AI Paper Atlas

Research map / Learning theory and algorithms / Optimal transport

Loss-Aware Distributionally Robust Optimization via Trainable Optimal Transport Ambiguity Sets

Jonas Ohnemus, Marta Fochesato, Riccardo Zuliani, John Lygeros

ICML 2026 · Learning theory and algorithms › Optimal transport · approach: Transport map estimation · 0 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
learn the Mahalanobis transport-cost geometry end-to-end via bilevel hypergradient instead of fixed isotropic costOT-DRO portfolio (CVaR) · Portfolio, k=2, J=30 single instanceworst-case CVaR and out-of-sample CVaRImproved
learn transport-cost geometry end-to-endOT-DRO portfolio (CVaR) · Portfolio, 50 experiments, k=3, 10 to 100 samplesrelative improvement of worst-case and out-of-sample CVaRImproved
fewer samples in the datasetlearned OT-DRO geometry (portfolio) · Portfolio, k=3relative improvementImproved
add the bootstrap coverage penalty (lambda_p > 0) (ablation)learned OT-DRO geometry · Portfolio, 10 to 100 samples, coverage d(P, P_hat; L)coverage of true distributionImproved
learn transport-cost geometry end-to-endDR linear regression (absolute error) · Linear regression, single instance, J=20worst-case and out-of-sample absolute errorImproved
learn transport-cost geometry end-to-endDR linear regression (absolute error) · Linear regression, 10 experiments, 10 to 50 samplesrelative improvementImproved

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