Engineering AI Systems and AI for Engineering: Compositionality and Physics in Learning

Explore the intersection of artificial intelligence and engineering systems in this insightful talk by Cyrus Neary. Delve into innovative approaches for transforming AI and machine learning capabilities into practical engineering solutions for robotics and autonomy. Discover compositional methodologies for AI system design that enable modular development and testing, facilitating reliable deployment in real-world scenarios. Learn about control-oriented learning algorithms that integrate data with physics knowledge, resulting in systems capable of effectively controlling hardware after minimal training. Examine experimental results on various robot platforms, from ground vehicles to hexacopters, showcasing the rapid and dependable transfer of simulation-and-data-driven AI algorithms to real-world environments. Gain valuable insights from Neary's research on creating engineering methodologies for AI-driven systems and developing learning algorithms tailored to the unique characteristics of engineering problems.