стажировка

[Vladimir Dobrynin]

Добрый день,

Я получмл персональное письмо от Константина Авраченкова из INRIA с
просьбой прорекламировать его тему, предлагаемую для стажировки для
наших студентов и аспирантов. Вот текст с описанием темы:

-----------------------------------------------------------
...
This year I've submitted a topic on large graph clustering
for the INRIA Internship programme (see the description below).

Please advertise it among the students.

With best wishes,
Konstantin

---

Topic number: 86

Title: Clustering large graphs by random walk methods

Abstract:
A goal of the internship is to develop and test random walk
based algorithms for clustering very large graphs. The main
idea is to run random walks from candidate centroid nodes to
detect tight clusters around these centroid nodes. Some variants
of the proposed approach can be found in the article [1].
The main advantage of the random walk clustering methods is
their low complexity. This fact allows one to apply the random walk
based methods for clustering very large graphs. Possible test
examples are the Web graph and the Wikipedia graph.

The candidate is expected to have a sound knowledge in Graph
Theory and Probability Theory and a practical knowledge in
Java and MATLAB.

The internship is intended for Master or PhD students and its
duration can be between 3 and 6 months.

[1] K. Avrachenkov, V. Dobrynin, D. Nemirovsky, S. Pham,
E. Smirnova, \"Pagerank based clustering of hypertext document
collections\", in Proceedings of ACM SIGIR 2008.

Any additional information can be obtained by e-mail
K.Avra...@sophia.inria.fr