Efficient parallel Stochastic Gradient Descent for matrix factorization using GPU.
Mohamed A. NassarLayla A. A. El-SayedYousry TahaPublished in: ICITST (2016)
Keyphrases
- matrix factorization
- stochastic gradient descent
- collaborative filtering
- loss function
- low rank
- recommender systems
- nonnegative matrix factorization
- negative matrix factorization
- least squares
- random forests
- step size
- probabilistic model
- importance sampling
- missing data
- lower bound
- regularization parameter
- linear combination
- weight vector
- maximum likelihood
- training data