Research map / Learning theory and algorithms / Prediction and decision losses
Minimizing Upper Confidence Bounds: A Data-Driven Framework for Stochastic Programming
Shixin Liu, Ming Gao, Jian Hu
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 |
|---|---|---|---|
| minimize APUB at higher nominal level instead of SAA | two-stage product-mix problem · small samples N=120 and 240 | out-of-sample cost (mean, 10th-90th percentile, extremes) | Improved |
| use APUB-M with suitable nominal level at N=120 instead of SAA-M | two-stage product-mix problem · N=120 vs SAA-M at N=240 | out-of-sample cost | Improved |
| increase sample size to 480 and 960 | APUB-M vs SAA-M · two-stage product-mix problem | out-of-sample cost gap | No clear effect |
| use APUB nominal level (1-alpha) as confidence knob instead of SAA | APUB-M · coverage probability, N up to 960 | empirical coverage probability | Improved |
| vary number of bootstrap scenarios M from 1000 to 5000 (ablation) | adapted L-shaped method · N=120 to 960, |I|=20, |J|=8 | runtime: N=960: 73.8 -> 75.3 -> 74.1 s (M=1000, 3000, 5000) | No clear effect |
| replace standard L-shaped for SAA-M with adapted L-shaped for APUB-M | two-stage product-mix problem · N=120, |I|=20, |J|=8 | runtime (s): 5.7 -> 7.2 | Hurt |
| use APUB-M instead of Wasserstein DRO with an inappropriate radius | two-stage product-mix problem with fixed recourse · N=30, 120, 480 | mean out-of-sample cost | Improved |
Closest papers by meaning
- New Sample Complexity Bounds for Sample Average Approximation in Heavy-Tailed Stochastic Programming (ICML 2024 · similarity 0.86)
- From Low Probability to High Confidence in Stochastic Convex Optimization (JMLR 2021 · similarity 0.86)
- Stochastic Conditional Gradient Methods: From Convex Minimization to Submodular Maximization (JMLR 2020 · similarity 0.86)
- Non-Asymptotic and Non-Lipschitzian Bounds on Optimal Values in Stochastic Optimization Under Heavy Tails (ICML 2025 · similarity 0.86)
- Statistical Properties of Robust Satisficing (ICML 2024 · similarity 0.85)
- BONSAI: Bayesian Optimization with Natural Simplicity and Interpretability (NeurIPS 2026 · similarity 0.85)
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.
