November 12, 2015: Seminar - Leman Akoglu: "3D?s of Anomaly Mining in Complex Graphs: Definition, Detection and Description"


3D?s of Anomaly Mining in Complex Graphs: Definition, Detection and Description

Dr. Leman Akoglu
Stony Brook University
November 12, 2015
11:00 a.m., Room 636 SEO Building


Anomaly mining is the task of finding irregularities in large complex datasets and finds a plethora of applications in security, finance, astronomy, biology, and so on. Despite its immense popularity, it is quite often challenging to define what an anomaly is, how to effectively detect it, and further how to describe or justify it.

In this talk I will introduce new definitions and descriptive techniques for anomaly mining in large complex graphs. First, I will focus on dynamic graphs and discuss recent work on ensemble methods for finding change points in time-evolving graphs. This work addresses challenges regarding how to integrate multiple heterogeneous detectors and how to assess their performance in order to effectively harness their output. I will follow up with a new model and algorithm to summarize individual node anomalies through the groups that they form in the graph. Next, I will shift focus to attributed graphs, and present new approaches for finding community outliers and anomalous neighborhoods. Our methods aim not only to?detect?anomalies, but also to be able to?describe?them, which helps the end users in explaining-away and sense-making of the detected anomalies.


Leman Akoglu is an Assistant Professor in the Department of Computer Science at Stony Brook University. She received her Ph.D. from the Computer Science Department at Carnegie Mellon University in 2012. She also spent summers at IBM T. J. Watson Research Labs and Microsoft Research at Redmond. Her research interests span a wide range of data mining and machine learning topics with a focus on algorithmic problems arising in graph mining, pattern discovery, social and information networks, and especially anomaly mining; outlier, fraud, and event detection. Dr. Akoglu's research has won 4 publication awards; Best Research Paper at SIAM SDM 2015, Best Paper at ADC 2014, Best Paper at PAKDD 2010, and Best Knowledge Discovery Paper at ECML/PKDD 2009. She also holds 3 U.S. patents filed by IBM T. J. Watson Research Labs. Dr. Akoglu is a recipient of the NSF CAREER award (2015) and Army Research Office Young Investigator award (2013). Her research is currently supported by the National Science Foundation, the US Army Research Office, DARPA, a gift from Northrop Grumman Aerospace Systems, and a gift from Facebook. More details can be found at?

Copyright 2016 The Board of Trustees
of the University of
Helping Women Faculty Advance
Funded by NSF