Research map / Bandits and online learning / Multi-armed bandits
Modeling Attrition in Recommender Systems with Departing Bandits
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 UCB-Hybrid over fixed-arm policies | Departing bandits with a single user type and general departure probabilities · Single user type, general departure probabilities | expected regret: O(√(KT log T + K log T)) | Helped |
| Use UCB-Hybrid over threshold policies | Departing bandits with two user types and two categories · Two user types, two categories, departure after first no-click | expected regret: O(√T ln T) | Helped |
| Restrict the learner to always recommending one fixed category | Departing bandits · Section 2.1 example | expected regret: Ω(T) | Hurt |
| Use a belief-dependent policy that switches categories instead of always recommending one category | Departing bandits with two user types and two categories · Section 2.1 example | expected return: E[V π∗]−E[V π1] > 0.0169; E[V π∗] − E[V π2] > 1.22 × 10−5 | Helped |
Closest papers by meaning
- Incentivized Bandit Learning with Self-Reinforcing User Preferences (ICML 2021 · similarity 0.91)
- Learning the Optimal Recommendation from Explorative Users (AAAI 2022 · similarity 0.91)
- Bandit Learning with Joint Effect of Incentivized Sampling, Delayed Sampling Feedback, and Self-Reinforcing User Preferences (ICLR 2022 · similarity 0.90)
- Fighting Boredom in Recommender Systems with Linear Reinforcement Learning (NeurIPS 2018 · similarity 0.89)
- Bandits Meet Mechanism Design to Combat Clickbait in Online Recommendation (ICLR 2024 · similarity 0.89)
- Bandits with Ranking Feedback (NeurIPS 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.