Reproduce “Digital Typhoon Dataset V3: Benchmarking for Multi-spectral Extension and Analyzing Systematic Bias in Intensity Regression” 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

Asanobu Kitamoto, Victor Le Louarn, Noelia Falczuk, Ngoc Anh Tong, Anja Veleva

The authors extend the Digital Typhoon dataset with three satellite channels and test whether multispectral inputs improve estimates of central pressure. On a matched four-channel dataset, ResNet-18’s RMSE is 6.9270 hPa with Channel 1 and 6.8308 hPa with Channels 1+2+3, a difference comparable to run variability. They find strong storms are underestimated; reweighting reduces, but does not remove, this bias.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.