Research map / Bandits and online learning / Multi-armed bandits
Learning the Optimal Recommendation from Explorative Users
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 BAIR's two-phase safe-sweeping and elimination strategy instead of Track-and-stop | Best-arm identification from revealed preferences · Synthetic bandits (K=20, δ=0.1, α=1) | stopping time: 3107 -> 2113 | Helped |
| Use BAIR instead of Track-and-stop | Best-arm identification from revealed preferences · Synthetic bandits (K=100, δ=0.01, α=1) | probability of success: 0.930 -> 1.000 | Helped |
| Use BAIR instead of Track-and-stop | Best-arm identification from revealed preferences · Synthetic bandits (K=20, δ=0.1, α=0.8) | stopping time: 2272 -> 2070 | Helped |
| Use BAIR instead of Track-and-stop | Best-arm identification from revealed preferences · Synthetic bandits (K=20, δ=0.1, α=2.0) | probability of success: 1.000 vs. 1.000 | No clear effect |
| Add click noise and wait for multiple rejections before eliminating an arm (ablation) | BAIR · Synthetic bandits (K=20, δ=0.1, click noise p=0.1) | probability of success: 1.000 -> 0.966 | Hurt |
| Use a smaller Phase-1 acceptance budget, N1=K, rather than the theoretical choice (ablation) | BAIR (α=0.8) · Synthetic bandits (K=20, δ=0.1) | probability of success: 1.000 -> 0.981 | Hurt |
Closest papers by meaning
- Modeling Attrition in Recommender Systems with Departing Bandits (AAAI 2022 · similarity 0.91)
- Fiduciary Bandits (ICML 2020 · similarity 0.89)
- Regret in Online Recommendation Systems (NeurIPS 2020 · similarity 0.89)
- Fighting Boredom in Recommender Systems with Linear Reinforcement Learning (NeurIPS 2018 · similarity 0.89)
- Learning from a Learning User for Optimal Recommendations (ICML 2022 · similarity 0.88)
- Would You Like Your Data to Be Trained? A User Controllable Recommendation Framework (AAAI 2024 · similarity 0.87)
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.