Research map / Learning theory and algorithms / Submodular and game algorithms
Noisy Submodular Maximization via Adaptive Sampling with Applications to Crowdsourced Image Collection Summarization
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 adaptive TopX sampling over possible top-l subset sizes instead of uniformly sampling all items | ExpGreedy · Synthetic submodular maximization with value queries | query budget to reach the desired guarantee | Helped |
| Use adaptive TopX sampling over possible top-l subset sizes instead of fixed top-1 identification (ablation) | ExpGreedy_G · Synthetic submodular maximization with value queries | sample complexity | Helped |
| Use adaptive TopX sampling over possible top-l subset sizes instead of fixed top-1 identification with approximation slack (ablation) | ExpGreedy_O · Synthetic submodular maximization with value queries | sample complexity | Helped |
| Replace adaptive TopX exploration with uniform sampling (ablation) | ExpGreedy · Synthetic submodular maximization with value queries (\sigma^2=10) and preference queries (\beta=0.5, \tau=1) | budget needed to achieve a desired total utility: up to 3 times more budget | Hurt |
| Increase noise variance in value-query responses | ExpGreedy · Synthetic submodular maximization with value queries | solution utility at a limited query budget | Hurt |
| Decrease preference-model noise parameter β | ExpGreedy · Synthetic submodular maximization with preference queries | solution utility at a limited query budget | Hurt |
| Use fewer pairwise comparisons per preference-query estimate (smaller τ) (ablation) | ExpGreedy · Synthetic submodular maximization with preference queries | solution utility at a limited query budget | Hurt |
| Use crowdsourced pairwise preferences to estimate item values | ExpGreedy with TopX · Crowdsourced image collection summarization: Venice, Venice Carnival, Venice Cathedrals | summary quality | Helped |
| Increase the allowed query budget | ExpGreedy · Synthetic submodular maximization with value and preference queries | solution utility | Helped |
Closest papers by meaning
- Data Summarization at Scale: A Two-Stage Submodular Approach (ICML 2018 · similarity 0.88)
- Fast Constrained Submodular Maximization: Personalized Data Summarization (ICML 2016 · similarity 0.88)
- Practical $0.385$-Approximation for Submodular Maximization Subject to a Cardinality Constraint (NeurIPS 2024 · similarity 0.85)
- Deletion-Robust Submodular Maximization with Knapsack Constraints (AAAI 2024 · similarity 0.85)
- Online Two-Stage Submodular Maximization (NeurIPS 2025 · similarity 0.85)
- Streaming Stochastic Submodular Maximization with On-Demand User Requests (NeurIPS 2025 · similarity 0.84)
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.