Booz Allen Hamilton Colloquium: Edward Raff, "Machine Learning for Cyber Security"
Friday, February 8, 2019
3:30 p.m.-4:30 p.m.
1110 Jeong H. Kim Engineering Building
301 405 4471
Machine Learning for Cyber Security: Challenges and Successes
Senior Lead Scientist,
Booz Allen Hamilton
Talk Abstract: Cyber Security is an important and growing problem, and a considerable amount of research effort has been applied in this space. However, successes can often be thwarted by a number of unique challenges that exist. In particular, many of the common basic assumptions of machine learning models are violated by the data in this space. Further still, the issue of adversarial attacks is made all the more pertinent due to real life adversaries that are writing malware. In this talk we will review a subset of this research for detecting malware, how these domain specific challenges have impacted research, and how results have deviated from expectations.
Speaker Bio: Edward Raff is a Senior Lead Scientist at Booz Allen Hamilton, and a Visiting Professor at the University of Maryland, Baltimore County. He received his Ph.D. in Computer Science from UMBC in 2018, and a M.S. and B.S. from Purdue in 2013 and 2012. Dr. Raff’s research includes work in new methods for malware detection and similarity analysis, high-performance machine learning, biometric fingerprint recognition, and algorithmic fairness. In his spare time, he is also the author of the JSAT machine learning library.