On the Generalization Power of Overfitted Two-Layer Neural Tangent Kernel Models

Mathematics

• Number 100%
• Class 100%
• Zeros 50%
• Bounds 50%
• Linear Models 50%
• Upper Bound 50%
• Fourier Matrix 50%
• Gaussian Distribution 50%
• Approach Infinity 50%
• Training Sample 50%
• Characteristics 50%
• Term 50%
• Neural Network 50%

Computer Science

• Model 100%
• Generalization Error 50%
• Class 50%
• Layer Neural Network 25%
• Generalization Performance 25%
• Training Sample 25%