Abstract:  High-performance cyber-physical systems rely on many sensors and hardware components for successful operation. Control strategies for these devices require an understanding of how unpredictability in these components might impair performance. In this talk, we aim to quantify the informational bottlenecks imposed by uncertain system models. We will compare systems with unreliable sensing (e.g., miscalibrated cameras) to systems with unreliable actuation (e.g., when motors on a drone cannot precisely execute control actions) as well as systems with unreliable growth (e.g. modeling or sampling errors). We can use this to quantify the value of side-information regarding the uncertainty in the system (in bits), in order to answer questions such as: “what is the value of adding an extra sensor to the system?”

The talk will include joint work with Jian Ding, Victoria Kostina, Yuval Peres, Mark Sellke and Alex Zhai.

Gireeja Ranade is an assistant teaching professor in the Department of Electrical Engineering and Computer Sciences at the University of California, Berkeley.

Before joining the faculty at UC Berkeley, Dr. Ranade was a Researcher at Microsoft Research AI in the Adaptive Systems and Interaction Group. She also designed and taught the first offering for the new course sequence EECS16A and EECS16B in the EECS department at UC Berkeley and received the 2017 UC Berkeley Electrical Engineering Award for Outstanding Teaching.

Dr. Ranade received her PhD in Electrical Engineering and Computer Science from the University of California, Berkeley, and her undergraduate degree from MIT in Cambridge, MA.