Global linear convergence of Newton's method without strong-convexity or Lipschitz gradients.
Sai Praneeth KarimireddySebastian U. StichMartin JaggiPublished in: CoRR (2018)
Keyphrases
- high accuracy
- cost function
- detection method
- segmentation method
- convergence rate
- closed form
- dynamic programming
- computational cost
- image registration
- classification method
- support vector machine svm
- probabilistic model
- computational complexity
- face recognition
- synthetic data
- experimental evaluation
- significant improvement
- preprocessing