Acquisition: Thompson Sampling vs UCB — DEJAN Codebase Explorer

Thompson Sampling selects claims by drawing random samples from each hypothesis's posterior Beta distributions, whereas Upper Confidence Bound deterministically selects the claim with the highest upper confidence interval calculated as \(\mu + c \sqrt{\frac{\ln(N)}{n}}\).

Referenced by

Thompson Sampling in Content Optimizer
Two-Stage Acquisition Policy
Acquisition Policy Formulas
Bayesian Optimization Loop Details
Bayesian Acquisition Policy Mechanics
Bayesian Optimization Engine Mechanics
Stage 2 Bayesian Acquisition
Seeder Hypothesis Generation
Snippet Optimization for Citations
Bayesian Acquisition Methods
Thompson Sampling Mechanics
Snippet vs Page Mode
Beta Posterior Mechanics
Two-Stage Acquisition Policy
Content Optimizer Acquisition Algorithm
Bayesian Acquisition Algorithms
Bayesian Optimization Loop Mechanics
Bayesian Acquisition Policy
Alternative Acquisition Methods in Optimizer