Neural Networks can Learn Representations with Gradient Descent

Keyphrases

• Neural Network 100%
• Gradient Descent 100%
• Sample Complexity 50%
• Kernel Methods 50%
• Class Function 25%
• Large Classes 25%
• Transfer Learning 25%
• Learning Representations 25%
• Two-layer Neural Network 25%
• Efficient Transfer Learning 25%

Computer Science

• Neural Network 100%
• Gradient Descent 100%
• Kernel Method 50%
• Transfer Learning 50%
• Representation Learning 25%
• Trained Neural Network 25%
• Layer Neural Network 25%

Mathematics

• Neural Network 100%
• Kernel Method 50%
• Transfer Learning 50%
• Polynomial 25%
• Representation Learning 25%