Research map / Bandits and online learning / Contextual bandits and Thompson sampling
MOTS: Minimax Optimal Thompson Sampling
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 clipped Gaussian sampling with an adaptive clipping threshold | Thompson sampling · 1-subGaussian K-armed bandit | problem-independent regret: O(\sqrt{KT}) | Helped |
| Use clipped 𝒥 sampling instead of clipped Gaussian sampling (ablation) | MOTS · Gaussian K-armed bandit | problem-independent regret: O\bigg(\sqrt{KT}+\sum_{i=2}^K \Delta_i\bigg) | Helped |
| Use clipped 𝒥 sampling instead of clipped Gaussian sampling (ablation) | MOTS · Gaussian bandit, K=50, ε=0.2, T=10^7 | regret: 9615 -> 9245 | Helped |
| Use clipped Gaussian sampling with ρ∈(1/2,1) rather than ρ=1 | Thompson sampling · K-armed bandit | problem-independent regret: O(\sqrt{KT\log K}) -> O(\sqrt{KT}) | Helped |
| Use time-varying ρ to approach 1 (ablation) | MOTS · 1-subGaussian K-armed bandit | asymptotic regret rate: \sum_{i:\Delta_i>0}\frac{2}{\rho \Delta_i} -> \sum_{i:\Delta_i>0}\frac{2}{\Delta_i} | Helped |
Closest papers by meaning
- Finite-Time Regret of Thompson Sampling Algorithms for Exponential Family Multi-Armed Bandits (NeurIPS 2022 · similarity 0.92)
- What Does Thompson Sampling Optimize? (ICML 2026 · similarity 0.91)
- Lenient Regret for Multi-Armed Bandits (AAAI 2021 · similarity 0.90)
- Thompson Sampling with Less Exploration is Fast and Optimal (ICML 2023 · similarity 0.90)
- Thompson Sampling for Multi-Objective Linear Contextual Bandit (NeurIPS 2025 · similarity 0.89)
- Thompson Sampling For Combinatorial Bandits: Polynomial Regret and Mismatched Sampling Paradox (NeurIPS 2024 · 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.