Stanford University

Invariance and Termination in Probabilistic Programs / Sriram Sankaranarayanan

Sriram Sankaranarayanan
Assistant Professor, University of Colorado Boulder
Sept. 25, 2014

Probabilistic programs involve standard imperative programs augmented with constructs for generating random values according to pre-specified distributions. As such, they arise in many domains including decision support systems, risk models, control systems, sensor fusion and randomized algorithms. In this talk, Sankaranarayanan will examine verification and synthesis problems for probabilistic programs, highlighting recent work in his group that uses martingales and concentration of measure inequalities to analyze the behavior of such programs. He will show show how these mathematical ideas combine to provide us with probabilistic analogues of invariants and ranking functions. He will conclude with a brief discussion that contrasts the symbolic approach presented in this talk with ideas based on statistical hypothesis testing.

The talk will be mostly self-contained. All relevant concepts will be introduced during the talk.

Sriram Sankaranarayanan is an assistant 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 Ph.D. in 2005 from Stanford University and worked subsequently as a research staff member at NEC research labs in Princeton, N.J. He has been on the faculty at CU-Boulder since 2009. Sriram has been the recipient of awards including the President’s Gold Medal from IIT Kharagpur (2000), Siebel Scholarship (2005), the CAREER award from NSF (2009), Dean’s awards for outstanding junior faculty (2012), outstanding teaching (2014), and the Provost faculty achievement award (2014).