Algorithms and Limits in Statistical Inference

Speaker: Jayadev Acharya
Affiliation: Post-Doc Researcher - MIT

Abstract: The massive explosion of data and the demand for fast inference has resulted in great interest and progress in various aspects of data science. I will discuss some problems at the heart of statistics and machine learning, such as hypothesis testing, distribution learning, and property estimation. I will contrast the classical themes and modern considerations in addressing these questions, drawing tools from information theory, computer science, machine learning, and statistics.

I will describe the problem of distribution estimation in greater detail, including a general framework that yields computationally efficient and statistically optimal algorithms for estimating a wide range of probability density functions. I will then describe some of the future directions that I am interested in.

Biography: Jayadev Acharya is a postdoctoral researcher in the EECS department at MIT, hosted by Piotr Indyk. His current research interests are in algorithmic statistics, machine learning, and information theory. He received his Ph.D. in ECE from UC San Diego, where he was advised by Alon Orlitsky. He is a recipient of the Jack Wolf student paper award at the IEEE International Symposium on Information Theory, the Shannon Graduate Fellowship from UC San Diego, and an MIT-Energy Initiative fellowship.

For more information, contact Prof. Christina Fragouli (christina.fragouli@ucla.edu)

Date/Time:
Date(s) - Mar 10, 2016
11:00 am - 12:15 pm

Location:
EE-IV Shannon Room #54-134
420 Westwood Plaza - 5th Flr., Los Angeles CA 90095