AI Paper Atlas

Research map / Learning theory and algorithms / Optimal transport

Low-rank Optimal Transport: Approximation, Statistics and Debiasing

Meyer Scetbon, marco cuturi

NeurIPS 2022 · Learning theory and algorithms › Optimal transport · approach: Transport map estimation · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use debiased DLOT instead of biased LOT as the gradient-flow objective (ablation)low-rank OT gradient flow · Gaussian to moon-shape distribution, r=100, 1000 samplesrecovery of target distributionImproved
increase the rank rplug-in LOT/DLOT estimator · mixture of 10 anisotropic Gaussiansstatistical convergence rateHurt
increase ground-space dimension dplug-in LOT/DLOT estimator · mixture of 10 anisotropic Gaussiansstatistical convergence rateNo clear effect
use k-means or generalized k-means initialization instead of rank-2 or random (ablation)mirror-descent LOT solver · Newsgroup20 embeddings, 50 runsavoidance of spurious local minima / stopping criterionImproved
vary initialization (k-means, generalized k-means, rank-2, random) (ablation)mirror-descent LOT solver · Newsgroup20 embeddingsfinal cost valueNo clear effect
use shortest-path graph cost instead of squared Euclidean cost for LOT clustering (ablation)LOT-based clustering · 6 scikit-learn synthetic datasetscluster recoveryImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy