Research map / Bandits and online learning / Contextual bandits and Thompson sampling
Double Thompson Sampling for Dueling 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 |
|---|---|---|---|
| Select both compared candidates using independent Thompson samples | D-TS · MSLR Condorcet 5-arm dataset | cumulative regret | Helped |
| Use D-TS+ regret-minimizing tie-breaking instead of D-TS random tie-breaking (ablation) | D-TS · MSLR non-Condorcet 5-arm dataset | cumulative regret | Helped |
| Use regret-minimizing tie-breaking to focus on a winner rather than randomly exploring all Copeland winners (ablation) | D-TS · Multiple-winner Copeland settings | cumulative regret | Helped |
| Remove RUCB-based elimination of likely non-winner first candidates (ablation) | D-TS · Copeland dueling bandits | cumulative regret | Hurt |
| Remove RLCB-based elimination of comparisons that are unlikely to provide information (ablation) | D-TS · Non-Condorcet dueling bandits | cumulative regret | Hurt |
| Use random tie-breaking rather than regret-minimizing tie-breaking (ablation) | D-TS · MSLR non-Condorcet 5-arm dataset | cumulative regret | Hurt |
| Use D-TS+ rather than CCB | Dueling bandit algorithms · MSLR non-Condorcet 5-arm dataset | cumulative regret: less than 10% of that of CCB | Helped |
| Use D-TS+ rather than ECW-RMED | Dueling bandit algorithms · MSLR Condorcet 5-arm dataset | cumulative regret | Helped |
| Use D-TS+ rather than ECW-RMED | Dueling bandit algorithms · MSLR non-Condorcet 5-arm dataset | regret standard deviation relative to mean: almost 200% -> 13.16% | Helped |
Closest papers by meaning
- Feel-Good Thompson Sampling for Contextual Dueling Bandits (ICML 2024 · similarity 0.92)
- Variance-aware Regret Bounds for Stochastic Contextual Dueling Bandits (ICLR 2024 · similarity 0.89)
- On Weak Regret Analysis for Dueling Bandits (NeurIPS 2024 · similarity 0.89)
- Thompson Sampling For Combinatorial Bandits: Polynomial Regret and Mismatched Sampling Paradox (NeurIPS 2024 · similarity 0.89)
- Borda Regret Minimization for Generalized Linear Dueling Bandits (ICML 2024 · similarity 0.89)
- Fusing Reward and Dueling Feedback in Stochastic Bandits (ICML 2025 · similarity 0.89)
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.