Local convergence of alternating low-rank optimization methods with overrelaxation.
Ivan V. OseledetsMaxim V. RakhubaAndré UschmajewPublished in: CoRR (2021)
Keyphrases
- optimization methods
- low rank
- global convergence
- quasi newton
- convex optimization
- matrix factorization
- linear combination
- simulated annealing
- missing data
- optimization method
- singular value decomposition
- optimization problems
- matrix completion
- low rank matrix
- high dimensional data
- semi supervised
- high order
- kernel matrix
- trace norm
- rank minimization
- convergence rate
- convex relaxation
- convergence speed
- minimization problems
- image processing
- efficient optimization
- linear program
- optimization algorithm
- feature extraction
- data sets