Title: Learning Hybrid Systems for Fun and Profit

Hybrid system identification techniques seek hybrid system models  of various forms that can approximate given observation data involving the states and outputs of the system.  They promise to derive relatively simple dynamical models that can be interpreted and analyzed using many of the available tools for safety, stability and controller synthesis developed by the HSCC community. Although hybrid system identification techniques have been well-studied,  the problem itself is known to be computationally hard. We motivate the continued need for efficient algorithms for identifying hybrid system models,  despite the successes enjoyed by neural network-based dynamical models identified using  variants of stochastic gradient descent. We present some recent results that combine  ideas from  areas such as approximation algorithms and non-smooth optimization.  Finally, we will conclude by examining some of the key open problems in this area.

Based on joint work with Guillaume Berger, Monal Narasimhamurthy, Kandai Watanabe and Morteza Lahijanian.

Biography:

Sriram Sankaranarayanan is a professor of Computer Science at the University of Colorado, Boulder. His research interests include automatic techniques for reasoning about the behavior of computer and cyber-physical systems. Sriram obtained a PhD in 2005 from Stanford University where he was advised by Zohar Manna and Henny Sipma. Subsequently he worked as a research staff member at NEC research labs in Princeton, NJ. He has been on the faculty at CU Boulder since 2009. Sriram has been awarded the prestigious Innovation Award from Coursera for his online “Data Structures and Algorithms” specialization, which has taught the fundamentals of data structures and algorithms with an emphasis on applications in data science to more than 10,000 students.