Research map / Learning theory and algorithms / Prediction and decision losses
Coresets for Wasserstein Distributionally Robust Optimization Problems
Ruomin Huang, Jiawei Huang, Wenjie Liu, Hu Ding
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use dual coreset sampling instead of uniform sampling | WDRO logistic regression · Letter, c=1% | worst-case risk: 0.72518 -> 0.68933 | Improved |
| use dual coreset sampling instead of uniform sampling | WDRO SVM · Letter, c=1% | worst-case risk: 0.68707 -> 0.67866 | Improved |
| use dual coreset sampling instead of uniform sampling | WDRO Huber regression · Appliances Energy, c=1% | worst-case risk: 28.57655 -> 28.57627 | Improved |
| use dual coreset instead of layer sampling | WDRO logistic regression · Letter, c=10% | worst-case risk: 0.60007 -> 0.60113 | No clear effect |
| use dual coreset instead of layer sampling | WDRO SVM · Letter, c=2% | worst-case risk: 0.58967 -> 0.59850 | Hurt |
| use importance sampling instead of uniform sampling | WDRO logistic regression · Letter, c=2% | worst-case risk: 0.64630 -> 0.65798 | Hurt |
| use importance sampling instead of uniform sampling | WDRO SVM · Letter, c=1% | worst-case risk: 0.68707 -> 0.66576 | Improved |
| increase compression rate c from 1% to 10% | DualCore on WDRO logistic regression · Letter | worst-case risk: 0.68933 -> 0.60113 | Improved |
| increase compression rate c from 1% to 10% | DualCore on WDRO SVM · Letter | worst-case risk: 0.67866 -> 0.51066 | Improved |
| dual coreset instead of uniform sampling | WDRO logistic regression · MNIST 0 vs 1, c=0.5% | test accuracy: 94.14 -> 94.37 | Improved |
| dual coreset instead of uniform sampling | WDRO logistic regression · MNIST 0 vs 2, c=0.5% | test accuracy: 87.1 -> 84.98 | Hurt |
| dual coreset instead of uniform sampling | WDRO logistic regression · MNIST 2 vs 7, c=0.5% | test accuracy: 78.54 -> 86.15 | Improved |
Closest papers by meaning
- Knowledge-Guided Wasserstein Distributionally Robust Optimization (ICML 2025 · similarity 0.86)
- Robust Sparsification via Sensitivity (ICML 2025 · similarity 0.85)
- Provable Robust Overfitting Mitigation in Wasserstein Distributionally Robust Optimization (ICLR 2025 · similarity 0.85)
- Group Distributionally Robust Optimization with Flexible Sample Queries (NeurIPS 2026 · similarity 0.82)
- Universal generalization guarantees for Wasserstein distributionally robust models (ICLR 2025 · similarity 0.82)
- Small coresets via negative dependence: DPPs, linear statistics, and concentration (NeurIPS 2024 · similarity 0.82)
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.
