Fast Amortized Fitting of Scientific Signals Across Time and Ensembles via Transferable Neural Fields
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910156809306112 |
|---|---|
| author | Zorek, Sophia Vyas, Kushal Liu, Yuhao Lenz, David Peterka, Tom Balakrishnan, Guha |
| author_facet | Zorek, Sophia Vyas, Kushal Liu, Yuhao Lenz, David Peterka, Tom Balakrishnan, Guha |
| contents | Neural fields, also known as implicit neural representations (INRs), offer a powerful framework for modeling continuous geometry, but their effectiveness in high-dimensional scientific settings is limited by slow convergence and scaling challenges. In this study, we extend INR models to handle spatiotemporal and multivariate signals and show how INR features can be transferred across scientific signals to enable efficient and scalable representation across time and ensemble runs in an amortized fashion. Across controlled transformation regimes (e.g., geometric transformations and localized perturbations of synthetic fields) and high-fidelity scientific domains-including turbulent flows, fluid-material impact dynamics, and astrophysical systems-we show that transferable features improve not only signal fidelity but also the accuracy of derived geometric and physical quantities, including density gradients and vorticity. In particular, transferable features reduce iterations to reach target reconstruction quality by up to an order of magnitude, increase early-stage reconstruction quality by multiple dB (with gains exceeding 10 dB in some cases), and consistently improve gradient-based physical accuracy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_19979 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Fast Amortized Fitting of Scientific Signals Across Time and Ensembles via Transferable Neural Fields Zorek, Sophia Vyas, Kushal Liu, Yuhao Lenz, David Peterka, Tom Balakrishnan, Guha Machine Learning Computational Engineering, Finance, and Science Computer Vision and Pattern Recognition Neural fields, also known as implicit neural representations (INRs), offer a powerful framework for modeling continuous geometry, but their effectiveness in high-dimensional scientific settings is limited by slow convergence and scaling challenges. In this study, we extend INR models to handle spatiotemporal and multivariate signals and show how INR features can be transferred across scientific signals to enable efficient and scalable representation across time and ensemble runs in an amortized fashion. Across controlled transformation regimes (e.g., geometric transformations and localized perturbations of synthetic fields) and high-fidelity scientific domains-including turbulent flows, fluid-material impact dynamics, and astrophysical systems-we show that transferable features improve not only signal fidelity but also the accuracy of derived geometric and physical quantities, including density gradients and vorticity. In particular, transferable features reduce iterations to reach target reconstruction quality by up to an order of magnitude, increase early-stage reconstruction quality by multiple dB (with gains exceeding 10 dB in some cases), and consistently improve gradient-based physical accuracy. |
| title | Fast Amortized Fitting of Scientific Signals Across Time and Ensembles via Transferable Neural Fields |
| topic | Machine Learning Computational Engineering, Finance, and Science Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2604.19979 |