An Optimal Algorithm for Adversarial Bandits with Arbitrary Delays.
Julian ZimmertYevgeny SeldinPublished in: AISTATS (2020)
Keyphrases
- dynamic programming
- detection algorithm
- worst case
- preprocessing
- optimal solution
- experimental evaluation
- improved algorithm
- np hard
- computational cost
- learning algorithm
- computationally efficient
- neural network
- exhaustive search
- significant improvement
- optimization algorithm
- optimality criterion
- recognition algorithm
- monte carlo
- times faster
- k means
- objective function
- globally optimal
- optimal path
- expectation maximization
- closed form
- path planning
- matching algorithm
- tree structure
- segmentation algorithm
- search space
- lower bound
- data structure
- genetic algorithm