A stochastic extra-step quasi-Newton method for nonsmooth nonconvex optimization

Keyphrases

• High-order 100%
• Nonsmooth Optimization 100%
• Quasi-Newton Method 100%
• Numerical Results 50%
• Learning Problems 50%
• State-of-the-art Techniques 50%
• Logistic Regression 50%
• Gradient Step 50%
• Variance Reduction 50%
• Nonsmooth 50%
• Large-scale Data 50%
• Smooth Part 50%
• Global Convergence 50%
• Regression Learning 50%
• Deep Learning 50%
• Fixed Point Equation 50%
• Proximal Methods 50%
• Big Data Applications 50%
• Convex Composite Optimization 50%
• Convergence to Stationary Points 50%
• Stochastic Oracles 50%
• Composite Optimization 50%
• Quasi-Newton Scheme 50%
• Stochastic Proximal Gradient 50%

Mathematics

• Stochastics 100%
• Quasi-Newton Method 100%
• Fixed Points 14%
• Stationary Point 14%
• Step Size 14%
• Objective Function 14%
• Logistic Regression 14%
• Variance Reduction Technique 14%
• Point Equation 14%
• Deep Learning Method 14%