Guaranteed Convergence of Training Convolutional Neural Networks via Accelerated Gradient Descent

Keyphrases

• Accelerated Gradient Method 100%
• Activation Function 33%
• Convolutional Neural Network 100%
• Feature Vector 33%
• Gaussian Distribution 33%
• Gradient Method 66%
• Guaranteed Convergence 100%
• Hidden Layer 33%
• Linear Regression 33%
• Loss Function 33%
• Non-overlapping 33%
• Nonconvex Optimization 33%
• Object Function 33%
• Optimization Problem 33%
• Rectified Linear Unit (ReLU) 33%
• Regression Problem 33%
• Training Data 33%
• Unknown Weight 33%

Computer Science

• Activation Function 33%
• Algorithm Converges 33%
• Convex Optimization 33%
• Convolutional Neural Network 100%
• Feature Vector 33%
• Gradient Descent 100%
• Object Function 33%
• Optimization Problem 66%
• Regression Problem 33%
• Training Data 33%

Mathematics

• Convolutional Neural Network 100%
• Feature Vector 50%
• Gaussian Distribution 50%
• Initial Point 50%
• Input Feature 50%
• Linear Regression 50%
• Loss Function 50%
• Training Data 50%