About Me

Why I do control theory


The age of AI. We live in the formative, productive age of control, machine learning, and artificial intelligence. These automation technologies underlie systems that are real (like robots, self-driving cars, and rockets), and virtual (scheduling, financial, and logistics systems). Many of these systems have nontrivial dynamics, and interact with the world in rich and interesting ways.

System-level thinking. The problem is many of our autonomous systems have become so sophisticated and so black-box that we have hit a complexity roadblock: nobody can keep track of it all. For example, it can be difficult to tell why a vision classifier or recommendation engine based on machine learning works — or fails to work — in any given application. Worse, when an algorithm does work, quantifying its limits when placed in the loop, as part of a larger safety-critical system, can be impossible.


Provable guarantees. We can do better. I believe we can make provably safe and reliable autonomous systems through proper and broad application of control theory and formal methods. These tools make it possible to know everything about the systems we create, because they force designers to work with well defined, constructive abstractions of their autonomous systems. If automation is our future, then further development of control theory and formal methods inside and outside the academic context is essential for continued safe technological advancement.

My role. Because I believe in the good that is possible when we slow down and do it right, I spend my days advancing the state of the art in theory, tools, and education related to safe and reliable autonomy.