Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Duration 21 hours
Course Outline
Introduction to Security in TinyML
- Security challenges inherent to resource-constrained ML systems
- Threat modeling for TinyML deployments
- Risk categorization for embedded AI applications
Data Privacy in Edge AI
- Privacy considerations for on-device data processing
- Strategies to minimize data exposure and transfer
- Approaches for decentralized data handling
Adversarial Attacks on TinyML Models
- Threats related to model evasion and poisoning
- Input manipulation via embedded sensors
- Evaluating vulnerabilities in constrained environments
Security Hardening for Embedded ML
- Firmware and hardware protection layers
- Access control and secure boot mechanisms
- Best practices for safeguarding inference pipelines
Privacy-Preserving TinyML Techniques
- Quantization and model design considerations for privacy
- Methods for on-device data anonymization
- Lightweight encryption and secure computation methods
Secure Deployment and Maintenance
- Secure provisioning of TinyML devices
- OTA update and patching strategies
- Monitoring and incident response at the edge
Testing and Validation of Secure TinyML Systems
- Security and privacy testing frameworks
- Simulating real-world attack scenarios
- Validation and compliance considerations
Case Studies and Applied Scenarios
- Security failures in edge AI ecosystems
- Designing resilient TinyML architectures
- Evaluating trade-offs between performance and protection
Summary and Next Steps
Requirements
- A solid understanding of embedded system architectures
- Practical experience with machine learning workflows
- Foundational knowledge of cybersecurity principles
Target Audience
- Security analysts
- AI developers
- Embedded engineers
Testimonials (1)
The exercises
