Research map / Bandits and online learning / Contextual bandits and Thompson sampling
Practical Contextual Bandits with Regression Oracles
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 RegCB instead of ε-Greedy | Contextual bandit learning · Unmodified UCI datasets | validation reward | Helped |
| Use RegCB instead of ILTCB | Contextual bandit learning · Unmodified UCI datasets | validation reward | Helped |
| Use RegCB instead of ε-Greedy or ILTCB | Contextual bandit learning with GB5 oracle · All datasets | normalized relative loss | Helped |
| Use Bootstrap instead of RegCB | Contextual bandit learning with GB5 oracle · Some datasets | validation reward | Mixed |
| Use RegCB instead of Bootstrap | Contextual bandit learning with linear oracle · All datasets | normalized relative loss | Helped |
| Use the elimination variant instead of the optimistic variant | RegCB · All datasets, aggregate performance | normalized relative loss | Helped |
| Use a fixed ensemble structure within each epoch | RegCB with GB5 oracle · All datasets | validation reward | Hurt |
| Use the optimistic RegCB variant | RegCB · letter | confidence-width decay slope: -0.34 | Helped |
| Use the optimistic RegCB variant | RegCB · letter-noise | confidence-width decay slope: -0.51 | Helped |
| Use the optimistic RegCB variant | RegCB · yahoo | confidence-width decay slope: -0.52 | Helped |
| Use Bootstrap instead of RegCB | Contextual bandit learning · letter | confidence-width decay slope: -0.05 | No clear effect |
Closest papers by meaning
- Beyond UCB: Optimal and Efficient Contextual Bandits with Regression Oracles (ICML 2020 · similarity 0.95)
- Adapting to misspecification in contextual bandits with offline regression oracles (ICML 2021 · similarity 0.94)
- Adapting to Misspecification in Contextual Bandits (NeurIPS 2020 · similarity 0.93)
- Contextual Online Decision Making with Infinite-Dimensional Functional Regression (ICML 2025 · similarity 0.90)
- Generalized Linear Bandits with Limited Adaptivity (NeurIPS 2024 · similarity 0.90)
- Regret Bounds for Adversarial Contextual Bandits with General Function Approximation and Delayed Feedback (NeurIPS 2025 · similarity 0.90)
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.