An efficient algorithm for contextual bandits with knapsacks, and an extension to concave objectives.
Shipra AgrawalNikhil R. DevanurLihong LiPublished in: COLT (2016)
Keyphrases
- computational complexity
- cost function
- improved algorithm
- objective function
- knapsack problem
- detection algorithm
- learning algorithm
- expectation maximization
- input data
- optimal solution
- dynamic programming
- computational cost
- probabilistic model
- high accuracy
- computationally efficient
- theoretical analysis
- memory efficient
- np hard
- estimation algorithm
- recognition algorithm
- convergence rate
- matching algorithm
- preprocessing
- optimization algorithm
- tree structure
- segmentation algorithm
- search space
- neural network
- highly efficient
- k means
- least squares