Research map / Bandits and online learning / Contextual bandits and Thompson sampling
Contextual Bandits with Smooth Regret: Efficient Learning in Continuous Action Spaces
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 SmoothIGW with CORRAL instead of empMOSS++ | New Yorker Cartoon Caption Contest bandit · New Yorker Cartoon Caption Contest | regret | No clear effect |
| use SmoothIGW and a linear argmin predictor instead of CATS | continuous-action contextual bandits · Cpu | average progressive loss (×1000): [55, 57] vs [40.6, 40.7] | Helped |
| use SmoothIGW and a linear argmin predictor instead of CATS | continuous-action contextual bandits · Fri | average progressive loss (×1000): [183, 187] vs [161, 163] | Helped |
| use SmoothIGW and a linear argmin predictor instead of CATS | continuous-action contextual bandits · Price | average progressive loss (×1000): [108, 110] vs [70.2, 70.5] | Helped |
| use SmoothIGW and a linear argmin predictor instead of CATS | continuous-action contextual bandits · Wis | average progressive loss (×1000): [172, 174] vs [138, 139] | Helped |
| use SmoothIGW and a linear argmin predictor instead of CATS | continuous-action contextual bandits · Zur | average progressive loss (×1000): [24, 26] vs [24.3, 24.4] | Mixed |
| replace the linear argmin predictor with a random-Fourier-feature predictor (ablation) | Ours (Linear) · Cpu | average progressive loss (×1000): [40.6, 40.7] -> [38.6, 38.7] | Helped |
| replace the linear argmin predictor with a random-Fourier-feature predictor (ablation) | Ours (Linear) · Fri | average progressive loss (×1000): [161, 163] -> [156, 157] | Helped |
| replace the linear argmin predictor with a random-Fourier-feature predictor (ablation) | Ours (Linear) · Price | average progressive loss (×1000): [70.2, 70.5] -> [66.1, 66.3] | Helped |
| replace the linear argmin predictor with a random-Fourier-feature predictor (ablation) | Ours (Linear) · Wis | average progressive loss (×1000): [138, 139] -> [136.2, 136.6] | Helped |
| replace the linear argmin predictor with a random-Fourier-feature predictor (ablation) | Ours (Linear) · Zur | average progressive loss (×1000): [24.3, 24.4] -> [25.4, 25.5] | Hurt |
Closest papers by meaning
- Contextual Bandits with Large Action Spaces: Made Practical (ICML 2022 · similarity 0.95)
- Context-lumpable stochastic bandits (NeurIPS 2023 · similarity 0.92)
- Active Context Selection Improves Simple Regret in Contextual Bandits (NeurIPS 2026 · similarity 0.92)
- Contextual Bandits for Unbounded Context Distributions (ICML 2025 · similarity 0.90)
- Adaptive Regret for Bandits Made Possible: Two Queries Suffice (ICLR 2024 · similarity 0.90)
- Optimal and Efficient Contextual Combinatorial Semi-bandits with General Function Approximation (NeurIPS 2026 · 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.