Randomly Initialized Alternating Least Squares: Fast Convergence for Matrix Sensing.
Kiryung LeeDominik StögerPublished in: SIAM J. Math. Data Sci. (2023)
Keyphrases
- alternating least squares
- negative matrix factorization
- low rank approximation
- nonnegative matrix factorization
- iterative algorithms
- stochastic gradient descent
- singular value decomposition
- convergence rate
- matrix factorization
- low rank
- step size
- least squares
- subspace learning
- kernel matrix
- convergence speed
- data representation
- spectral clustering
- sparse representation
- regularization parameter