Search papers, labs, and topics across Lattice.
This research establishes a novel connection between numerical solvers for partial differential equations (PDEs) and neural emulators, revealing that neural architectures can effectively mirror classical discretizations and their errors can be analyzed using spectral methods. The study introduces APEBench, a benchmarking suite designed for autoregressive neural emulators that leverages fast differentiable pseudo-spectral solvers, alongside an exploration of how unconverged solvers impact surrogate training. Key findings suggest that understanding the interplay between solver errors and architectural biases can enhance the performance of neural emulators, leading to more efficient simulations in engineering and science.
Neural emulators can outperform traditional PDE solvers by leveraging insights from both paradigms, revealing a surprising synergy that enhances simulation efficiency.
Simulation is central to modern engineering and science, but the cost of numerical solvers for partial differential equations (PDEs) remains a bottleneck whenever fast or many-query evaluations are required. Neural emulators trained on solver-generated data promise significant speedups, yet they are usually framed as opaque alternatives to the very methods that produce their training signal. This thesis argues the two paradigms are more alike than different: neural architectures mirror classical discretizations, their errors are amenable to the same spectral analysis, and insight flows profitably in both directions. We approach the relationship by disentangling the multiple roles a solver plays in the emulator learning pipeline. Mode-wise Fourier analysis then provides a common language in which solver errors, architectural inductive biases, and training objectives can all be read off simultaneously. Taken together, this allows synthesizing three contributions. (1) APEBench, a comprehensive benchmarking suite for autoregressive neural emulators of PDEs that uses fast differentiable pseudo-spectral solvers in JAX. (2) Progressively Refined Differentiable Physics, an investigation of the effect of unconverged solvers on surrogate training. (3) Neural Emulator Superiority, an analysis of the influence of numerical errors and architectural inductive biases.
