Get Started
Research questionHow can batched entropic optimal transport stop updating converged instances without changing its stopping rule?In a static batch of independent entropic optimal transport problems, every update continues until the slowest instance meets tolerance. Computation is therefore spent updating instances that have already converged.
Inference Optimization
Machine Learning
Statistical Machine Learning
Latest papersRecent research connected to this question, newest first.DrainSinkhorn: Safe Elimination for Batched Entropic Optimal TransportEvidence concerns an active-packing and compaction layer with a Sinkhorn-specific screen and verifier under the configured two-sided residual check. Results cover tested Flash-backed, OTT-JAX, and PyKeOps backends and heterogeneous workloads, with reported speedups and residual, output, and training-quality checks limited to those settings.research paper · Sep 3, 2026
Related questions
How can batched Lipschitz narrowing achieve low regret when the zooming dimension is unknown?How can iterated one-step approximations control error in strongly continuous convex monotone semigroups?When can reasoning LLMs reliably optimize numerical and semantically rich discrete search spaces in batches?How can SGMCMC hyperparameters be tuned without Metropolis-Hastings acceptance rates when computation is limited?
Home
Topics
Search
Library