AI engineering glossary · Reinforcement learning
What is thompson sampling?
Thompson sampling (also called Posterior sampling): A bandit method that keeps a probability distribution over each option's true average reward, draws one sample from each distribution, and picks the option with the highest sample. Uncertain options get sampled high sometimes, so exploration happens naturally.
Thompson sampling: an example
For click rates with Beta(successes+1, failures+1) beliefs, an arm with 3 clicks in 10 views often gets sampled above one with 30 in 100.
Related terms
More Reinforcement learning terms
- Transition probability
- Twin delayed DDPG
- Upper confidence bound
- Value iteration
- World model
- A3C
- Action-value function
- Actor-critic
Where you learn it
Thompson sampling is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.