An adaptive algorithm for high-dimensional integrals on heterogeneous CPU-GPU systems.
Giuliano LaccettiMarco LapegnaValeria MeleRaffaele MontellaPublished in: Concurr. Comput. Pract. Exp. (2019)
Keyphrases
- high dimensional
- dynamic programming
- learning algorithm
- preprocessing
- k means
- graphics processing units
- parallel implementation
- times faster
- high accuracy
- experimental evaluation
- matching algorithm
- np hard
- significant improvement
- search space
- optimal solution
- real time
- computational cost
- gpu implementation
- dimensionality reduction
- objective function
- computational complexity
- noisy data
- gpu accelerated
- cost function
- decision trees
- recognition algorithm
- detection algorithm
- segmentation algorithm
- lower bound
- general purpose