Reproduce “Adaptive Randomized Neural Networks for PDEs: Structure-Aware Feature Redistribution and Residual-Driven Block Enrichment” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
You Yang, Fei Wang
SFRBE–RaNN adapts randomized neural networks for PDE approximation by redistributing a fixed feature budget toward localized structures, then adding features aligned with residual directions the current space misses. On a two-peak Poisson problem, its ten-seed mean relative L2 error was 1.164×10−6, versus 4.545×10−4 for SFR–RaNN and 5.897×10−3 for AD–RaNN at comparable widths; this is benchmark-specific, not a general guarantee.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.