Reproduce “RankFormer: Rethinking Time Series Transformers through Representation Capacity Analysis and Progressive Decoding” 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
Peiwang Tang, Cheng HE, Xijie Liang, Defu Lian
RankFormer addresses multivariate time-series forecasting with patch tokens, hourglass attention, and a decoder that exchanges information across channels and refines forecasts. Its rank analysis treats matrix rank as a proxy for representation diversity, not forecast accuracy. On long-term benchmarks with fixed input length 96, it has the best MSE and MAE on seven of eight datasets; SEER leads on Traffic.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.