Seminars: Spring 2017
Friend or Foe? Your Wearable Devices Reveal Your Personal PIN
Speaker:Yingying (Jennifer) Chen, Stevens Institute of Technology
Time: 11:00 am - 12:00 pm Jan 26, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
The proliferation of wearable devices, e.g., smart watches and activity trackers, with embedded sensors has already shown its great potential on monitoring and inferring human daily activities. Our work reveals a serious security breach of wearable devices in the context of divulging secret information (i.e., key entries) while people accessing key-based security systems. Existing methods of obtaining such secret information relies on installations of dedicated hardware (e.g., video camera or fake keypad), or training with labeled data from body sensors, which restrict use cases in practical adversary scenarios. In this work, we show that a wearable device can be exploited to discriminate mm-level distances and directions of the user’s fine-grained hand movements, which enable attackers to reproduce the trajectories of the user’s hand and further to recover the secret key entries. In particular, our system confirms the possibility of using embedded sensors in wearable devices, i.e., accelerometers, gyroscopes, and magnetometers, to derive the moving distance of the user’s hand between consecutive key entries regardless of the pose of the hand. Our Backward PIN-Sequence Inference algorithm exploits the inherent physical constraints between key entries to infer the complete user key entry sequence. Extensive experiments are conducted with over 5000 key entry traces collected from 20 participants for key-based security systems (i.e. ATM keypads and regular keyboards) through testing on different kinds of wearables. Results demonstrate that such a technique can achieve 80% accuracy with only one try and more than 90% accuracy with three tries, which to our knowledge, is the first technique that reveals personal PINs leveraging wearable devices without the need for labeled training data and contextual information.
About the Speaker: Yingying (Jennifer) Chen is a Professor in the Department of Electrical and Computer Engineering at Stevens Institute of Technology. Her research interests include cyber security and privacy, Internet of Things, smart healthcare and mobile computing and sensing. She has published over 100 journals and referred conference papers in these areas. She received her Ph.D. degree in Computer Science from Rutgers University. Prior to joining Stevens, she was with Alcatel-Lucent at Murray Hill, New Jersy. She is the recipient of the NSF CAREER Award and Google Faculty Research Award. She also received NJ Inventors Hall of Fame Innovator Award. She is the recipient of the Best Paper Awards from ACM AsiaCCCS 2016, IEEE CNS 2014 and ACM MobiCom 2011. She also received the IEEE Outstanding Contribution Award from IEEE New Jersey Coast Section each year 2005 - 2009. Her research has been reported in numerous media outlets including MIT Technology Review, Fox News Channel, Wall Street Journal, and National Public Radio. She serves on the editorial boards of IEEE Transactions on Mobile Computing (IEEE TMC), IEEE Transactions on Wireless Communications (IEEE TWireless), and IEEE Network Magazine.
Cinematic Virtual Reality: Immersive Video for Head-Mounted Displays
Speaker:Bernd Girod, Stanford University
Time: 11:00 am - 12:00 pm Feb 2, 2017
Location: 2 MetroTech Center, 10th floor, MAGNET Lecture Hall, Brooklyn, NY
The 2014 acquisition of a fledging head-mounted display company for $2B has reignited the excitement in virtual reality, not just for rendered 3d computer graphics but primarily for immersive “cinematic” video captured by means of special camera rigs. In this talk, we will review the principles of representing immersive video for head-mounted displays and the challenges that arise for efficient coding and delivery. How many pixels are needed to cover the full field of view with retina resolution? How can we provide binocular stereo in all directions? How can we accommodate head motion? How can we provide defocus cues to overcome the conflict between stereo vergence and lens accommodation? And what are best video representations for compact storage and transmission that support all of the above? We show that significant technology challenges remain for cinematic virtual reality to live up to its high expectations, some of them familiar and some new.
About the Speaker: Bernd Girod is the Robert L. and Audrey S. Hancock Professor of Electrical Engineering at Stanford University, California. He received the Engineering Doctorate degree from University of Hannover, Germany, and the M.S. degree from the Georgia Institute of Technology. Until 1999, he was a Professor with the Electrical Engineering Department, University of Erlangen– Nuremberg. He has authored over 600 conference and journal papers and six books, receiving the EURASIP Signal Processing Best Paper Award in 2002, the IEEE Multimedia Communication Best Paper Award in 2007, the EURASIP Image Communication Best Paper Award in 2008, the EURASIP Signal Processing Most Cited Paper Award in 2008, the EURASIP Technical Achievement Award in 2004, and the Technical Achievement Award of the IEEE Signal Processing Society in 2011. His research interests are in the area of image, video, and multimedia systems. As an entrepreneur, he was involved in numerous startup ventures, among them Polycom, Vivo Software, 8x8, and RealNetworks. He is a Fellow of the IEEE, a EURASIP Fellow, a member of the the National Academy of Engineering, and a member of the German National Academy of Sciences (Leopoldina).
Social-Sensed Multimedia Computing
Speaker:Wenwu Zhu, Tsinghua University, China
Time: 11:00 am - 12:00 pm Feb 13, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Most multimedia applications try to deliver multimedia contents to end users according to their information needs. Thus, multimedia computing actually plays the role of a bridge between multimedia data and user needs. In the past years, however, the multimedia research community mostly focus on multimedia content analysis and understanding; while the user needs over multimedia data are rarely researched, which results in the well-known Intention Gap problem. With the emergence of online social networks, billions of users proactively interact with huge volumes of multimedia data, thereby user behaviors, user relations, user influences etc. can be sensed from the social networks. In this talk, we first will present the social-sensed multimedia computing framework, which tries to inject social factors into traditional multimedia computing and bridges the multimedia content with users. Then, several exemplary applications of this framework, such as social-sensed image search, social-sensed multimedia recommendation, social-sensed video delivery, etc. will be presented. Finally, the future research directions will be discussed.
About the Speaker: Wenwu Zhu is currently a Professor and Deputy Head of Computer Science Department of Tsinghua University. Prior to his current position, he was a Senior Researcher and Research Manager at Microsoft Research Asia. He was the Chief Scientist and Director at Intel Research China from 2004 to 2008. He worked at Bell Labs New Jersey as Member of Technical Staff during 1996-1999. He received his Ph.D. degree from New York University Polytechnic School of Engineering in 1996. He is an IEEE Fellow, AAAS Fellow, SPIE Fellow and ACM Distinguished Scientist. His research interests are in the areas of multimedia computing, communications and networking, as well as big data computing. He served in editorial boards for various journals such as Proceedings of the IEEE, JSAC, TMM, TMC, TCSVT, TBD, etc.. He served in the Steering Committee for IEEE Transactions on Multimedia (2016) and IEEE Transactions on Multimedia Computing (2007-2010), respectively. He will serve as the EIC for the IEEE Transactions on Multimedia starting on Jan.1, 2017. He received five Best Paper Awards including ACM Multimedia 2012 and IEEE Transactions on Circuits and Systems for Video Technology in 2001.
Two-dimensional epitaxial superconductor-semiconductor heterostructures: A platform for topological superconducting networks
Speaker:Javad Shabani, City College of New York
Time: 11:00 am - 12:00 pm Feb 15, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Progress in the emergent field of topological superconductivity relies on synthesis of new material combining superconductivity, low density, and spin-orbit coupling. Recently, epitaxial growth of Al on InAs nanowires was shown to yield a high quality superconductor-semiconductor (S-Sm) system with uniformly transparent interfaces and a hard induced gap, indicted by strongly suppressed subgap tunneling conductance. Here we report the realization of a two-dimensional (2D) InAs/InGaAs heterostructure with epitaxial Al, yielding a planar S-Sm system with structural and transport characteristics as good as the epitaxial wires. The realization of 2D epitaxial S-Sm systems represent a significant advance over wires, allowing extended networks via top-down processing. Among numerous potential applications, this new material system can serve as a platform for complex networks of topological superconductors with gate-controlled junctions. We demonstrate gateable Josephson junctions and a highly transparent 2D S-Sm interface based on the product of excess current and normal state resistance.
About the Speaker: Shabani received his PhD in Electrical Engineering from Princeton University 2011. After two years of research on semiconductor-based qubits at Harvard University, he joined UC Santa Barbara in 2012. There, he worked closely with Microsoft research on hybrid semiconductors/superconductors heterostructures to study topological superconductivity. He is an expert in quantum materials and devices for computation technologies. He is currently an assistant professor at City college of New York and recipient of a number of awards including young investigator awards from Army and Airforce research offices.
Data-Driven Techniques Against Cybercrime Operations
Speaker:Shuang Hao, University of California, Santa Barbara
Time: 11:00 am - 12:00 pm Feb 21, 2017
Location: LC400, 5 MetroTech Center, Brooklyn, NY
Modern day cybercrime has become a heavy burden on the Internet ecosystem and caused increasing loss to end users, ranging from scam websites to message abuse to compromises of bank accounts. The explosive growth of online services and user-generated data facilitate the engagement in illegal activities. Mitigating emerging threats requires both gathering empirical measurement to understand the operation logistics of attacks and developing practical defense mechanisms.
In this talk, I will describe my work on using data-driven approaches to detect and disrupt cybercrime operations. First, I will present PREDATOR, a proactive DNS reputation system that we developed to detect malicious domains early, at time-of-registration, rather than later at time-of-use. I will then describe two of our studies on mitigating misuse of cloud services for illicit web hosting and disrupting a monetization scheme that cybercriminals rely on to turn stolen credit cards into cash. Finally, I will briefly discuss other threats and highlight the potential of leveraging data aggregated from various vantage points to address a wide range of security and privacy problems.
About the Speaker: Shuang Hao is a Postdoctoral Researcher in Computer Science at the University of California, Santa Barbara working with Prof. Christopher Kruegel and Prof. Giovanni Vigna. His research interests include anomaly detection, underground economics, DNS analysis, and web and mobile security. His research uses data-driven techniques to measure and detect attacks and vulnerabilities in networking and distributed systems. His work has been covered by various major media including MIT Technology Review, Wall Street Journal, WIRED, Slashdot, and KrebsOnSecurity. He was awarded the Yahoo! Key Scientific Challenges Program Award. He obtained his Ph.D. in the School of Computer Science at the Georgia Institute of Technology under the supervision of Prof. Nick Feamster.
Privacy in the Internet (Without Giving up Everything Else)
Speaker:David Naylor, Carnegie Mellon University
Time: 11:00 am - 12:00 pm Feb 22, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Using the Internet inherently entails privacy risks. Each packet, potentially carrying information that users would rather keep private, is exposed to a network infrastructure operated by a number of third parties the user may not trust and likely cannot even identify. In some cases, the user may not even trust the recipient.
Techniques exist to protect user privacy, but they typically do so at the expense of other desirable properties. For example, anonymity services like Tor hide a packet's true sender, but weaken accountability by making it difficult for network administrators or law enforcement to track down malicious senders. Similarly, encryption hides application data from third parties, but prevents the use of middleboxes---devices that process packets in the network to improve performance (like caches) or security (like intrusion detection systems).
In this talk, I’ll present techniques for managing these "Privacy vs. X" conflicts, including a new network architecture that re-thinks basic networking building blocks like packet source addresses and new secure communication protocols that explicitly balance data privacy with the benefits of middleboxes.
About the Speaker: David is a Ph.D. student at Carnegie Mellon University, where he is advised by Peter Steenkiste. His primary research interests are computer networking, security, and privacy, but he is also interested in Web measurement and performance (http://isthewebhttp2yet.com and https://eyeorg.net). David received his B.S. from the University of Iowa in 2011, where he created the DDR inspired "Scrub Scrub Revolution," a handwashing training game for healthcare professionals. He is an NDSEG fellow and received an ACM SIGCOMM best paper award.
Energy-efficient, high-fidelity, and broadband RF/millimeter-wave signal generation in silicon
Speaker:Song Hu, Georgia Institute of Technology
Time: 11:00 am - 12:00 pm Feb 23, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Wireless electronic devices are everywhere in our daily lives with applications ranging from data communication to energy transfer and biomedical sensing. Power amplifier (PA) interfaces the antenna and governs the energy efficiency of a wireless transceiver. Its linearity is also of paramount importance to ensure the signal fidelity. Moreover, its broadband operation is highly desired for high-speed communication and high-resolution sensing. However, integrating a PA in silicon entails challenges due to the PA’s nature of large-signal, non-linear, and highly dynamic operation. Meanwhile, the silicon platform has become one of the finest manufacturing capabilities that human beings have ever created. Device scaling in the front-end- of-line processes has empowered the silicon platform unprecedented mixed-signal (digital and analog) processing capabilities and the back-end- of-line processes offer the super-flexible and ultra-fine lithography of interconnections.
My research exploits the energy-efficient mixed-signal computation and novel on-chip electromagnetic structures to enable intelligent RF/mm-wave large-signal operation in silicon. This talk will present silicon implementations based on these design methodologies. I will first introduce a digital Doherty PA in silicon. Flexible and precise digital control optimizes in-field Doherty efficiency enhancement and enables robustness against antenna mismatch. Then, I will introduce hybrid PA efficiency enhancement techniques enabled by digital-intensive architectures in silicon. Mixed-signal linearization eliminates the trade-off between efficiency and linearity in these architectures. Lastly, I will present a multiband millimeter-wave PA in silicon for fifth-generation (5G) communication. Mixed-signal reconfiguration and a novel on-chip power combiner enable its broadband operation. These examples extend the boundary of RF/millimeter-wave large-signal circuits in silicon from various perspectives and illustrate the introduced paradigm-shift design methodologies that open doors to future silicon solutions.
About the Speaker: Song Hu is a Ph.D. candidate in the School of Electrical and Computer Engineering at the Georgia Institute of Technology. He received the B.Eng. degree in information engineering (with honors) from Southeast University, Nanjing, China, in 2009 and the M.Sc. degree in microelectronics from Fudan University, Shanghai, China, in 2012. His current research interests are high-speed circuits and systems for communication, sensing, and imaging.
Mr. Hu was the winner of the Best Student Paper Award (First Place) of the 2014 IEEE Radio Frequency Integrated Circuits (RFIC) Symposium and the Best Student Paper Award of the 2011 IEEE Radio Frequency Integration Technology (RFIT) Symposium. He was also the co-recipient of the 2016 Microwave Magazine Best Paper Award and the Best Student Paper Award (Second Place) of the 2015 IEEE Custom Integrated Circuits Conference (CICC).
Mr. Hu received the 2016 Georgia Tech ECE Roger P. Webb Graduate Research Assistant Excellence Award, the 2015-2016 IEEE Solid-State Circuits Society (SSCS) Predoctoral Achievement Award, the 2015 IEEE Microwave Theory and Techniques Society (MTT-S) Graduate Fellowship, and the 2014 Analog Devices Inc. Outstanding Student Designer Award. He was also the recipient of the Southeast University President’s Scholarship in 2006, the National Scholarship in 2007, the first prize in the National Undergraduate Electronic Design Contest in 2008, the Microsoft Young Fellowship Award from Microsoft Research Asia in 2008, the Marvell Fellowship in 2011, the Shanghai Outstanding Graduate Student Award in 2012, and the Shanghai Excellent Master Thesis Award in 2013.
Engineering W-band phased array transmit/receive (T/R) module for maximum power efficiency toward sustainable large phased array systems.
Speaker:Kwang Jin Koh, Virginia Tech
Time: 11:00 am - 12:00 pm Feb 24, 2017
Location: LC400, 5 MetroTech Center, Brooklyn, NY
Historically, phased arrays have been widely used for defense radar applications at a high cost per element ($200-2000). As an effort to reduce the volume and cost of the antenna array systems, there have been substantial research efforts to integrate the phased arrays in a cost-effective silicon process during the past decade. This results in a product reality in commercial 60 GHz wireless applications with a low-to- medium level array size, typically 16 to 64 elements integration level. For a high resolution radar sensor, especially for defense applications at W-band (94 GHz), the required array size is large, ranging several hundreds to several tens of thousands. In such large arrays, it is critically important for the phased array electronics to maintain a high power efficiency to contain thermal management issues with an affordable manner. While the realization of a silicon-based large array at W-band is still in its infancy, recent exploratory designs under a DARPA program exhibit extremely low power efficiency, typically less than 1-2% and raising a significant concern on the thermal management task.
During the past several years, under the support of the same defense program Dr. Koh has been conducting a focused research for the efficiency breakthrough in the silicon phased arrays. This talk is a summary of those engineering effort and focuses on the high-efficiency transmit/receive (T/R) module designs in 0.13-µm SiGe BiCMOS technology for the W-band phased array applications in the context of large antenna arrays. The SiGe process is still preferable choice over nano-scale CMOS process for better system reliability in the defense electronics due to a high breakdown voltage. The phase shifter and power amplifier are two major power hungry building blocks and it is important to design them with a high power efficiency to improve overall T/R module energy efficiency. After a brief introduction on the past phased array integration efforts toward a large scale array, this talk will discuss details on the high-efficiency phase shifter and power amplifier architectures and implementation efforts. The power-efficient T/R module designs in this talk can be scaled properly to other millimeter bands applications. Thus, this talk also includes an introduction of impending further work on the construction of a scalable large array based on the high-efficiency phased array T/R component designs for a various promising millimeter-wave radio/radar applications.
About the Speaker: Dr. Koh received his Ph.D degree in electrical and computer engineering (ECE) from the University of California, San Diego in 2008. After several years stint in the semiconductor industry, he joined in the Virginia Tech ECE Department as an assistant professor and started teaching from the 2012 Spring semester. Dr. Koh and his students received the first place Best Student Paper award in the 2016 IEEE/MTT-S International Microwave Symposium (IMS) and the second place Best Student Paper award in the 2015 IEEE/MTT-S IMS. Dr. Koh was also nominated for the R.W.P. King best paper award from the IEEE Antenna and Propagation Society in 2015. Dr. Koh was the recipient of a best paper award of the IEEE Solid-State Circuits and Electron Device Societies, Seoul Chapter in 2002. He and his student also received the 2015-2017 NASA Virginia Space Grant Consortium STEM research fellowship. Dr. Koh is the recipient of the 2014 Outstanding Assistant Professor Award and the 2012 Junior Faculty Research Award from the College of Engineering, Virginia Tech, and the 2010 Team of the Year Award from Teledyne Technology Inc. (formerly, Teledyne Scientific Corp.). Dr. Koh has been serving as a technical program committee member of the IEEE Bipolar/BiCMOS Circuits and Technology Meeting (BCTM) and IEEE Custom Integrated Circuis Conference (CICC).
On the Security and Scalability of Proof of Work Blockchains
Speaker:Arthur Gervais, Institute of Information Security at ETH Zürich, Germany
Time: 11:00 am - 12:00 pm Feb 27, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
The security properties of blockchain technology allow for the shifting of trust assumptions, e.g., to remove trusted third parties; they, however, create new challenges for security and scalability, which have not yet been fully understood and that we investigate in this talk. The blockchain’s security, for example, affects the ability of participants to exchange monetary value or to participate in the network communication and the consensus process. Our first contribution provides a quantitative framework to objectively compare the security and performance characteristics of Proof of Work-based blockchains under adversaries with optimal strategies. Our work allows us to increase Bitcoin’s transaction throughput by a factor of ten, given only one parameter change and without deteriorating the security of the underlying blockchain. In our second contribution, we highlight previously unconsidered impacts of the PoW blockchain’s scalability on its security and propose design modifications that are now implemented in the primary Bitcoin client. Because blockchain technology is still in its infancy, we conclude the talk with an outline of future work towards an open, scalable, privacy-preserving and decentralized blockchain.
About the Speaker: Arthur Gervais's research interests revolve around the security and privacy of blockchain technology, and he also worked on web privacy. He just defended in December 2016 his Ph.D. in the Institute of Information Security at ETH Zürich. During his Ph.D., he performed a 3-months internship at Intel Labs, Oregon, working on blockchain technology. He obtained his Master degrees from KTH Stockholm (Sweden) and Aalto University (Finland) in 2012. Furthermore, he holds a diplôme d'ingénieur from INSA de Lyon (France) from 2012. His Master's thesis was on the security of industrial control systems (SCADA).
Diversity and Resilience in Robot Networks
Speaker:Amanda Prorok, University of Pennsylvania
Time: 11:00 am - 12:00 pm Mar 1, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Recent years have seen falling costs of communication and storage technologies and advances in fabrication methods. Sensors, actuators, and processors are being integrated into globally accessible information networks. These trends are promoting a profusion of networked robotic platforms with distinct features and unique capabilities. As we aspire to harness this diverse array of robots to solve increasingly complex problems, heterogeneity and diversity become design features. However, we still lack a fundamental understanding of how to compose and control large-scale systems of heterogeneous robots. Moreover, as we program diverse robots to exploit their technical complementarities, we create interdependencies and critical links. Such collaborative algorithms give rise to new sources of internal and external threats that lead to unintended failure modes. As a consequence, we need new mechanisms that ensure resilience.
I begin my talk by formalizing diversity in the context of dynamic task allocation for large-scale heterogeneous multi-robot systems. In light of this setting, I show how optimal control policies are impacted by the heterogeneity of the robot team. In the second part of the talk, my focus shifts to the question of how to provide resilience to internal failures through precautionary collaboration mechanisms. By building on foundational concepts of network science and security, I show how we can achieve resilience, allowing robot teams to function in the presence of defective and/or malicious robots. Finally, I consider the importance of providing system-wide protection against external threats, and introduce some new ideas that touch upon privacy.
About the Speaker: Amanda Prorok is a Postdoctoral Researcher in the General Robotics, Automation, Sensing and Perception (GRASP) Laboratory at the University of Pennsylvania, where she works with Prof. Vijay Kumar on heterogeneous networked robotic systems. She completed her PhD at EPFL, Switzerland, where she addressed the topic of localization with ultra-wideband sensing for robotic networks. Her dissertation was awarded the Asea Brown Boveri (ABB) award for the best thesis at EPFL in the fields of Computer Sciences, Automatics and Telecommunications. She was selected as an MIT Rising Star in 2015, and won a Best Paper Award at the 9th International Conference on Bio-inspired Information and Communications Technologies, 2015.
Website:
http://www.prorok.me
Rapid diagnostics: When Electrochemical Nanobiosensors probe Pathogens
Speaker:Aida Ebrahimi, Purdue University
Time: 11:00 am - 12:00 pm Mar 2, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Considering their scalability and compatibility with integrated circuit technology and microfluidics, electrochemical nanobiosensors are among the most promising paradigms to address some of the 21st century’s grand challenges, especially for low-cost personalized health monitoring and rapid, point-of- care diagnostics. Many exciting functionalities can be created by integration of electronics with nanomaterials, microfluidics, system engineering, etc. In this talk, I will show you how nature inspired us to develop a low-cost, array-formatted electrochemical sensor for rapid detection of biomolecules. The method utilizes droplet evaporation to concentrate the analyte and reduces the detection time from days to mere minutes. I will highlight the power of time-dependent electrical analysis to sort alive and dead bacterial cells in only a few minutes, as well as on-chip probing of some of their fundamental behaviors in response to important environmental input. These findings demonstrate the potential of electrochemical nanobiosensors for development of next generation of low-cost and accurate biological interrogation systems to address today’s challenges, especially the antibiotic resistance issue, and offer exciting opportunities for innovating health monitoring and therapeutic methods.
About the Speaker: Aida Ebrahimi is a PhD candidate in the Electrical and Computer Engineering Department at Purdue University. Her research interests include electronic sensors and systems for point-of-care diagnostics, lab-on- chip, flexible bioelectronics, and cellular biophysics. Her research intersects various disciplines from device engineering, electrochemistry, and micro and nanofabrication to microbiology and cellular biophysics. She has been selected as a Rising Star in EECS which recognizes top 60 EECS graduate and postdoctoral women from USA and Europe (MIT, 2015). She is also one of the two recipients of the prestigious Bilsland Dissertation Fellowship Award in the department of ECE at Purdue University. This award recognizes a PhD candidate’s superior academic abilities and scholarly achievements. Her research has been published in prestigious journals and conference proceedings, including Proceedings of National Academy of Sciences, Lab on a Chip, Applied Physics Letters, and others.
Modeling, Sensing and Control of Unstable Physical Human-Machine Interactions: A Rider-Bikebot Example
Speaker:Jingang Yi, Rutgers University
Time: 11:00 am - 12:00 pm Mar 3, 2017
Location: LC400, 5 MetroTech Center, Brooklyn, NY
Human with trained motor skills can fluidly and flexibly interact with machines while smart machines can also provide motor assistance and enhancement to facilitate human’s motor skills learning. However, we currently lack theories and design tools to effectively model and tune human motor control and its interactions with machines. In this talk, I will discuss recent developments modeling, sensing and control of human motor skills through unstable physical human-machine interactions (upHMI). Rider-bikebot (i.e., bicycle-like robot) interactions is used as an upHMI paradigm to examine a sensorimotor theory for modeling of human motor control relevant to balancing motor activities. I will first present a novel control-theoretic physical/learning modeling framework of extracting and characterizing human control strategies in a lower-dimensional space. Then, I will discuss the development of the in-situ sensing design to estimate the poses of the rider and the bicycle in natural environment with wearable and onboard sensors. Finally, I will briefly present balancing control design and stability analysis for the rider-bicycle interactions. If time permits, I will also briefly present various other research projects at the Robotics, Automation and Mechatronics (RAM) Lab at Rutgers.
About the Speaker: Professor Jingang Yi received the B.S. degree in electrical engineering from Zhejiang University (China) in 1993, the M.Eng. degree in precision instruments from Tsinghua University (China) in 1996, and the M.A. degree in mathematics and the Ph.D. degree in mechanical engineering from the University of California, Berkeley, in 2001 and 2002, respectively. He is currently an Associate Professor in mechanical engineering and a Graduate Faculty member in electrical and computer engineering at Rutgers University. His research interests include autonomous robotic and vehicle systems, dynamic systems and control, mechatronics, automation science and engineering, with applications to biomedical, transportation and civil infrastructure systems. Prof. Yi is a Fellow of ASME and a senior member of IEEE. He has published more than 150 papers in international journal and conferences and received several awards, including the 2014 ASCE Charles Pankow Award for Innovation, the 2013 Rutgers Board of Trustees Research Fellowship for Scholarly Excellence, and the 2010 NSF CAREER Award. He also held visiting positions as a Global Guest Professor at Keio University, Japan in 2016 and a Distinguished Visiting Scholars, University of Technology, Sydney (UTS), Australia in 2015. He has coauthored several best papers, including the 2015 Best New Application Paper in IEEE Transactions on Automation Science and Engineering and the best papers at the IEEE/ASME AIM, ASME DSCC, and IEEE ICRA. He currently serves as an Associate Editor for IEEE/ASME Transactions on Mechatronics, IEEE Transactions on Automation Science and Engineering, IFAC journal Control Engineering Practice, IFAC journal Mechatronics, the ASME Journal of Dynamic Systems, Measurement and Control, International Journal of Intelligent Robotics and Applications, and the IEEE Robotics and Automation Society (RAS) Conference Editorial Board (since 2008). His research has been supported by the NSF, NIST, FHWA, NASA, and other agencies in US and the NSFC in China.
Safe by design: Soft robots for human-machine interaction
Speaker:Michael Wehner, Harvard Microrobotics Lab
Time: 11:00 am - 12:00 pm Mar 7, 2017
Location: LC400, 5 MetroTech Center, Brooklyn, NY
In recent years, the nascent field of soft robotics has emerged as an exciting area of research that stands to revolutionize our interaction with machines. Soft robots possess many attributes that are difficult, if not impossible, to achieve with conventional robots composed of rigid materials. Yet, despite recent advances, soft robots still require traditional control and power systems, thus they remain either tethered to remote hardware, or they are soft-rigid hybrid systems. Recent work has explored possible soft analogs for these standard rigid components. While new challenges arise from the incorporation of these new components, new possibilities arise as well. Eliminating rigid components allows the shrinking of length scales and developing of new form factors impossible with traditional motors, batteries, and electronic controllers. We look at the use of novel soft controllers, alternate fuel source, and soft actuators and explore possible form factors and applications. These components are brought together via a design and rapid fabrication approach, which lays the foundation for a new class of completely soft, autonomous robots.
About the Speaker: Michael Wehner received his Ph.D. in Mechanical Engineering at the University of California at Berkeley, where his research focused on human machine interaction and the development of an exoskeleton system to reduce back forces during lifting. In 2011, Michael returned to academia as a post-doctoral fellow in the Harvard Microrobotics lab with Rob Wood. In this group, he explored soft alternatives to conventional rigid orthotics for disabled children, and developed the first engineered soft exosuit. Working with the Jennifer Lewis and George Whitesides labs, he also developed Octobot, the first entirely soft untethered robot, replacing rigid battery, pump, and controller with monopropellant fluid power and microfluidic soft logic. Michael Wehner has also worked extensively in industry in various engineering, consultant, and management roles in the semiconductor capital equipment, medical device, and consumer goods fields.
Iterative Algorithms for Audio Reconstruction
Speaker:İlker Bayram, Istanbul Technical University, Istanbul
Time: 11:00 am - 12:00 pm Mar 16, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Recovery of a signal from its distorted and noisy observations is known as an inverse problem. Energy (or, cost) minimization formulations constitute one of the main approaches for addressing inverse problems. Even though convexity of the cost function is a desirable feature, such a restriction may not lead to a high fidelity reconstruction when the desired signal is an audio signal. Nevertheless, convexity can be still be employed in an iterative manner (e.g., each step of an iterative scheme may require the solution of a convex minimization problem) so as to ensure the stability of the methods. In this talk, I will discuss two different applications, involving audio signals, that make use of convex optimization techniques. In the first part, I will introduce a weakly-convex penalty function that respects the harmonic structure of audio signals. Even though the introduced penalty function is not convex, its weak convexity allows for the construction of iterative algorithms that are provably convergent. The second part of the talk will be on blind reverberant source separation with a microphone array. In this problem, we are given multiple observations of the desired audio signals, distorted with different linear time-invariant systems with unknown impulse responses. I will discuss how this reverberant separation problem can be reduced to a non-reverberant one, using a sparsity promoting energy minimization formulation, and describe an algorithm based on the Douglas-Rachford algorithm for solving the minimization formulation.
About the Speaker: İlker Bayram received the B.Sc. and M.Sc. degrees in Electrical and Electronics Engineering from Middle East Technical University (METU), Ankara, Turkey, in 2002, and 2004 respectively. He recieved the Ph.D. degree in Electrical Engineering from Polytechnic Institute of New York University in 2009. In the following year, he was with the Biomedical Imaging Group at Ecole Polytechnique Federale de Lausanne (EPFL), Switzerland, as a post-doctoral researcher. In 2010, he joined Istanbul Technical University, Dept. of Electronics and Communications Engineering, as an assistant professor. He is currently an associate professor in the same department. His research interests are in time-frequency frames, sparse signal processing, algorithms for reconstruction problems, and machine learning for signal processing applications.
Smart Management, Monitoring, and Learning for Sustainable Power Grids
Speaker:Yu Zhang, University of California-Berkeley
Time: 11:00 am - 12:00 pm Mar 23, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Current electrical grids are undergoing dramatic transformations due to high-penetration renewables and flexible loads. With the growing deployment of two-way communications, we are facing grand challenges in the form of increased uncertainty and cybersecurity threats. In this talk, accounting for the stochastic nature of renewable generation, I will start with the robust and distributed energy management for grid-connected microgrids. The second part of the talk deals with the non-convex power flow and power system state estimation problems, which play an indispensable role in grid monitoring and operation. Leveraging the sparsity of the underlying power network, I will present a novel convex relaxation framework and scalable algorithms with guaranteed performance. If time allows, I will finally highlight efficient approaches for energy data analytics that include forecasting electricity prices and wind power.
About the Speaker: Yu Zhang is a postdoctoral scholar in the Department of Industrial Engineering and Operations Research at UC Berkeley. He received his Ph.D. degree in Electrical and Computer Engineering from the University of Minnesota in 2015. Before that, he got the B.Eng. and M.Sc. degrees in Electrical Engineering from Wuhan University of Technology and Shanghai Jiao Tong University in China, respectively. His research interests span the area of cyber-physical systems including smart power grids, wireless communications, and geo-distributed data centers. Dr. Yu Zhang is a recipient of the ECE Department Fellowship at the University of Minnesota, as well as Huawei Scholarship and Infineon Scholarship at Shanghai Jiao Tong University. He has been awarded student travel grants from the IEEE Signal Processing Society and the SIAM.
Data-Driven Dynamic Robust Resource Allocation for Efficient Transportation
Speaker:Fei Miao, University of Pennsylvania
Time: 11:00 am - 12:00 pm Mar 28, 2017
Location: LC400, 5 MetroTech Center, Brooklyn, NY
Ubiquitous sensing in smart cities enables large-scale multi-source data collected in real-time, poses several challenges and requires a paradigm-shift to capture the ever growing complexity and dynamics of systems. Data-driven cyber-physical systems(CPSs) integrating machine learning, statistical methods, optimization, and control are highly desirable for this paradigm- shift, since existing model-based techniques of CPSs become inadequate. For instance, how to identify, analyze the dynamical interplay between urban-scale phenomena (such as mobility demand and supply) from data, and take actions to improve system-level service efficiency is still a challenging and unsolved problem in transportation systems. In this talk, we present a unified data-driven dynamic robust resource allocation framework to match supply towards spatial- temporally uncertain demand, while seeking to reduce total resource allocation cost in real-time. First, we present a receding horizon control framework that incorporates large-scale historical and real-time sensing data in demand prediction and dispatch decisions under practical constraints. However, demand prediction error is not negligible and affects the system’s performance. Therefore, with spatial-temporal demand uncertainty models constructed from data, we then propose two computationally tractable or real-time robust resource allocation methods to provide probabilistic guarantees for the system’s worst-case and expected performances. As a case study, we evaluated the proposed framework using real taxi operational data, and showed that the data- driven robust resource allocation methods reduce the average total idle distance in the city by 55%. Lastly, I will provide an overview of my research that uses the knowledge of the system dynamics for guarantee security and resiliency properties of CPSs and smart cities. I will introduce my research of coding schemes for stealthy data injection attacks detection, and stochastic game schemes for resilient control of CPSs.
About the Speaker: Fei Miao received the B.Sc. degree in Automation from Shanghai Jiao Tong University, Shanghai, China, in June 2010. She received the dual M.A. degree in Statistics in August 2015, and the Ph.D. degree in Electrical and Systems Engineering in May 2016, both from the University of Pennsylvania. Currently, she is a postdoc researcher at the Department of Electrical and Systems Engineering, University of Pennsylvania. She is also a researcher of the CMU-UPENN Safe and Efficient Transportation Center. Her broad research agenda is to develop the foundations for the science of data-driven cyber-physical systems and autonomous transportation systems to assure safety, efficiency and security. Dr. Miao received the “Charles Hallac and Sarah Keil Wolf Award for Best Doctoral Dissertation” in Electrical and Systems Engineering Department from University of Pennsylvania, in 2016, and she was a Best Paper Award Finalist at the 6th ACM/IEEE International Conference on Cyber-Physical Systems (ICCPS) in 2015.
Energy‐Efficient Digital Hardware for Learning Complex Systems
Speaker:Jae Ha Kung, Georgia Institute of Technology
Time: 11:00 am - 12:00 pm Apr 6, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
System learning is the most fundamental research area in engineering domain. It is a modeling method to map external inputs to the corresponding outputs with/without physically analyzing the system between them. The system can be simple enough, e.g. a linear time‐invariant system, to be easily identified by a simple mathematical model. However, it can be a more complex system, such as a nonlinear dynamic system, which is highly difficult to understand with mathematical representations.
In this presentation, energy‐efficient digital hardware to understand the wide range of complex systems using different approaches will be presented. As a data‐driven approach, several neural network algorithms are selected for the system learning. The focused system is related to vision tasks such as image or video processing. Several design algorithms and analysis to realize low‐power neural network accelerators will be discussed. The proposed low‐power design methods are not limited to certain tasks, but are based on algorithmic analysis for general applicability. For a model‐based approach, a programmable and efficient hardware for simulating dynamical systems will be presented. The proposed platform accelerates the computation of solving a wide class of differential equations by utilizing a computing model called cellular nonlinear network with novel system architecture.
About the Speaker: Jae Ha Kung received the B.S. degree in electrical engineering from Korea University, Seoul, Korea, in 2010, and the M.S. degree in electrical engineering from KAIST in Daejeon, Korea, in 2012. He will receive the doctoral degree in electrical and computer engineering at Georgia Institute of Technology, Atlanta, GA in May, 2017. His research interests include energy-efficient digital accelerators for deep learning, high-performance solver for dynamical systems, and real-time thermal management.
Structural Results for Coding Over Communication Networks
Speaker:Farhad Shirani, University of Michigan
Time: 4:00 pm - 5:00 pm Apr 6, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Due to the continuously growing necessity for efficient and reliable data compression and communication, information theory has garnered great interest, and has inspired hundreds of researchers over the past decades. In Point-to- Point (PtP) communications, the optimal achievable performance was characterized by Shannon. However, the problem of characterizing and achieving optimal performance in network communications has remained open.
In this talk, we investigate the short-comings of the existing unstructured and structured coding strategies for networks. We prove the necessity of structure in optimality achieving codes. We consider two categories of structured codes. First, we study codes with algebraic structure. These include linear codes, lattice codes, quasi-linear codes and group codes. We prove the necessity of these structures in multi-terminal communications and derive improved achievable regions for the problem of reliable communication over the interference channel, the distributed source coding problem, the multiple-descriptions problem, and the problem of computation over the multiple-access channel. In the second part, we consider codes constructed over finite effective-lengths. The assumption that the performance of coding schemes improves with increasing effective-length has been a hallmark of multi-terminal communication strategies. We prove that this assumption is incorrect in various multi-terminal settings.
About the Speaker: Farhad Shirani is a postdoctoral researcher working with Professor Sandeep Pradhan in the Electrical Engineering and Computer Science Department at the University of Michigan. His research interests are in classical information theory, quantum information theory, privacy and network anonymity. Farhad received his Ph.D. from the University of Michigan in 2016. Prior to his doctoral work he received his M.S. in Electrical Engineering and M.S. in Applied Mathematics from the University of Michigan, and his B.Sc. from the Sharif University of Technology.
Utility of the Future: New York’s Reforming the Energy Vision (REV) program in Con Edison
Speaker:Stephen Wemple, Con Edison
Time: 11:00 am - 12:00 pm Apr 13, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Stephen Wemple, General Manager of Con Edison’s Utility of the Future team, will provide an overview of New York’s Reforming the Energy Vision (REV) which is transitioning the electricity distribution grid into a dynamic, bidirectional system with increased deployment of customer-sited Distributed Energy Resources (DER). Mr. Wemple will explain some of the existing plans to deploy smart meters throughout the Con Edison and O&R systems, which will provide more visibility into and control of the distribution grid, as well as a recently completed study with EPRI and SCE on the Time and Locational Value of DER. The EPRI study modelled and analyzed power-flows on one the Con Edison networks to demonstrates the opportunities and challenges to defer traditional infrastructure investment with DER. The EPRI work also demonstrated how inverter-based technologies can cause voltage issues, which, without smart-inverter capability, will limit and/or increase the cost of additional DER deployment.
About the Speaker: Stephen Wemple is the General Manager of Con Edison’s Utility of the Future Team responsible for policy issues associated with New York’s Reforming the Energy Vision (REV) proceeding. Prior to joining the Utility of Future Team, Mr. Wemple was Vice President of Regulatory Affairs representing Con Edison’s non-utility affiliates, Con Edison Development, Con Edison Energy and Con Edison Solutions in State and Federal regulatory proceedings.
Mr. Wemple has been an active participant in the New York, New England and PJM wholesale markets and served as Chair of the NYISO’s Business Issues Committee in 2002. He has worked for the Con Edison family of companies for 30 years with responsibilities ranging from resource planning for steam-electric generation, the design and implementation of energy efficiency programs, the development of retail access programs as well as marketing and business development for the wholesale and retail commodity businesses.
Mr. Wemple received his Bachelor of Science and Master of Engineering degrees from Cornell University and is a volunteer firefighter.
Massive MIMO and Beyond
Speaker:Thomas L. Marzetta, Bell Labs
Time: 11:00 am - 12:00 pm Apr 26, 2017
Location: LC400, 5 MetroTech Center, Brooklyn, NY
Massive MIMO (Multiple Input, Multiple Output) wireless communications creates virtual parallel circuits, each occupying the full spectral bandwidth, between a multiplicity of single-antenna terminals and a numerically large array of individually controlled antennas, with huge benefits in terms of area spectral efficiency, uniformly great service, and energy efficiency. Massive MIMO is emerging as the premier fifth-generation wireless physical layer technology.
A pertinent and highly speculative research question is whether 10x improvements over Massive MIMO are possible. A serious attack on this problem entails discarding the overly simplistic propagation models favored by communication theorists, and rigorously combining electromagnetic theory with communication theory.
There is considerable scope and opportunity for innovative research in non-cellular deployments of both Massive MIMO and the earlier Point-to-Point MIMO. Future neural microelectrode arrays will require entirely new wireless transmission schemes to keep up with the data rates. The uplink version of Massive MIMO could permit continuous wireless collection of signals from vast number of sensors, with possible applications to 3D reflection seismology, monitoring of active volcanoes, and structural health monitoring. The principles of MIMO are not limited solely to electromagnetic propagation: acoustic and elastic wave propagation remain to be exploited, as well as communication in media governed by parabolic (heat equation) or elliptic (electrical conductivity) partial differential equations.
About the Speaker: Thomas Marzetta was born in Washington, D.C. He received the PhD and SB in Electrical Engineering from Massachusetts Institute of Technology in 1978 and 1972, and the MS in Systems Engineering from University of Pennsylvania in 1973. After careers in petroleum exploration at Schlumberger-Doll Research and defense at Nichols Research Corporation, he joined Bell Labs in 1995 where he is currently a Bell Labs Fellow. Previously he directed the Communications and Statistical Sciences Department within the former Mathematical Sciences Research Center. He is the lead author of the book “Fundamentals of Massive MIMO”.
Dr. Marzetta was on the Advisory Board of MAMMOET (Massive MIMO for Efficient Transmission), an EU-sponsored FP7 project, and he was Coordinator of the GreenTouch Consortium’s Large Scale Antenna Systems Project. He has received awards including the 2015 IEEE Stephen O. Rice Prize, the 2015 IEEE W. R. G. Baker Award, and the 2013 IEEE Guglielmo Marconi Prize Paper Award. He was elected a Fellow of the IEEE in 2003, and he received an Honorary Doctorate from Linköping University in 2015.
High Capacity Wireless Networks Architectures Through Collaboration and Intelligent Information Storage
Speaker:Leandros Tassiulas, Yale University
Time: 11:00 am - 12:00 pm Apr 27, 2017
Location: 2MTC, 8th floor, Room 800, Brooklyn, NY
A significant portion of today's network traffic is due to recurring downloads of popular content (e.g., movies, video clips and daily news). It has been observed that replicating the latter in caches installed at the network edge -close to the users- can drastically reduces network bandwidth usage and improve content access delay. The key technical issues in emergent caching architectures relate to the following questions: where to install caches, what content and for how long to cache, and how to manage the routing of content within the network. In this talk, an overview of caching is provided, starting with generic architectures that can be applied to different networking environments, and moving to emerging architectures that enable caching in wireless networks (e.g., at cellular base stations and WiFi access points). Novel challenges arise in the latter due to the inadequacy of wireless resources and their broadcast nature, the frequent hand-offs between different cells for mobile users, as well as the specific requirements of different types of user applications, such as video streaming. We will present our recent results on innovative caching approaches that (i) harvest idle user-owned cache space and bandwidth, (ii) leverage the broadcast nature of the wireless medium to serve concurrent requests for content (iii) exploit the regularity of user mobility patterns, and (iv) apply advanced video encoding techniques to support multiple video qualities (e.g., screen sizes, frame rates, or signal-to-noise ratio (SNR) qualities). These are cutting-edge approaches that can achieve significant performance and cost-reduction benefits over the state-of-the-art methods.
This is a part of the Jack Wolf Lecture Series.
About the Speaker: Leandros Tassiulas is the John C. Malone Professor of Electrical Engineering at Yale University. His research interests are in the field of computer and communication networks with emphasis on fundamental mathematical models and algorithms of complex networks, architectures and protocols of wireless systems, sensor networks, novel internet architectures and experimental platforms for network research. His most notable contributions include the max-weight scheduling algorithm and the back-pressure network control policy, opportunistic scheduling in wireless, the maximum lifetime approach for wireless network energy management, and the consideration of joint access control and antenna transmission management in multiple antenna wireless systems. Dr. Tassiulas is a Fellow of IEEE (2007). His research has been recognized by several awards including the IEEE Koji Kobayashi computer and communications award 2016, the inaugural INFOCOM 2007 Achievement Award “for fundamental contributions to resource allocation in communication networks,” the INFOCOM 1994 best paper award, a National Science Foundation (NSF) Research Initiation Award (1992), an NSF CAREER Award (1995), an Office of Naval Research Young Investigator Award (1997) and a Bodossaki Foundation award (1999). He holds a Ph.D. in Electrical Engineering from the University of Maryland, College Park (1991). He has held faculty positions at Polytechnic University, New York, University of Maryland, College Park, and University of Thessaly, Greece.
The Panorama of LARX: The Systems, the Data and the Games
Speaker:Quanyan Zhu, New York University
Time: 11:00 am - 12:00 pm May 4, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
In this talk, we will give an introduction to the laboratory of agile and resilient complex systems (LARX). The theme of the lab centers around the recent development in the area of systems, control and games, and their roles in the emerging applications in cyber-physical security, Internet of Things (IoT), smart communities and network economics. We will show the gradual convergence of multi-disciplinary system sciences including systems and control theory, data science, information theory, game theory, signal processing, communications, economics and operations research. We will use recent examples of cyber-physical security of cloud-enabled robotics, contract design for the Internet of Controlled Things, and attack-aware cyber insurance as prime examples to demonstrate the recent development of multi-disciplinary system sciences on complex systems. Students are encouraged to take advanced system courses and participate in research projects to develop their system thinking in problem solving.
About the Speaker: Quanyan Zhu is a faculty member of the ECE department. He received B. Eng. in Honors Electrical Engineering from McGill University in 2006, M.A.Sc. from University of Toronto in 2008, and Ph.D. from the University of Illinois at Urbana-Champaign (UIUC) in 2013. After a short stint at Princeton University, he joined the Department of Electrical and Computer Engineering at New York University (NYU) as an assistant professor in 2014.
Improving Resolution, Quality and Productivity for Scanning Electron Microscope Images by the Use of Deconvolution and Regularization
Speaker:Eric Lifshin, SUNY Polytechnic Institute
Time: 1:00 pm - 2:00 pm May 17, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Scanning Electron Microscopes (SEMs) have been available for about 50 years. They are used extensively in many fields including life and materials science as well as engineering, failure analysis and quality control. The first instruments used thermionic sources and after some development achieved a maximum obtainable resolution of 2 or 3 nm under ideal conditions at moderate beam energies (20 keV), however the signal to noise ratio drops below acceptable values at low beam energies (below 2 keV). Such energies are required for improved surface sensitivity or to minimize damage in the examination of a variety of samples including tissue specimens and certain microelectronic devices. To address this challenge Schottky source and cold field emission instruments were developed that give 1 nm or better resolution at low keV, but generally cost twice as much as those with thermionic sources.
An alternative to expensive hardware improvements is the use of advanced software for image noise reduction and improved resolution through deconvolution. This approach is now applied extensively in surveillance, medical imaging and astronomy to restore images that have been altered by the observation process itself, as in the case of atmospheric and noise effects that degrade astronomical telescope images. These methods are much less expensive than specialized hardware improvements and are well suited to low cost workstations. This talk will describe how this approach can now be applied to SEM image restoration and point out the opportunities and challenges.
About the Speaker: Eric Lifshin is a Professor at the College of Nanoscale Science and Engineering (CNSE) at SUNY Polytechnic Institute in Albany, New York. He holds a BS in Physics, and MS and PhD in Materials Engineering all from Rensselaer Polytechnic Institute. Prior to joining CNSE he was the Manager of the Materials Characterization and Environmental Technology Laboratory at GE Corporate R&D. His group consisted of about 80 scientists, engineers, technicians and administrative assistants involved with a range of activities of interest to GE. He began his career as a research staff scientist specializing in a variety of microanalysis methods particularly scanning electron microscopy and x-ray microanalysis and has stayed very active in those fields.
Semi-Supervised Graph Classifier Learning with Negative Edge Weights
Speaker:Gene Cheung, National Institute of Informatics, Japan
Time: 11:00 am - 12:00 pm Jun 5, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
In a semi-supervised learning scenario, (possibly noisy) partially observed labels are used as input to train a classifier, in order to assign labels to unclassified samples. We construct a complete graph-based binary classifier given only samples' feature vectors and partial labels. Specifically, we first build appropriate similarity graphs with positive and negative edge weights connecting all samples based on inter-node feature distances. By viewing a binary classifier as a piecewise constant graph-signal, we cast classifier learning as a signal restoration problem via a classical maximum a posteriori (MAP) formulation. One unfortunate consequence of negative edge weights is that the graph Laplacian matrix L can be indefinite, and previously proposed graph-signal smoothness prior x^T L x for candidate signal x can lead to pathological solutions. In response, we derive a minimum-norm perturbation matrix Del that preserves L's eigen-structure---based on a fast lower-bound computation of L's smallest eigenvalue via a novel application of the Haynsworth inertia additivity formula---so that L + Del is positive semi-definite, resulting in a stable signal prior. Finally, we propose an algorithm based on iterative reweighted least squares (IRLS) that solves the posed MAP problem efficiently. Extensive simulation results show that our proposed algorithm outperforms both SVM variants and previous graph-based classifiers using positive-edge graphs noticeably.
About the Speaker: Gene Cheung received the B.S. degree in electrical engineering from Cornell University in 1995, and the M.S. and Ph.D. degrees in electrical engineering and computer science from the University of California, Berkeley, in 1998 and 2000, respectively. He was a senior researcher in Hewlett-Packard Laboratories Japan, Tokyo, from 2000 till 2009. He is now an associate professor in National Institute of Informatics in Tokyo, Japan. He has been an adjunct associate professor in the Hong Kong University of Science & Technology (HKUST) since 2015.
His research interests include 3D image processing, graph signal processing, and signal processing for sleep analysis. He currently serves as associate editor for IEEE Transactions on Image Processing (2015--present), IEEE Transactions on Circuits and Systems for Video Technology (2016--present) and APSIPA Journal on Signal & Information Processing (2011--present), and as area editor for EURASIP Signal Processing: Image Communication (2011--present). He is a distinguished lecturer in APSIPA (2016--2017). He served as a member of the Multimedia Signal Processing Technical Committee (MMSP-TC) in IEEE Signal Processing Society (2012--2014), and a member of the Image, Video, and Multidimensional Signal Processing Technical Committee (IVMSP-TC) (2015--2017). He has also served as technical program co-chair of International Packet Video Workshop (PV) 2010 and IEEE International Workshop on Multimedia Signal Processing (MMSP) 2015, and symposium co-chair for CSSMA Symposium in IEEE GLOBECOM 2012. He is a co-author of the best student paper award in IEEE Workshop on Streaming and Media Communications 2011 (in conjunction with ICME 2011), ICIP 2013 and IVMSP 2016, best paper runner-up award in ICME 2012, best paper finalists in ICME 2011, ICIP 2011 and ICME 2015, and IEEE Signal Processing Society (SPS) Japan best paper award 2016.
Toggle MUX: How X-Optimism Can Lead to Malicious Hardware
Speaker:Christian Krieg, TU Wien, Austria
Time: 12:00 pm - 1:00 pm Jun 29, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
To highlight a potential threat to hardware security, we propose a methodology to derive a trigger signal from the behavior of Verilog simulation models of field-programmable gate array (FPGA) primitives that behave X-optimistic. We demonstrate our methodology with an example trigger that is implemented using Xilinx 7 Series FPGAs. Experimental results show that it is easily possible to create a trigger signal that is ‘0’ in simulation (pre- and post-synthesis), and ‘1’ in hardware. We show that this kind of trigger is neither detectable by formal equivalence checks, nor by recent Trojan detection techniques. As a countermeasure, we propose to carefully reconsider the utilization of X-optimism in FPGA simulation models.
About the Speaker: Christian Krieg received the bachelor's and master's degree in electrical engineering from TU Wien and is now pursuing his PhD studies on hardware security at TU Wien. His research focuses on design-level hardware Trojan design and detection. He also works on reasonable threat models for hardware Trojan attacks. Christian recently received the ICCAD William McCalla best paper award for a novel hardware Trojan implementation. At a wider scope, Christian's research interests include security-driven design understanding, cyber-physical systems security and IoT security.