Research map / Learning theory and algorithms / Submodular and game algorithms
Hitting the High Notes: Subset Selection for Maximizing Expected Order Statistics
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 the theoretically weaker Quantile score instead of the KR score for expected maximum | score-based subset selection · Synthetic normal distributions, n=500 | expected maximum | No clear effect |
| Use KR score instead of picking by expectation | score-based subset selection · Synthetic normal distributions, n=500 | expected maximum | Helped |
| Vary the quantile used by the Quantile algorithm (ablation) | Quantile algorithm · Synthetic normal distributions, k=10,20,30 | expected maximum | Mixed |
| Use theory-prescribed quantiles (1-1/sqrt(k) for Quantile, 1-1/k for KR) versus greedy submodular | Quantile, KR and Greedy · Synthetic normal distributions | expected maximum | No clear effect |
| Give some candidates only 10 samples instead of 5000 (data scarcity) | all selection methods · Synthetic selection-bias experiment | percentage of small-data candidates selected | Mixed |
| Use quantile-loss or KR regression instead of squared-loss regression | neural network scorer · Twitter likes, n=500 chunks, k=10 and 30 | expected maximum likes | Helped |
Closest papers by meaning
- Tight Asymptotics of Extreme Order Statistics (NeurIPS 2025 · similarity 0.85)
- Maxing and Ranking with Few Assumptions (NeurIPS 2017 · similarity 0.84)
- A Principled Approach to Randomized Selection under Uncertainty: Applications to Peer Review and Grant Funding (NeurIPS 2025 · similarity 0.84)
- Sharp Non-Asymptotic Analysis of the Penalized Challenger in $\beta$-EB-TCI for Bernoulli Bandits (NeurIPS 2026 · similarity 0.84)
- Learning Safe Strategies for Value Maximizing Buyers in Uniform Price Auctions (ICML 2025 · similarity 0.83)
- Random Exploration in Bayesian Optimization: Order-Optimal Regret and Computational Efficiency (ICML 2024 · similarity 0.83)
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.