InKS, a Programming Model to Decouple Performance from Algorithm in HPC Codes.
Ksander EjjaaouaniOlivier AumageJulien BigotMichel MehrenbergerHitoshi MuraiMasahiro NakaoMitsuhisa SatoPublished in: Euro-Par Workshops (2018)
Keyphrases
- probabilistic model
- mathematical model
- objective function
- algorithm employs
- algorithm has been developed
- cost function
- theoretical analysis
- recognition algorithm
- model free
- classification algorithm
- expectation maximization
- parameter estimation
- final result
- learning algorithm
- em algorithm
- optimal solution
- high level
- computational model
- search space
- kalman filter
- computational cost
- computational complexity
- filtering algorithm
- input data
- preprocessing
- estimation algorithm
- linear model
- optimization method
- hidden markov models
- linear programming
- particle swarm optimization
- closed form
- bayesian framework
- artificial neural networks
- selection algorithm
- tree structure
- significant improvement
- detection algorithm
- optimization algorithm