An efficient method for adapting step-size parameters of primal-dual hybrid gradient method in application to total variation regularization.
Tatsuya YokotaHidekata HontaniPublished in: APSIPA (2017)
Keyphrases
- gradient method
- step size
- convergence rate
- cost function
- primal dual
- negative matrix factorization
- convex optimization
- computational complexity
- constrained optimization
- special case
- similarity measure
- optimization method
- cross validation
- optimization algorithm
- convergence speed
- linear programming
- dynamic programming
- regularization parameter
- pairwise
- total variation regularization