Formalizing Generalization and Adversarial Robustness of Neural Networks to Weight Perturbations

Keyphrases

• Neural Network 100%
• Adversarial Robustness 100%
• Weight Perturbation 100%
• Loss Function 20%
• Model Performance 20%
• Active Research 20%
• Feedforward Neural Network 20%
• Gap Measurement 20%
• Model Compression 20%
• Machine Learning Tasks 20%
• Theory-driven 20%
• Adversarial Attack 20%
• First Integral 20%
• Generalization Gap 20%
• Generalization Behavior 20%
• Robust Neural Networks 20%

Computer Science

• Neural Network 100%
• Research Topic 25%
• Feed Forward Neural Networks 25%
• Performance Model 25%
• Active Research 25%
• Model Compression 25%
• Adversarial Machine Learning 25%
• Machine Learning 25%
• Learning System 25%

Mathematics

• Neural Network 100%
• Wide Range 20%
• Loss Function 20%
• Learning Task 20%
• Performance Model 20%