PDEfuncta: Spectrally-Aware Neural Representation for PDE Solution Modeling

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Jo, Minju, Cho, Woojin, Mudiyanselage, Uvini Balasuriya, Lee, Seungjun, Park, Noseong, Lee, Kookjin
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918059973804032
author Jo, Minju
Cho, Woojin
Mudiyanselage, Uvini Balasuriya
Lee, Seungjun
Park, Noseong
Lee, Kookjin
author_facet Jo, Minju
Cho, Woojin
Mudiyanselage, Uvini Balasuriya
Lee, Seungjun
Park, Noseong
Lee, Kookjin
contents Scientific machine learning often involves representing complex solution fields that exhibit high-frequency features such as sharp transitions, fine-scale oscillations, and localized structures. While implicit neural representations (INRs) have shown promise for continuous function modeling, capturing such high-frequency behavior remains a challenge-especially when modeling multiple solution fields with a shared network. Prior work addressing spectral bias in INRs has primarily focused on single-instance settings, limiting scalability and generalization. In this work, we propose Global Fourier Modulation (GFM), a novel modulation technique that injects high-frequency information at each layer of the INR through Fourier-based reparameterization. This enables compact and accurate representation of multiple solution fields using low-dimensional latent vectors. Building upon GFM, we introduce PDEfuncta, a meta-learning framework designed to learn multi-modal solution fields and support generalization to new tasks. Through empirical studies on diverse scientific problems, we demonstrate that our method not only improves representational quality but also shows potential for forward and inverse inference tasks without the need for retraining.
format Preprint
id arxiv_https___arxiv_org_abs_2506_12790
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PDEfuncta: Spectrally-Aware Neural Representation for PDE Solution Modeling
Jo, Minju
Cho, Woojin
Mudiyanselage, Uvini Balasuriya
Lee, Seungjun
Park, Noseong
Lee, Kookjin
Machine Learning
Numerical Analysis
Computational Physics
Scientific machine learning often involves representing complex solution fields that exhibit high-frequency features such as sharp transitions, fine-scale oscillations, and localized structures. While implicit neural representations (INRs) have shown promise for continuous function modeling, capturing such high-frequency behavior remains a challenge-especially when modeling multiple solution fields with a shared network. Prior work addressing spectral bias in INRs has primarily focused on single-instance settings, limiting scalability and generalization. In this work, we propose Global Fourier Modulation (GFM), a novel modulation technique that injects high-frequency information at each layer of the INR through Fourier-based reparameterization. This enables compact and accurate representation of multiple solution fields using low-dimensional latent vectors. Building upon GFM, we introduce PDEfuncta, a meta-learning framework designed to learn multi-modal solution fields and support generalization to new tasks. Through empirical studies on diverse scientific problems, we demonstrate that our method not only improves representational quality but also shows potential for forward and inverse inference tasks without the need for retraining.
title PDEfuncta: Spectrally-Aware Neural Representation for PDE Solution Modeling
topic Machine Learning
Numerical Analysis
Computational Physics
url https://arxiv.org/abs/2506.12790