Bandit Convex Optimization: \(\sqrt{T}\) Regret in One Dimension.
Sébastien BubeckOfer DekelTomer KorenYuval PeresPublished in: COLT (2015)
Keyphrases
- convex optimization
- regret bounds
- lower bound
- online convex optimization
- bandit problems
- weighted majority
- online learning
- interior point methods
- upper confidence bound
- worst case
- multi armed bandit
- linear regression
- multi armed bandit problems
- low rank
- total variation
- upper bound
- convex relaxation
- primal dual
- convex formulation
- online algorithms
- convex optimization problems
- convex sets
- semidefinite program
- norm minimization
- optimal solution
- operator splitting
- image denoising
- active learning
- feature space