Alternating Optimization Method Based on Nonnegative Matrix Factorizations for Deep Neural Networks.
Tetsuya SakuraiAkira ImakuraYuto InoueYasunori FutamuraPublished in: ICONIP (4) (2016)
Keyphrases
- optimization method
- matrix factorization
- nonnegative matrix factorization
- neural network
- genetic algorithm
- tensor factorization
- factor matrices
- optimization algorithm
- data matrix
- optimization methods
- particle swarm
- negative matrix factorization
- collaborative filtering
- evolutionary algorithm
- simulated annealing
- optimization process
- neural network training
- recommender systems
- artificial neural networks
- differential evolution
- low rank
- data representation
- global optimum
- missing data
- optimization procedure
- metaheuristic
- factorization methods
- nonlinear optimization
- nelder mead simplex
- multi objective
- objective function
- back propagation
- quasi newton
- least squares
- binary matrix
- neural network model
- genetic algorithm ga
- fitness function
- mixture model