SOTAVerified

Thompson Sampling

Thompson sampling, named after William R. Thompson, is a heuristic for choosing actions that addresses the exploration-exploitation dilemma in the multi-armed bandit problem. It consists of choosing the action that maximizes the expected reward with respect to a randomly drawn belief.

Papers

Showing 501–525 of 655 papers

TitleStatusHype
Policy Gradient Optimization of Thompson Sampling Policies—0
Position-Based Multiple-Play Bandits with Thompson Sampling—0
Posterior Sampling-Based Bayesian Optimization with Tighter Bayesian Regret Bounds—0
Posterior sampling for reinforcement learning: worst-case regret bounds—0
Posterior Sampling via Autoregressive Generation—0
Practical Adversarial Attacks on Stochastic Bandits via Fake Data Injection—0
Preferential Multi-Objective Bayesian Optimization—0
Prior-free and prior-dependent regret bounds for Thompson Sampling—0
Probabilistic Inference in Reinforcement Learning Done Right—0
Profitable Bandits—0
QoS-Aware Multi-Armed Bandits—0
Racing Thompson: an Efficient Algorithm for Thompson Sampling with Non-conjugate Priors—0
Random Effect Bandits—0
Random Hypervolume Scalarizations for Provable Multi-Objective Black Box Optimization—0
Randomised Bayesian Least-Squares Policy Iteration—0
Randomized Exploration in Cooperative Multi-Agent Reinforcement Learning—0
Regenerative Particle Thompson Sampling—0
Regret Analysis of Bandit Problems with Causal Background Knowledge—0
Regret Analysis of the Finite-Horizon Gittins Index Strategy for Multi-Armed Bandits—0
Regret Bounds for Information-Directed Reinforcement Learning—0
Regularized-OFU: an efficient algorithm for general contextual bandit with optimization oracles—0
Reinforcement Learning for Efficient and Tuning-Free Link Adaptation—0
Reinforcement learning techniques for Outer Loop Link Adaptation in 4G/5G systems—0
Reinforcement Learning with Subspaces using Free Energy Paradigm—0
Reinforcement Learning with Trajectory Feedback—0
Show:102550
← PrevPage 21 of 27Next →

No leaderboard results yet.