Group-wise dynamic dropout based on latent semantic variations

Keyphrases

• Dropout 100%
• Group-wise 100%
• Semantic Variation 100%
• Latent Semantic 100%
• Adaptive Dropout 66%
• Co-adaptation 66%
• Deep Neural Network 33%
• Public Dataset 33%
• Classification Performance 33%
• Overfitting 33%
• Local Features 33%
• Semantic Information 33%
• Hidden Unit 33%
• CIFAR-10 33%
• Probability-based 33%
• Adversarial Examples 33%
• Network Robustness 33%
• CIFAR-100 33%
• Fashion-MNIST 33%
• Dropout Regularization 33%
• Feature Detector 33%
• Feature Discriminability 33%
• Feature Density 33%

Computer Science

• Deep Neural Network 100%
• Regularization 100%
• Classification Performance 100%
• local feature 100%
• Discriminability 100%
• Adversarial Example 100%
• Feature Detector 100%
• Degraded Information 100%