RobCaps: Evaluating the Robustness of Capsule Networks against Affine Transformations and Adversarial Attacks

Keyphrases

• Affine Transformation 100%
• Adversarial Attack 100%
• Capsule Network 100%
• Convolutional Neural Network 50%
• Comprehensive Comparison 16%
• Affine 16%
• Image Classification 16%
• Classification Task 16%
• Multi-object 16%
• MNIST 16%
• Number of Parameters 16%
• Safety-critical Applications 16%
• Hierarchical Feature Learning 16%
• German Traffic Sign Recognition Benchmark (GTSRB) 16%
• Dynamic Routing 16%
• Network Robustness 16%
• CIFAR10 Dataset 16%
• Convolutional Neural Network Model 16%
• Adversarial Examples 16%
• CapsNet 16%
• Input Transformation 16%

Computer Science

• Convolutional Neural Network 100%
• Affine Transformation 100%
• Adversarial Machine Learning 100%
• Dynamic Routing 25%
• Image Classification 25%
• Classification Task 25%
• Representation Learning 25%
• Neural Network Model 25%
• Critical Application 25%
• Adversarial Example 25%