Reproduce “Invertible Fourier Neural Operators for Tackling Both Forward and Inverse Problems” 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
Da Long, Zhitong Xu, Qiwei Yuan, Yin Yang, Shandian Zhe
iFNO makes Fourier neural operators bidirectional by using reversible transformations in latent channels, with a β-VAE to represent variation in inverse predictions. It learns forward and inverse function-to-function mappings through staged training. On seven simulated benchmarks, it often outperforms alternatives; in noisy D-LINE training, error is 0.0899 ± 0.0014, without isolating which component causes gains.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.