Learning one-hidden-layer neural networks with landscape design

Keyphrases

• Analytic Formula 33%
• Complexity Results 33%
• Finite Sample 33%
• Full Rank 33%
• Gaussian Distribution 33%
• Global Minimum 100%
• Hidden Layer 100%
• Landscape Design 100%
• Local Minima 33%
• Low-rank Tensor 33%
• Neural Network 100%
• Noise Vectors 33%
• Nonconvex Objective Function 33%
• Population Risk 33%
• Ranked Positional Weight 33%
• Sample Complexity 33%
• Squared Loss 33%
• Stochastic Gradient Descent 33%
• Weight Matrix 33%

Mathematics

• Analytic Formula 50%
• Gaussian Distribution 50%
• Local Minimum 50%
• Neural Network 100%
• Objective Function 50%
• Rank Tensor 50%
• Stochastic Property 50%
• Weight Matrix 50%

Computer Science

• Complexity Result 50%
• Gradient Descent 50%
• Layer Neural Network 100%
• Local Minimum 50%
• Objective Function 50%
• Risk Population 50%