptive rank instead of full rankHyper-LR-PINN · Convection beta=40relative error: 0.0655 -> 0.0722Hurt
replace vanilla PINN with Hyper-LR-PINNPINN · Convection, general CDRrelative error: 0.0327 -> 0.0085Improved
replace MAML with Hyper-LR-PINNMAML · Conv-Diff-Reacrelative error: 0.0144 -> 0.0329Hurt
replace MAML with Hyper-LR-PINNMAML · Reactionrelative error: 0.0057 -> 0.0045Improved
replace PINN with low-rank Hyper-LR-PINN (parameter count)PINN · model size# trainable parameters: 10,401 -> 351Improved
low-rank format with hypernetwork vs vanilla PINNPINN · training timetraining time: more than 4x speed upImproved
use fixed instead of trainable basis vectors (ablation)Hyper-LR-PINN · ConvectionaccuracyImproved
add orthogonality constraint on basis vectors (ablation)Hyper-LR-PINN · ConvectionaccuracyImproved

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