A note on the Bayesian regret of Thompson Sampling with an arbitrary prior
Sébastien BubeckChe-Yu LiuPublished in: CoRR (2013)
Keyphrases
- prior distribution
- proposal distribution
- bayesian model
- markov chain monte carlo
- bayesian methods
- prior knowledge
- prior probabilities
- posterior distribution
- bayesian networks
- loss function
- lower bound
- likelihood model
- metropolis hastings algorithm
- fully bayesian
- multi armed bandit
- bayesian learning
- sample size
- worst case
- online learning
- sequential monte carlo
- bayesian models
- random sampling
- particle filter
- dirichlet prior
- maximum a posteriori
- sampling algorithm
- bayesian inference
- upper bound
- expert advice
- posterior probability
- maximum likelihood
- sampled data
- conjugate priors
- state space
- bayesian estimation
- prior information
- long run
- gaussian processes