Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Thompson Sampling via Local Uncertainty
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 context-dependent local latent uncertainty instead of making the latent variable global (ablation) | LU-Gauss · Eight contextual bandit datasets | normalized cumulative regret: Mushroom 31.40 -> 21.28; Financial 13.10 -> 12.35; Statlog 8.73 -> 7.45; Jester 70.19 -> 69.98; Wheel 52.90 -> 64.03; Co | Mixed |
| Use context-dependent local latent uncertainty instead of making the latent variable global (ablation) | LU-SIVI · Eight contextual bandit datasets | normalized cumulative regret: Mushroom 14.84 -> 20.32; Financial 8.28 -> 9.01; Statlog 7.62 -> 6.34; Jester 71.64 -> 71.10; Wheel 63.06 -> 79.33; Cove | Mixed |
| Use a semi-implicit rather than diagonal-Gaussian local variational posterior | Local-uncertainty Thompson sampling · Mushroom | normalized cumulative regret: 31.40 -> 14.84 | Helped |
| Use a semi-implicit rather than diagonal-Gaussian local variational posterior | Local-uncertainty Thompson sampling · Eight contextual bandit datasets | mean value of normalized cumulative regret: 46.03 -> 45.42 | Helped |
| Use local rather than global uncertainty (ablation) | Thompson sampling · Eight contextual bandit datasets | mean value of normalized cumulative regret | Helped |
Closest papers by meaning
- Feel-Good Thompson Sampling for Contextual Bandits: a Markov Chain Monte Carlo Showdown (NeurIPS 2025 · similarity 0.92)
- The Choice of Noninformative Priors for Thompson Sampling in Multiparameter Bandit Models (AAAI 2024 · similarity 0.89)
- $\mathtt{VITS}$ : Variational Inference Thompson Sampling for contextual bandits (ICML 2024 · similarity 0.89)
- No-Regret Thompson Sampling for Finite-Horizon Markov Decision Processes with Gaussian Processes (NeurIPS 2025 · similarity 0.89)
- Contextual Thompson Sampling via Generation of Missing Data (NeurIPS 2025 · similarity 0.89)
- Variance-Aware Feel-Good Thompson Sampling for Contextual Bandits (NeurIPS 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.