The Global R-linear Convergence of Nesterov's Accelerated Gradient Method with Unknown Strongly Convex Parameter.
Chenglong BaoLiang ChenJiahong LiPublished in: CoRR (2023)
Keyphrases
- gradient method
- convergence rate
- convex formulation
- log likelihood function
- semidefinite programming
- semidefinite
- step size
- primal dual
- convex concave
- convergence speed
- learning rate
- optimization methods
- negative matrix factorization
- saddle point
- convex sets
- natural gradient learning
- variational inequalities
- potts model
- convex relaxation
- image classification
- linear programming