Course Outline
Introduction to Edge AI and Embedded Systems
- An overview of Edge AI, including its applications and limitations
- Examination of edge hardware platforms and associated software stacks
- Analysis of security challenges in embedded and decentralized environments for government
Threat Landscape for Edge AI
- Risks associated with physical access and tampering
- Vulnerabilities to adversarial examples and model manipulation
- Concerns regarding data leakage and model inversion threats for government operations
Securing the Model
- Strategies for model hardening and quantization
- Techniques for watermarking and fingerprinting models to ensure integrity
- Methods for defensive distillation and pruning to enhance security
Encrypted Inference and Secure Execution
- Utilization of trusted execution environments (TEEs) for AI applications
- Implementation of secure enclaves and confidential computing solutions
- Application of homomorphic encryption or secure multi-party computation (SMPC) for encrypted inference
Tamper Detection and Device-Level Controls
- Secure boot processes and firmware integrity checks to prevent unauthorized access
- Sensor validation and anomaly detection mechanisms to ensure data accuracy
- Remote attestation and device health monitoring for continuous security oversight
Edge-to-Cloud Security Integration
- Secure data transmission protocols and key management practices
- End-to-end encryption techniques and comprehensive data lifecycle protection
- Cloud AI orchestration strategies that incorporate edge security constraints for government
Best Practices and Risk Mitigation Strategy
- Threat modeling methodologies tailored to edge AI systems
- Security design principles for embedded intelligence in public sector applications
- Incident response frameworks and firmware update management processes
Summary and Next Steps
Requirements
- Knowledge of embedded systems or environments for deploying edge AI
- Experience with Python and machine learning frameworks (e.g., TensorFlow Lite, PyTorch Mobile)
- Basic understanding of cybersecurity principles or IoT threat models
Audience for Government
- Embedded AI developers
- IoT security specialists
- Engineers responsible for deploying machine learning models on edge or resource-constrained devices
Testimonials (2)
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
