Course Outline
Introduction to AI Threat Modeling
- Factors contributing to AI system vulnerability.
- Comparison of the AI attack surface with that of traditional systems.
- Primary attack vectors across data, model, output, and interface layers.
Adversarial Attacks on AI Models
- Exploration of adversarial examples and perturbation techniques.
- Distinction between white-box and black-box attacks.
- Analysis of FGSM, PGD, and DeepFool methods.
- Visualization and creation of adversarial samples.
Model Inversion and Privacy Leakage
- Techniques for inferring training data from model outputs.
- Understanding membership inference attacks.
- Privacy risks associated with classification and generative models.
Data Poisoning and Backdoor Injections
- Impact of poisoned data on model behavior.
- Mechanisms of trigger-based backdoors and Trojan attacks.
- Strategies for detection and data sanitization.
Robustness and Defense Techniques
- Application of adversarial training and data augmentation.
- Utilization of gradient masking and input preprocessing.
- Implementation of model smoothing and regularization techniques.
Privacy-Preserving AI Defenses
- Overview of differential privacy.
- Management of noise injection and privacy budgets.
- Integration of federated learning and secure aggregation.
AI Security in Practice
- Execution of threat-aware model evaluation and deployment.
- Practical use of ART (Adversarial Robustness Toolbox) in real-world settings.
- Examination of industry case studies involving real-world breaches and mitigation strategies.
Summary and Next Steps
Requirements
- Knowledge of machine learning workflows and model training processes.
- Proficiency in Python and common ML frameworks such as PyTorch or TensorFlow.
- Basic familiarity with security or threat modeling concepts is beneficial.
Audience
- Machine learning engineers.
- Cybersecurity analysts.
- AI researchers and model validation teams.
Testimonials (3)
inventory and identifying the different risk exposures within AI
Gary Cook - Cybersecurity and Information Technology Risk Division
Course - Introduction to AI Trust, Risk, and Security Management (AI TRiSM)
I really enjoyed learning about AI attacks and the tools out there to begin practicing and actively using for security testing. I took a lot of knowledge away which I didn't have at the beginning and the course met what I hoped it would be. My favorite part shown from the training was Comet Browser and was amazed at what it could do. Definitely something will be looking into more. Overall it was a great course and enjoyed learning all OWASP GenAI Top 10.
Patrick Collins - Optum
Course - OWASP GenAI Security
The profesional knolage and the way how he presented it before us