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.
Course Outline
Introduction to Privacy-Preserving ML
- Motivations and risks in sensitive data environments
- Overview of privacy-preserving ML techniques
- Threat models and regulatory considerations (e.g., GDPR, HIPAA)
Federated Learning
- Concept and architecture of federated learning
- Client-server synchronization and aggregation
- Implementation using PySyft and Flower
Differential Privacy
- Mathematics of differential privacy
- Applying DP in data queries and model training
- Using Opacus and TensorFlow Privacy
Secure Multiparty Computation (SMPC)
- SMPC protocols and use cases
- Encryption-based vs secret-sharing approaches
- Secure computation workflows with CrypTen or PySyft
Homomorphic Encryption
- Fully vs partially homomorphic encryption
- Encrypted inference for sensitive workloads
- Hands-on with TenSEAL and Microsoft SEAL
Applications and Industry Case Studies
- Privacy in healthcare: federated learning for medical AI
- Secure collaboration in finance: risk models and compliance
- Defense and government use cases
Summary and Next Steps
Requirements
- An understanding of machine learning principles
- Experience with Python and ML libraries (e.g., PyTorch, TensorFlow)
- Familiarity with data privacy or cybersecurity concepts is helpful
Audience
- AI researchers
- Data protection and privacy compliance teams
- Security engineers working in regulated industries
14 Hours
