Fast Amortized Fitting of Scientific Signals Across Time and Ensembles via Transferable Neural Fields
This work addresses efficiency and scalability issues for researchers modeling complex scientific signals like turbulent flows and astrophysical systems, representing an incremental improvement through feature transfer.
The paper tackles the slow convergence and scaling challenges of neural fields in high-dimensional scientific settings by extending them to spatiotemporal and multivariate signals with transferable features, achieving up to an order of magnitude reduction in iterations for target reconstruction quality and gains exceeding 10 dB in early-stage reconstruction.
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.