Research map / Bandits and online learning / Contextual bandits and Thompson sampling
Factorization Bandits for Interactive Recommendation
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 |
|---|---|---|---|
| model user dependency graph W (ablation) | factorUCB · synthetic simulation | cumulated regret | Helped |
| model user dependency graph W (ablation) | factorUCB · Yahoo Today Module | normalized CTR | Mixed |
| model user dependency graph W (ablation) | factorUCB · LastFM | normalized cumulated reward | Helped |
| add latent item factors to contextual bandit | contextual bandits (GOB.Lin, CoLin) · synthetic simulation | cumulated regret | Helped |
| add latent item factors to contextual bandit | contextual bandits · Yahoo Today Module | normalized CTR | Helped |
| use observed context features in factorization bandit | PTS (pure factorization) · LastFM | normalized cumulated reward | Helped |
| UCB confidence-based exploration instead of epsilon-greedy | ALS factorization bandit · Yahoo Today Module | normalized CTR | Helped |
| hide most informative rather than least informative factors as latent (ablation) | factorUCB · synthetic simulation | cumulated regret | Hurt |
| empirically tuned exploration parameters instead of theoretical values (ablation) | factorUCB · synthetic simulation | cumulated regret | No clear effect |
Closest papers by meaning
- Blocked Collaborative Bandits: Online Collaborative Filtering with Per-Item Budget Constraints (NeurIPS 2023 · similarity 0.88)
- On Context-Dependent Clustering of Bandits (ICML 2017 · similarity 0.88)
- Recommendations with Sparse Comparison Data: Provably Fast Convergence for Nonconvex Matrix Factorization (ICML 2025 · similarity 0.87)
- Learning-Efficient Yet Generalizable Collaborative Filtering for Item Recommendation (ICML 2024 · similarity 0.86)
- Online Clustering of Dueling Bandits (ICML 2025 · similarity 0.86)
- Would You Like Your Data to Be Trained? A User Controllable Recommendation Framework (AAAI 2024 · 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.