My current research is primarily focused on Trustworthy AI and Safe Autonomy.
Autonomous systems increasingly rely on machine learning for making decisions. When such systems are used in safety-critical
applications such as driverless cars and medical devices, we need guarantees regarding their safety. This motivates research
on integrating logical reasoning in learning algorithms. Representative research directions are:
How to formally verify closed-loop systems that include a neural-network-based controller? Sample publications: Verisig, EMSOFT'21; Talk (Jan 20)
At the undergraduate level, I teach
CIS 2620: Automata, Computability, and Complexity, usually each Fall semester.
To help students learn the basic model of deterministic finite automata, we have developed
an interactive tool AutomataTutor with novel features
for automatically generated feedback and automatic grading.
If you are an instructor teaching a course on this topic,
please encourage your students to use this tool, and if you want to set up
a homework assignment using our tool, send us an email!
To train Masters students in the cross-disciplinary skills required for the emerging area of embedded
and cyber-physical systems, Penn Engineering started a new program
EMBS that I directed for many years.
For many years, I taught CIS 540: Principles of Embedded Computation
for this program. On this topic, I have written the textbook
Principles of Cyber-Physical Systems which was
published by MIT Press in April 2015. There are lots of teaching resources such as
exercises, solutions, slides, and projects, available if you want to teach
a similar course.