AI for Safety Critical Control

Overview

Explore theoretical foundations of AI in safety-critical control systems with Claire Tomlin from UC Berkeley, focusing on trustworthiness in high-risk applications.

Syllabus

• Introduction to AI for Safety Critical Control
Overview of safety-critical systems and their importance
Introduction to trustworthiness in AI-driven applications
• Theoretical Foundations
Basics of control systems and AI intersections
Introduction to dynamical systems
Stability and safety in control systems
• AI Techniques in Control
Machine learning methods for control systems
Reinforcement learning in safety-critical environments
Model predictive control using AI
• Trustworthiness and Reliability
Defining trustworthiness in AI
Verifiable AI methods
Assurance cases and argumentation frameworks
• Risk Analysis and Management
Risk assessment techniques in AI control systems
Mitigation strategies for AI-induced risks
• Human-AI Interaction
Human factors in AI control loop
Designing for human oversight and intervention
• Applications in High-Risk Sectors
AI in aerospace and automotive systems
AI-driven medical devices
Robotics and automation in safety-critical environments
• Case Studies and Real-World Examples
Success stories and lessons learned
Failures and their implications for AI trustworthiness
• Future Trends and Research Directions
Emerging techniques and technologies
Policy and ethical considerations in AI safety
• Course Conclusion
Summary of key learning points
Final project presentations and discussions
• Additional Resources
Recommended readings
Online tools and platforms for further learning

Subjects

Computer Science