A Relative Exponential Weighing Algorithm for Adversarial Utility-based Dueling Bandits.
Pratik GajaneTanguy UrvoyFabrice ClérotPublished in: CoRR (2016)
Keyphrases
- preprocessing
- experimental evaluation
- worst case
- objective function
- detection algorithm
- theoretical analysis
- simulated annealing
- np hard
- high accuracy
- linear complexity
- estimation algorithm
- matching algorithm
- expectation maximization
- computational cost
- cost function
- optimization algorithm
- learning algorithm
- image sequences
- recognition algorithm
- path planning
- optimal solution
- k means
- dynamic programming
- artificial neural networks
- probabilistic model
- computationally efficient
- similarity measure
- convergence rate
- linear programming
- improved algorithm
- least squares