Bio: Professor S.Y. Kung received his Ph.D. Degree in Electrical Engineering from Stanford University in 1977. In 1974, he was an Associate Engineer of Amdahl Corporation, Sunnyvale, CA. From 1977 to 1987, he was a Professor of Electrical Engineering-Systems of the University of Southern California, L.A. Since 1987, he has been a Professor of Electrical Engineering at the Princeton University. In addition, he held a Visiting Professorship at the Stanford University (1984); and a Visiting Professorship at the Delft University of Technology (1984); a Toshiba Chair Professorship at the Waseda University, Japan (1984); an Honorary Professorship at the Central China University of Science and Technology (1994); and a Distinguished Chair Professorship at the Hong Kong Polytechnic University since 2001. His research interests include VLSI array processors, system modelling and identification, neural networks, wireless communication, sensor array processing, multimedia signal processing, bioinformatic data mining and biometric authentication. Professor Kung has authored more than 400 technical publications and numereous textbooks, Professor Kung has co-authored more than 400 technical publications and numerous textbooks including "VLSI and Modern Signal Processing," with Russian translation, Prentice-Hall (1985), "VLSI Array Processors", with Russian and Chinese translations, Prentice-Hall (1988); "Digital Neural Networks", Prentice-Hall (1993) ; "Principal Component Neural Networks", John-Wiley (1996); and "Biometric Authentication: A Machine Learning Approach", Prentice-Hall (2004). Professor Kung is a Fellow of IEEE since 1988. He served as a Member of the Board of Governors of the IEEE Signal Processing Society (1989-1991). He was a founding member of several Technical Committees (TC) of the IEEE Signal Processing Society , including VLSI Signal Processing TC (1984), Neural Networks for Signal Processing TC (1991) and Multimedia Signal Processing TC (1998), and was appointed as the first Associate Editor in VLSI Area (1984) and later the first Associate Editor in Neural Network (1991) for the IEEE Transactions on Signal Processing. He presently serves on Technical Committees on Multimedia Signal Processing. Since 1990, he has been the Editor-In-Chief of the Journal of VLSI Signal Processing Systems. Professor Kung was a recipient of IEEE Signal Processing Society's Technical Achievement Award for his contributions on "parallel processing and neural network algorithms for signal processing" (1992); a Distinguished Lecturer of IEEE Signal Processing Society (1994) ; a recipient of IEEE Signal Processing Society's Best Paper Award for his publication on principal component neural networks (1996); and a recipient of the IEEE Third Millennium Medal (2000).

Title: Compressive Privacy Based on Joint Optimization of Differential Utility/Cost

Abstract:
With the rapidly growing internet commerce, personal data are being collected, stored, and circulated around the internet - often without the data owner's knowledge. As such, protection of privacy of internet data has become a vital research field. Conventionally, the task of data protection is left entirely to the cloud server, rendering data privacy extremely vulnerable to hacker attack or unauthorized leakage. To rectify this problem, a key is to let data owners (not cloud servers) control the data privacy.

This talk explores the rich synergy among signal processing, information theory, estimation theory, and machine learning and, thereafter, presents a novel privacy preserving methodology, named compressive privacy (CS). The objective is to build user-collaborative machine learning systems to facilitate the intended function while protecting the privacy of owner's sensitive information. It involves the joint optimization over three design spaces: (i) Feature Space (ii) Utility Subspace; and (iii) Cost Subspace (i.e. Privacy Subspace). Mathematically, the optimal query can be derived from the joint optimization formulation where the query should be chosen to simultaneously maximize the utility and minimize the cost. In order to derive a closed form analysis/solution, we recast the information theoretical criterion (such as the log-likelihood or mutual information) in terms of (differential) error covariance matrix used in the estimation theory. More exactly, the optimal query search (or feature selection) becomes a problem of maximizing a Differential Utility/Cost (DUC) ratio, a criterion function commonly adopted by economists. More exactly, DUC is defined as the ratio between Differential Utility and Differential Cost. The DUC formulation can be extended to Machine Learning applications, where the Differential Utility and Cost are characterized by the given supervised training dataset. Furthermore, the DUC optimization can be reformulated in the kernel learning models, where nonlinear kernels afford a much expanded search space to enhance the optimized DUC ratio. Simulation results based on facial image classification (utility) and reconstruction (privacy) will be demonstrated.

Acknowledgement: This material is based upon work supported in part by the Brandeis Program of the Defense Advanced Research Project Agency (DARPA) and Space and Naval Warfare System Center Pacific (SSC Pacific) under Contract No. 66001-15-C-4068.

Bio: Xiaodong Zhang is the Robert M. Critchfield Professor in Engineering and Chair of the Computer Science and Engineering Department at the Ohio State University. His research interests focus on data management in computer and distributed systems. He has made strong efforts to transfer his academic research into advanced technology to update the design and implementation of major general-purpose computing systems. He received his Ph.D. in Computer Science from University of Colorado at Boulder, where he received Distinguished Engineering Alumni Award in 2011. He is a Fellow of the ACM, and a Fellow of the IEEE.

Title: Fast Data Access Service in Cloud Systems.

Abstract: Cloud computing provides us with basic and indispensable service in daily life. Taking the iPhone as an example, each daily used app on the phone, such as calendar, weather, text message, Wechat, and many others, connects to a data center. Today’s concept of computing is data processing in practice. This presentation will focus on fundamental issues of data processing in the context of big data.

A major goal of algorithms analysis and implementation in data processing is to read and write data records from memory or storage in high speed at a low cost for a given data storage format. As the data volume generated in the society continues to grow in an increasingly rapid way, we have reevaluated several commonly used data accessing methods including LSM-tree for sequentially archived data, relational data model, and storing/retrieving methods for key-value stored data. In this talk, I will show their limits and inabilities to handle big volume of data in a scalable way. I will also present three new research results: (1) re-enabling buffer caching capability for LSM-tree to achieve high performance of both reads and writes to process sequentially archived data, (2) balancing both network bandwidths and storage transfers for relational tables in large clusters, and (3) maximizing throughput of in-memory key-value stores by GPUs. All the related algorithms and software implementations are open sourced, some of which have been adopted in production systems.