Overview
Explore the OWASP Top 10 for Machine Learning Security in this 57-minute conference talk from DevSecCon. Gain practical insights into key security challenges and best practices specific to machine learning. Delve into an in-depth overview of each of the top ten vulnerabilities, including input manipulation, data poisoning, model inversion, model stealing, AI supply chain attacks, transfer learning attacks, model skewing, output integrity attacks, and model poisoning. Learn from real-world examples and case studies illustrating how these vulnerabilities manifest. Discover actionable recommendations for mitigating risks and implementing strategies to ensure robust and secure ML deployments. Equip yourself with essential knowledge to enhance the security posture of machine learning projects, whether you're a developer, data scientist, or security professional.
Syllabus
Introduction
Input Manipulation Attack
Data Poisoning Attack
Model Inversion Attack
Model Stealing
AI Supply Chain Attack
Transfer Learning Attack
Model Skewing Attack
Output Integrity Attack
Model Poisoning Attack
Conclusion
Taught by
DevSecCon