Why do larger models generalize better? A theoretical perspective via the XOR problem

Keyphrases

• Global Minimum 100%
• Small Networks 100%
• Large Animal Model 100%
• Convolutional Neural Network 100%
• 3-layer 100%
• XOR Problem 100%
• Neural Network 50%
• Max Pooling 50%
• MNIST 50%
• Generalization Performance 50%
• Gradient Descent 50%
• Weighted Clustering 50%
• Convolutional Networks 50%
• ReLU Activation 50%
• Feature Exploration 50%

Computer Science

• Convolutional Neural Network 100%
• Neural Network 50%
• Generalization Performance 50%
• Gradient Descent 50%
• Convolutional Network 50%

Mathematics

• XOR 100%
• Convolutional Neural Network 100%
• Neural Network 50%