Explainability of Image Classifiers for Targeted Adversarial Attack

Mathematics

• Classifier 100%
• Attack 90%
• Interpretability 58%
• Neural Networks 41%
• Model 29%
• Real-world Applications 28%
• Computer Vision 26%
• Learning Algorithm 26%
• Architecture 24%
• Prey 23%
• Robustness 18%
• Class 6%

Engineering & Materials Science

• Classifiers 62%
• Deep neural networks 43%
• Computer vision 18%
• Learning algorithms 18%