An Investigation on Fragility of Machine Learning Classifiers in Android Malware Detection

Keyphrases

• Machine Learning Classifiers 100%
• Android Malware Detection 100%
• Oracle 50%
• Generative Adversarial Networks 50%
• Adversarial Training 50%
• Adversarial Samples 50%
• Adversarial Evasion Attack 50%
• Fool 25%
• Static Analysis 25%
• Detection Accuracy 25%
• Detection Efficiency 25%
• Machine Learning Based 25%
• Attacker 25%
• Linear Classifier 25%
• Learning Classifiers 25%
• Android Application 25%
• Network Applications 25%
• Training Scheme 25%
• Proposed Methodology 25%
• Mimicry Attacks 25%
• Public Dataset 25%
• Persuasion Detection 25%

Computer Science

• Android Malware 100%
• Malware Detection 100%
• Machine Learning 100%
• Learning System 100%
• Generative Adversarial Networks 40%
• Adversarial Machine Learning 40%
• Experimental Result 20%
• Static Program Analysis 20%
• Attackers 20%
• Detection Accuracy 20%
• Android Application 20%
• Linear Classifier 20%