Reproduce “A Controlled Visual-Backbone Benchmark for Multimodal Short-Term Solar Irradiance Forecasting” 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

Oshadha Samarakoon, Dushan Herath, Ishara Ranmandala, Dilshara Herath, Roshan Godaliyadda, Parakrama Ekanayake, Vijitha Herath

The authors compare visual backbones for ten-minute-ahead solar forecasting while holding the multimodal pipeline fixed, pairing sky images with weather history. On Folsom, every evaluated visual model beats smart persistence, with VMamba Small lowest at 65.39 W/m². On NREL’s 313-sample test split, persistence beats all visual models; these single-seed results establish no universal architecture winner.

Summary by alphaXiv.

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