Cubic Regularized Newton Method for the Saddle Point Models: A Global and Local Convergence Analysis

Keyphrases

• Point Model 100%
• Convex-concave 100%
• Lipschitz 100%
• Type Errors 100%
• Error Bound Property 100%
• Homotopy Continuation 100%
• Regularized Newton Method 100%
• Global Convergence Analysis 100%
• Local Convergence Analysis 100%
• Iteration Complexity Bound 100%
• Continuation Approach 50%
• Search Direction 50%
• Path Following 50%
• Point Function 50%
• Saddle Point Problem 50%
• Strongly Convex 50%
• Superlinear Convergence Rate 50%
• Homotopy Path 50%
• Minimax Saddle-point 50%
• Local Superlinear Convergence 50%

Mathematics

• Point Model 100%
• Local Convergence Analysis 100%
• Newton's Method 100%
• Saddle Point 100%
• Error Bound 40%
• Homotopy 40%
• Point Function 20%
• Minimax 20%
• Convergence Rate 20%
• Search Direction 20%
• Subproblem 20%