Research map / Image generation / Fast diffusion sampling
Thompson Sampling with Diffusion Generative Prior
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 a diffusion prior with Thompson sampling instead of the evaluated baseline priors and UCB1 | DiffTS · Popular and Niche; iPinYou Bidding; 2D Maze, perfect and imperfect data | regret | Helped |
| Use a diffusion prior rather than the best-performing baseline | DiffTS · Popular and Niche; 2D Maze | regret: around two times smaller | Helped |
| Use a GMM prior instead of a full-covariance Gaussian prior | Thompson sampling · Popular and Niche; 2D Maze | regret | Hurt |
| Use a GMM prior instead of a full-covariance Gaussian prior | Thompson sampling · iPinYou Bidding | regret | No clear effect |
| Remove variance calibration (ablation) | DiffTS · Bandit problems | regret | Hurt |
| Remove the EM-like procedure for training from imperfect data (ablation) | Diffusion prior trained on imperfect data · Bandit problems | regret | Hurt |
| Add SURE-based regularization to the imperfect-data training loss (ablation) | Diffusion prior trained on imperfect data · Bandit problems | regret | Helped |
| Vary the regularization parameter (ablation) | SURE-based imperfect-data training loss · Bandit problems | regret | Mixed |
Closest papers by meaning
- Diffusion Models Meet Contextual Bandits (NeurIPS 2025 · similarity 0.90)
- Online Posterior Sampling with a Diffusion Prior (NeurIPS 2024 · similarity 0.90)
- Prior-Guided Diffusion Planning for Offline Reinforcement Learning (NeurIPS 2025 · similarity 0.89)
- Learning to Sample From Diffusion Models via Inverse Reinforcement Learning (NeurIPS 2026 · similarity 0.89)
- Feel-Good Thompson Sampling for Contextual Bandits: a Markov Chain Monte Carlo Showdown (NeurIPS 2025 · similarity 0.89)
- The Choice of Noninformative Priors for Thompson Sampling in Multiparameter Bandit Models (AAAI 2024 · similarity 0.88)
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.