Login
Talks & Seminars
Title: Approximating Probabilistic Inference Without Losing Guarantees: Combining Hashing and Feasibility
Mr. Kuldeep Meel, Rice University
Date & Time: August 20, 2014 15:30
Venue: Lecture Hall, B Block, 02nd/03rd Floor, Department of Computer Science and Engineering, Kanwal Rekhi (KReSIT) Building
Abstract:
With the rise of big data and complex modeling, probabilistic inference (i.e. computing probability of an event given some observation) has emerged as a key problem. The current state of art techniques are either exact and face scalability issues or provide very weak approximation guarantees. We introduce a new computational paradigm, which makes only few feasibility queries on models augmented with random constraints, that provides (ϵ-δ) guarantees and scalable practical algorithms. Our new approach builds upon prior work in computational theory and techniques based on universal hashing. I will discuss how we can further use these techniques to generate samples from complex distributions which has been key to constrained-random verification over last two decades.
Speaker Profile:
Kuldeep Meel is a PhD student in Rice working with Prof. Moshe Vardi and Prof. Supratik Chakraborty. His research broadly falls into the intersection of probabilistic reasoning, computer-aided verification and formal methods. He is the recipient of 2013-14 Andrew Ladd Fellowship. He received his M.S. from Rice in 2014 and B.Tech. from IIT Bombay in 2012.
List of Talks

Webmail

Username:
Password:
Faculty CSE IT
Forgot Password
    [+] Sitemap     Feedback