Dapadv: Differentiated adversarial perturbation generation method in problem space for android malware detection

Key Points

• The method improves android malware detection accuracy through adversarial perturbation.
• Utilizing differentiated generation, the detection algorithm achieves a notable performance increase.
• The approach leverages nuances in problem space to generate effective perturbations for enhancing detection.
• Implications for this methodology may lead to more resilient malware defenses in mobile applications.