Fall 2015 SIP Seminars
Prof. Dan Yang
Title: Bilinear Regression with Matrix Covariates and Applications in Neuroimaging Data Analysis
Abstract: Traditional functional linear regression usually takes a one dimensional functional predictor as input and estimates the continuous coefficient function. Modern applications often generate two dimensional covariates, which when observed at grid points are matrices. To avoid inefficiency of the classical method involving estimation of a two dimensional coefficient function, we propose a bilinear regression model and obtain estimates via a smoothness regularization method. The proposed estimator exhibits minimax optimal property for prediction under the framework of Reproducing Kernel Hilbert Space. The merits of the method are further demonstrated by numerical experiments and an application on real imaging data.
Biography: Dan Yang has a PhD in Statistics from University of Pennsylvania, and an undergraduate degree in Statistics from Peking University. She is currently an assistant professor in the Department of Statistics and Biostatistics at Rutgers University. Her research interests include high dimensional data analysis, functional data analysis, multivariate data analysis, nonparametric statistics, analysis of observational studies, and imaging data.
Prof. Venu Veeravalli
Title: Quickest Detection and Isolation of Line Outages in Power Systems
Abstract: The problem of detecting abrupt changes in stochastic systems and time series, often referred to as the quickest change detection (QCD) problem, arises in various branches of science and engineering. It is assumed that the observations undergo achange in distribution in response to a change or disruption in the environment or, more generally, to changes in certain patterns. The observations are obtained sequentially and, as long as their behavior is consistent with the normal state, the process is allowed to continue. If the state changes, then it is of interest to detect the change as soon as possible, subject to false alarm constraints, and take any necessary action in response to the change. In the first part of this talk, an up-to-date overview of the results on the QCD problem will be provided, including some recent results on data-efficient QCD. A number of applications of QCD will be discussed. In the second part of the talk, the focus will be on the problem of detecting and isolating line outages in power systems using phasor measurement unit (PMU) measurements taken at the buses. It is shown that QCD based algorithms are tailor-made for this problem and significantly outperform existing methods.
Biography: Dr. Veeravalli, received the Ph.D. degree in Electrical Engineering from the University of Illinois at Urbana-Champaign in 1992, the M.S. degree from Carnegie-Mellon University in 1987, and the B.Techdegree from Indian Institute of Technology, Bombay (Silver Medal Honors) in 1985. He is currently a Professor in the department of Electrical and Computer Engineering (ECE), the Coordinated Science Laboratory (CSL) and the Information Trust Institute (ITI) at the University of Illinois at Urbana-Champaign. He was on the faculty of the School of ECE at Cornell University before he joined Illinois in 2000. He served as a program director for communications research at the U.S. National Science Foundation in Arlington, VA during 2003-2005. His research interests span the theoretical areas of detection and estimation, information theory, statistical learning, and stochastic control, with applications to wireless communication systems and networks, sensor networks, cyberphysicalsystems, big data, and genomics. He is a Fellow of the IEEE, and a recipient of the 1996 IEEE Browder J. Thompson Best Paper Award and the U.S. Presidential Early Career Award for Scientists and Engineers (PECASE). He served as a distinguished lecturer for the IEEE Signal Processing Society during 2010-2011.
Title: Distributed Proportional Stochastic Coordinate Descent With Social Sampling
Abstract: We consider stochastic message passing algorithms that limit the communication required for decentralized and distributed convex optimization and provide convergence guarantees on the objective value. We first propose a centralized method that modifies the coordinate-sampling distribution for stochastic coordinate descent, which we call proportional stochastic coordinate descent. This method treats the gradient of the function as a probability distribution to sample the coordinates, and may be useful in so-called lock-free decentralized optimization schemes. For general distributed optimization in which agents jointly minimize the sum of local objectives we propose treating the iterates as gradients and propose a stochastic coordinate-wise primal averaging algorithm for optimization.
Biography: Mohsen Ghassemi has an undergraduate degree in electrical engineering from University of Tehran, Iran. He joined Rutgers in fall 2013 as a PhD student. He is currently working under supervision of Prof. Anand Sarwate on stochastic and distributed optimization algorithms for large scale machine learning problems.
Title: Symmetric Matrix Perturbation For Differentially-Private Principal Component Analysis
Abstract: Differential privacy is a strong, cryptographically-motivated definition of privacy that has recently received a significant amount of research attention for its robustness to known attacks. The principal component analysis (PCA) algorithm is frequently used in signal processing, machine learning, statistics pipelines. In this paper, we propose a new algorithm for differentially-private computation of PCA and compare the performance empirically with some recent state-of-the-art algorithms on different data sets. We intend to investigate the performance of these algorithms with varying privacy parameters and database parameters. We show that our proposed algorithm, despite guaranteeing stricter privacy, provides very good utility for different data sets.
Biography: Hafiz Imtiaz earned his B.Sc. and M.Sc. degrees in Electrical and Electronic Engineering from Bangladesh University of Engineering and Technology, Dhaka, Bangladesh. He joined Rutgers in Fall 2014 as a Ph.D. student under the supervision of Prof. Anand Sarwate. He is currently working on differential privacy and related applications.