Preconditioned Gradient Descent for Over-Parameterized Nonconvex Matrix Factorization.
Jialun ZhangSalar FattahiRichard Y. ZhangPublished in: NeurIPS (2021)
Keyphrases
- matrix factorization
- stochastic gradient descent
- objective function
- iterative methods
- collaborative filtering
- low rank
- low rank matrices
- cost function
- recommender systems
- missing data
- nonnegative matrix factorization
- factorization methods
- negative matrix factorization
- convex optimization
- factor analysis
- data sparsity
- variational bayesian
- probabilistic matrix factorization
- data matrix
- factorization method
- tensor factorization
- item recommendation
- matrix completion
- loss function
- personalized ranking
- probabilistic model
- conjugate gradient
- latent factor models
- face recognition