Mission:
We study engineering design principles for machines that learn and interact with human beings.
Focused research topics include
- (1) methods for identifying worst-case risks
(e.g., corner cases in self-driving and adversarial attacks against learning machines such as neural networks),
- (2) theories for enabling competence awareness of intelligent systems
(e.g., risk certification of self-driving cars during interactions with human beings),
- and (3) new educational tools to prepare future engineers and designers with
knowledge to conduct risk specifications and design for intelligent engineering products.
Challenges lie across disciplines including system engineering, machine learning,
optimization, optimal control, and game theory.
Real-world values of our research include socially adept human-robot interactions,
accelerated computational material discovery, and robust machine learning.