TITLE: PageRank: some methods, models and regularizations

SPEAKER:  Prof. Boris Polyak

ABSTRACT:

PageRank problem is one of challenging problems in information technologies and numerical analysis due to its huge dimension and wide range of applications. It also attracts great attention of experts in control theory; it is closely related to consensus in multiagent systems. The traditional approach to PR goes back to the pioneering paper by Brin and Page. The original problem is replaced with finding the eigenvector of the modified matrix which can be

 effectively solved by the power method. In this paper we demonstrate that the solution of the modified problem can be far enough from the original one and propose an iterative regularization method which allows to find the desired solution. We also propose an l1-regularization; its solution of the PR problem ignores low-ranking pages. All methods are illustrated on two examples of PR problems which have many attractive features as simulation tests.

 Bio:

 Boris Polyak currently serves as the head of Adaptive and Robust Control Systems Laboratory at the Institute for Control Sciences, Russian Academy of Sciences and as a professor at the Department of Engineering and Cybernetics of Moscow Institute of Physics and Technology. He earned his PhD in Mathematics (1963) from Moscow State University and Doctor of Science degree in Engineering (1978) from the Institute of Control Sciences. After working as a researcher at the Moscow State University for several years, Boris joined the Institute of Control Sciences at 1971. He is an author of 4 monographs and more than 180 journal papers in different fields including mathematical programming, control, numerical analysis, stochastic optimization, and statistics. Boris is an IFAC fellow. His numerous awards include Meyerhoff Fellowship in 1991 and Andronov's Award of Russian Academy of Science in 1994. Last but not the least, he is a nice and cheerful fellow.