Reproduce “Learning Optimal Transport Plans Via Autoregressive Token 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

Ivan Jaen Marquez, Sotirios Panagiotis Chytas, Vikas Singh

ToR-OT generates sparse transport plans for discrete optimal-transport problems, avoiding a new optimization loop at test time. It serializes nonzero entries row by row, conditioning on earlier allocations and target capacities, but does not enforce marginal constraints. On 30-bin squared-Euclidean examples, mean MAPE was 0.056 versus Meta OT’s 0.229; each method was scored against its own reference solution, and feasibility is not guaranteed.

Summary by alphaXiv.

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