LEAPSANDBOUNDS: A Method for Approximately Optimal Algorithm Configuration.
Gellért WeiszAndrás GyörgyCsaba SzepesváriPublished in: ICML (2018)
Keyphrases
- high accuracy
- dynamic programming
- experimental evaluation
- cost function
- computational cost
- preprocessing
- significant improvement
- improved algorithm
- optimization algorithm
- k means
- segmentation method
- synthetic and real images
- detection algorithm
- high efficiency
- computational complexity
- computationally efficient
- clustering method
- objective function
- recognition algorithm
- detection method
- support vector machine svm
- tree structure
- computational efficiency
- single pass
- input data
- segmentation algorithm
- optimization method
- convergence rate
- estimation algorithm
- approximately optimal
- matching algorithm
- similarity measure
- classification algorithm
- theoretical analysis
- combinatorial optimization
- np hard
- energy function
- selection algorithm
- reconstruction method
- learning algorithm
- classification method
- probabilistic model
- search space
- linear programming
- support vector machine
- clustering algorithm