AI Paper Atlas

Research map / Learning theory and algorithms / Optimal transport

Riemannian Metric Learning via Optimal Transport

Christopher Scarvelis, Justin Solomon

ICLR 2023 · Learning theory and algorithms › Optimal transport · approach: Neural optimal transport · 2 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 the learned Riemannian metric for scRNA trajectory inference instead of the non-regularized baselineTrajectory inference model · scRNA trajectory inference, k ≥ 10mean EMD: 19.7% mean reduction in EMDImproved
Use the learned Riemannian metric instead of Euclidean geodesicsSnow goose trajectory inference · Snow goose trajectoriesmean DTW distance: 26.9% reduction in mean DTW distanceImproved
Use the Euclidean metric regularizer instead of no regularizerscRNA trajectory inference · scRNA trajectory inferencemean EMDHurt
Increase regularization coefficient λ to 10^5 (ablation)Learned Riemannian metric · Circular synthetic metric recoveryeigenspace alignment: alignment score falls to 0.857Hurt
Vary regularization coefficient λ across the tested values, excluding the largest value (ablation)Learned Riemannian metric · Circular synthetic metric recoveryeigenspace alignmentNo clear effect
Remove the optimal transport term (ablation)Learned Riemannian metric · Circular synthetic metric recoverydata-informed metric recoveryHurt

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