Chairman, Preston M. Green Department of Electrical & System Engineering
The Eugene and Martha Lohman Professor
Washington University in St. Louis, USA
► Personal homepage
Biography:Arye Nehorai (IEEE S'80--M'83--SM'90--F'94) is the Eugene and Martha Lohman Professor and Chair of the Preston M. Green Department of Electrical and Systems Engineering (ESE), Professor in the Department of Biomedical Engineering (by courtesy) and in the Division of Biology and Biomedical Studies (DBBS) at Washington University in St. Louis (WUSTL). He serves as Director of the Center for Sensor Signal and Information Processing at WUSTL. Under his leadership as department chair, the undergraduate enrollment has more than tripled in the last four years. Earlier, he was a faculty member at Yale University and the University of Illinois at Chicago. He received the B.Sc. and M.Sc. degrees from the Technion, Israel and the Ph.D. from Stanford University, California.
Abstract:The last decade has witnessed burgeoning developments in the reconstruction of signals based on exploiting their low-dimensional structures, particularly their sparsity, block-sparsity, and low-rankness. The reconstruction performance of these signals is heavily dependent on the structure of the operating matrix used in sensing. The quality of these matrices in the context of signal recovery is usually quantified by the restricted isometry constant and its variants. However, the restricted isometry constant and its variants are extremely difficult to compute.
We present a framework for analytically computing the performance of the recovery of signals with sparsity structures. We define a family of incoherence measures to quantify the goodness of arbitrary sensing matrices. Our primary contribution is the design of efficient algorithms, based on linear programming and second order cone programming, to compute these incoherence measures. As a by-product, we implement efficient algorithms to verify sufficient conditions for exact signal recovery in the noise-free case. The utility of the proposed incoherence measures lies in their relationship to the performance of reconstruction methods. We derive closed-form expressions of bounds on the recovery errors of convex relaxation algorithms in terms of these measures.
Information and Data Sciences (IDS) Group
Department of Electrical and Computer Engineering
Boston University, USA
► Personal homepage
Biography:Venkatesh Saligrama is a faculty member in the Electrical and Computer Engineering Department at Boston University. He holds a PhD from MIT. His research interests are in Statistical Signal Processing, Statistical Learning, Video Analysis, Information and Decision theory. He has edited a book on Networked Sensing, Information and Control. He has served as an Associate Editor for IEEE Transactions on Information Theory, IEEE Transactions on Signal Processing and has been on Technical Program Committees of several IEEE and Machine Learning conferences. He is the recipient of numerous awards including the Presidential Early Career Award(PECASE), ONR Young Investigator Award, and the NSF Career Award. More information about his work is available here.
Abstract:In a wide spectrum of problems in science and engineering that includes hyperspectral imaging, gene expression analysis, and metabolic networks, the observed data is high-dimensional and can be modeled as arising from an unknown mixture of a small set of unknown shared latent factors. Our approach is based on a natural separability property of the shared latent factors. Our separability property posits that every latent factor contains at least one component that is dominant in that factor. We first establish that this property is not only natural but an inevitable consequence of high-dimensionality, and satisfied by the estimates produced by popular nonparametric Bayes approaches. We show that geometrically these dominant latent factors can be associated with extreme points in a suitable space. We leverage this geometric insight to develop a suite of efficient algorithms for a diverse set of latent variable problems. The proposed random-projections-based algorithm is naturally amenable to a low communication-cost distributed implementation that is attractive for modern web-scale distributed data mining applications. We then establish statistical and computational efficiency guarantees for learning in high-dimensional latent variable models.
Sr. Director of Cyber Security / Head of Security Analytics, Prelert
► Linkedin page
Biography:Oleg currently serves as Senior Director of Cyber Security / Head of Security Analytics for Prelert. He has 15+ years of experience in the area of Information Security, including senior technical, research, and executive roles. Previously, he served as senior director at a security analytics company, where he developed a product to detect Advanced Persistent Threats (APT) using behavioral analytics, big data analysis, and machine learning technology. Earlier, he had also served in director-level roles for more than eight years, building and leading teams of security professionals. His publications include research papers in Anomaly Detection, Network Security, and Intrusion Detection and Prevention in such conferences as IEEE Security & Privacy and USENIX Security. Oleg received his M.S. degree in Information Security from Georgia Tech. He is the author of two issued US patents and a book in Network Security.
Abstract:"In theory, there is no difference between theory and practice. In practice, there is." (Yogi Berra, 1925).
The focus of this talk will be on practical applications of Machine-Learning-based Anomaly Detection (MLAD) in the area of Information Security/Cyber Security, including detection of Advanced Cyber Threats. The talk is based on the lessons learned from deploying MLAD to detect advanced cyber threats in a number of different real-world security environments. In the first part of the talk, we will discuss some of the most important use cases for MLAD that are useful for Advanced Cyber Threat detection. We will also briefly go over the methodology that we have been leveraging to analyze and model advanced threats for MLAD purposes as well as the different types of relevant anomalies.
In the second part of the talk, we will present a live demo of how MLAD works when applied to various types of real-world datasets to detect different classes of cyber threats and malicious threat actor activities associated with different stages of the Lockheed Martin Cyber Kill Chain (Hutchins et al, 2010). The types of cyber threats/malicious threat actor activities demonstrated we are going to discuss include malicious network threats/process behavior (based on a security team @services company deployment), data infiltration/exfiltration of credit card information (based on a malicious implant associated with one of the high-profile breaches), covert tunneling and command and control (C2) used by attackers to evade detection, lateral movement/use of compromised credentials etc.