Reproduce “Statistical Optimal Transport via Factored Couplings” 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
Aden Forrow, Jan-Christian Hütter, Mor Nitzan, Philippe Rigollet, Geoffrey Schiebinger, Jonathan Weed
The authors introduce FactoredOT to align cell-expression datasets across laboratories and sequencing protocols for label transfer. It represents transport as a mixture of a few product distributions and estimates the coupling through a Wasserstein barycenter. In one hematopoietic dataset pair, label-transfer mean error was 14.10%, versus 17.47% for ordinary optimal transport; this result is limited to the evaluated task.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.