Composite difference-max programs for modern statistical estimation problems

Keyphrases

• Statistical Estimation 100%
• Estimation Problem 100%
• Loss Function 66%
• Regularizer 66%
• Difference of Convex Functions 66%
• Piecewise Affine Regression 66%
• Statistical Model 33%
• Optimization Problem 33%
• Nonconvex 33%
• Deep Learning 33%
• Composite Structures 33%
• Major Components 33%
• Numerical Results 33%
• Activation Function 33%
• Overfitting 33%
• Stationary Solution 33%
• Convergence Results 33%
• Output Variable 33%
• Non-monotone 33%
• Component Functions 33%
• Nondifferentiable Optimization 33%
• Majorization-minimization 33%
• Empirical Risk Minimization 33%
• Piecewise Affine 33%
• Semismooth Newton Method 33%
• Majorization-minimization Algorithm 33%

Mathematics

• Loss Function 100%
• Majorization 100%
• Linear Models 50%
• Pointwise 50%
• Statistical Model 50%
• Superiority 50%
• Convergence Result 50%
• Stationary Solution 50%
• Convex Programming 50%
• Empirical Risk Minimization 50%
• Component Function 50%
• Nondifferentiable Optimization 50%
• Data Point 50%
• Newton's Method 50%
• Input Feature 50%
• Deep Learning Method 50%
• Subproblem 50%

Engineering

• Max 100%
• Loss Function 100%
• Directional 50%
• Optimisation Problem 50%
• Newton's Method 50%
• Major Component 50%
• Predicted Model 50%
• Illustrates 50%
• Model Parameter 50%
• Data Point 50%
• Subproblem 50%
• Input Feature 50%
• Activation Function 50%
• Majorization-Minimization Algorithm 50%
• Deep Learning Method 50%
• Statistical Model 50%