Seminars: Fall 2017
Recent NETMODE Activities on Internet Research & Experimentation
Speaker: Vasilis Maglaris
Director of the Network Management & Optimal Design (NETMODE) Laboratory, School of Electrical & Computer Engineering, National Technical University of Athens (NTUA), Greece
Time: 11:00 am - 12:00 pm Sep 7, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
The Network Management & Optimal Design Laboratory (NETMODE) is established within the School of Electrical & Electronic Engineering of the National Technical University of Athens (NTUA). NETMODE supports undergraduate and graduate teaching & research on Internet Technologies - Distributed Network/System Management, Future Internet Control & Management Plane Protocols & Architectures and e-Learning Multimedia Toolkits. The lab facilities include an SDN/NFV Testbed with OpenFlow devices and SDN controllers, and a Wireless Testbed. Apart from its usage as proof of concept on novel networking protocols and architectures, it is hosting facilities of global emulation platforms, e.g. PlanetLab, Fed4FIRE, NOVI and FEDERICA.
The talk will briefly introduce recent research activities in the following topics:
Federated e-Infrastructures (NOVI Concept of Data, Control & Management Stitching)
Policy Based Resource Management (NFV model of Policy Orchestration)
Anomaly Detection & Mitigation (Extending Remotely Triggered Black Hole – RTBH; Classification of Malicious Source IP Prefixes; Cooperative Schema for Multi-domain SDN Environments; Collaborative Schema for Exchanging Attack Data; Applying Emerging Tools for Network Security – Advanced Statistical Methods & Machine Learning)
Multi-tenant Monitoring as VNF (Monitoring Architecture for Research in Internet Experimentation; Monitoring in SDN Multi-tenant Environments; Scalable Monitoring-as- a-Service; Application in a Federated Environment of the GÉANT Testbed Service; Monitoring Data Mining via the OmniDisco Collector)
Mobile Broadband Carrier Selection & Offloading by SDN-enabled User Equipment (Monitoring & Analysis for Radio Interface Selection via Open vSwitch Control functionality)
About the Speaker:
Vasilis Maglaris was born in Athens, Greece in 1952. He holds an Engineering Degree from the National Technical University of Athens – NTUA (Athens, Greece, 1974) and a Ph.D. degree from Columbia University (New York NY, USA, 1979). Between 1979 and 1981 he was with the Network Analysis Corporation (Great Neck NY, USA) working on electronic communications advanced projects. From 1981 to 1989 he was with the faculty of Electrical Engineering & Computer Science of Polytechnic University (Brooklyn NY, USA - now New York University Tandon School of Engineering). In 1989 he joined the faculty of the School of Electrical & Computer Engineering at NTUA, teaching and performing research on Internet technologies.
Professor Maglaris established the Network Management & Optimal Design (NETMODE) Laboratory within the School of Electrical & Computer Engineering at NTUA; the laboratory supports undergraduate, graduate and post-doctoral teaching and research, co-financed by European Union and National projects. He authored more than 150 research papers, regularly delivers lectures on Internet advances and supervised 20 Doctoral theses in the USA and Greece.
In addition to his academic duties, he was responsible for the development of the NTUA Campus LAN and for the establishment of GRNET (the Greek National Research & Education Network - NREN), serving as its Chairman from 1998 to 2004. He was on the board of the Greek National Regulatory Authority on Telecommunications and Posts for two five-year terms (1995 – 2005) and an independent expert of the Greek Parliamentary Committee on Secure Communications (1996 – 2002).
From 1994 to 1996, he was the Executive Director of the National Hellenic Research Foundation (NHRF), a front research organization consisting of six Research Institutes, also hosting the Greek National Documentation Centre and operating the largest library of scientific periodicals in Greece.
From October 2004 to June 2012, he served as the Chairman of the National Research & Education Networks Policy Committee (NREN PC). The NREN PC was harmonizing policies amongst the 37 NRENs in the extended European Research Area; it was also responsible for the governance of GÉANT, the Future Internet initiative of the European Research & Academic community.
From July 2012 to June 2013, he served as General Secretary for Research & Technology (GSRT), Ministry of Education, on hold from his duties as NTUA Professor. During his tenure as GSRT he was responsible for setting and monitoring Research & Innovation policies of the Greek coalition Government within national and EU frameworks. This included coordinated planning of University, Research Centre and Private Business R&D related to the EU Horizon 2020 and the Greek Structural Funds Programme 2014-2020.
Reliably Precise State Estimation for Autonomous Highway Vehicles
Speaker: Jay A. Farrell
Department of Electrical and Computer Engineering University of California, Riverside
Time: 11:00 am - 12:00 pm Sep 11, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
Abstract: Key among the challenges inhibiting effective commercial deployment of autonomous and connected vehicles is accurate and reliable awareness of world interactions. Awareness arises from onboard sensors and ubiquitous communication between vehicles and infrastructure. Vehicle coordination and safety necessitate reliable “where-in- lane” knowledge of vehicle position. This presentation will address sensor fusion for high-bandwidth vehicle state estimation with a focus on high precision and reliability.
Standard Extended Kalman Filtering (EKF) is not sufficiently reliable at removing the effects of spurious measurements because it must decide at the time each measurement arrives whether or not it is valid. When that decision is wrong, either measurement information is lost or the state and covariance estimates are corrupted. Either situation can result in EKF divergence.
An alternative new approach is to extract the Bayesian optimal trajectory from a window of sensor data to minimize risk subject to accuracy constraints. We refer to this approach as a Contemplative Real-Time (CRT) estimator, as it is able to reconsider outlier assumptions for measurements within the window. This presentation will review the interrelationships between the EKF, Iterated Extended Kalman Filter (IEKF), and the CRT. It will also discuss its relationship with Moving Horizon Estimation (MHE), Simultaneous Localization and Mapping (SLAM), and sparse signal processing. Both theoretical and experimental results will be presented.
Biography:Jay A. Farrell is a Professor in the Department of Electrical and Computer Engineering at the University of California, Riverside. His research focuses on state estimation, control, and planning for nonlinear systems, particularly autonomous vehicles. He served the IEEE Control Systems Society (CSS) as Finance Chair for three IEEE CDC`s, on the Board of Governors for two terms, as Vice President Finance and Vice President of Technical Activities, as General Chair of IEEE CDC 2012, and as President in 2014. In 2018, he will serve as Vice President of the American Automatic Control Council. He was a GPS World Magazine GNSS Leader to Watch for 2009-2010 and a winner of the U.S. Department of Transportation`s Connected Vehicle Technology Challenge in 2011. He is author of over 250 technical publications and three books, a Distinguished Member of IEEE CSS, a Fellow of AAAS, and a Fellow of the IEEE.
Revisiting Downlink Scheduling in a Multi-Cell OFDMA Network: From Full Base-Station Coordination to Practical Schemes
Speaker: Catherine Rosenberg
Department of Electrical and Computer Engineering, University of Waterloo, Canada
Time: 11:00 am - 12:00 pm Sep 12, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
Abstract: We revisit the user scheduling problem on the downlink of OFDMA cellular networks. Our aim is to find the optimal system-wide schedule of a multi-cell system to understand how much a coordinated scheduling could improve performance if full coordination between base-stations was possible. We make no a priori assumption on how power is allocated to resource blocks and how resource blocks are allocated to users and we use a realistic modulation and coding scheme rate function that is piecewise constant. Hence, the corresponding scheduling problem is very general. It is a highly non-convex mixed integer non-linear problem that includes many binary variables. To the best of our knowledge, such a general problem has never been solved before. We propose a method to upper bound this problem by a signomial programming problem that can be solved efficiently. The solution to the signomial problem can be used to derive a feasible solution to the original global scheduling problem. We show numerically that the gap between the upper bound and that feasible solution is very small, implying that the upper bound is very tight.
We then provide results both for homogeneous and heterogeneous networks. In the homogeneous case, we use the feasible solutions to the joint problem to derive a practical scheme based on a well-parameterized soft frequency reuse and a simple local scheduler. We show that the coordinated scheduling performs only 20% better than the (local) practical scheme and hence, one might question the need for coordination due to its inherent complexity. In the heterogeneous case, we also compare the solution obtained by fully coordinating the scheduling of all the base-stations with a practical scheme that we design and find that the performances of that scheme are not too far from the performances of the fully coordinated case.This work was done with Yigit Ozcan.
Biography: Catherine Rosenberg is a Professor in Electrical and Computer Engineering at the University of Waterloo. Since June 2010, she holds the Tier 1 Canada Research Chair in the Future Internet. From 1999 to 2004, Prof. Rosenberg was a Professor in the School of Electrical and Computer Engineering at Purdue University.
Prof. Rosenberg was a member of the Scientific Advisory Board of Orange (formerly France-Telecom) from 2007 to 2015 and its President in the last 2 years. She is the president of the Scientific Advisory Board of the French IRT (Research and Technology Institute) B<>COM on multimedia and networking since January 2014. She was elected an IEEE Fellow for contributions to resource management in wireless and satellite networks on 2011 and a Fellow of the Canadian Academy of Engineering in 2013. Her research interests are mainly in three areas: the Internet, Wireless Networks, and Energy Systems.
An Information Theoretic Perspective of Fronthaul Constrained Cloud and Fog Radio Access Networks
Speaker: Shlomo Shamai
The Andrew and Erna Viterbi Department of Electrical Engineering, Technion-Israel Institute of Technology
Time: 11:00 am - 12:00 pm Sep 14, 2017
Location: 2MTC, 10th floor, Room 10.099, Brooklyn, NY
Cloud radio access networks (C-RANs) emerge as appealing architectures for next-generation wireless/cellular systems whereby the processing/decoding is migrated from the local base-stations/radio units (RUs) to a control/central units (CU) in the "cloud". Fog radio access networks (F-RAN) address the case where the RUs are enhanced by having the ability of local caching of popular contents. The network operates via fronthaul digital links connecting the CU and the RUs.In this talk we will address basic information theoretic aspects of such networks, with emphasis of simple oblivious processing. Theoretical results illustrate the considerable performance gains to be expected for different cellular models. Some interesting theoretical directions conclude the presentation.
About the Speaker: Shlomo Shamai (Shitz) received the B.Sc., M.Sc., and Ph.D. degrees in electrical engineering from the Technion—Israel Institute of Technology, in 1975, 1981 and 1986 respectively.
During 1975-1985 he was with the Communications Research Labs in the capacity of a Senior Research Engineer. Since 1986 he is with the Department of Electrical Engineering, Technion-Israel Institute of Technology, where he is now a Technion Distinguished Professor, and holds the William Fondiller Chair of Telecommunications.
Dr. Shamai (Shitz) is an IEEE Fellow, a Member of the Israeli Academy of Sciences and Humanities and a Foreign Member of the US National Academy of Engineering. He is the recepient of the 2011 Claude E. Shannon Award, the 2014 Rothschild Prize in Mathematics/Computer Sciences and Engineering and the 2017 IEEE Richard W. Hamming Medal.
His research work encompasses a wide spectrum of topics in information theory and statistical communications to which he has contributed fundamentally. Some highlights of his scientific work comprise: Conclusive results on the capacity of the multi-input-multi-output broadcast Channels; Establishing basic connections between information theory and statistical estimation theory; Introducing pioneering concepts of interference alignment for communications networks and currently providing a unified information theoretic framework of cloud and fog radio access networks.
He has been awarded the 1999 van der Pol Gold Medal of the Union Radio Scientifique Internationale (URSI), and is a co-recipient of the 2000 IEEE Donald G. Fink Prize Paper Award, the 2003, and the 2004 joint IT/COM societies paper award, the 2007 IEEE Information Theory Society Paper Award, the 2009 and 2015 European Commission FP7, Network of Excellence in Wireless COMmunications Best Paper Awards, the 2010 Thomson Reuters Award for International Excellence in Scientific Research, the 2014 Europian Association for Signal Processing (EURASIP) Best Paper Award, and the 2015 IEEE Communications Society Best Tutorial Paper Award. He is a Highly Cited researcher and is listed in the 2015 Thomson Reuters “The World’s Most Influential Scientific Minds”. He is also the recipient of 1985 Alon Grant for distinguished young scientists and the 2000 Technion Henry Taub Prize for Excellence in Research. He has served as Associate Editor for the Shannon Theory of the IEEE Transactions on Information Theory, and twice on the Board of Governors of the Information Theory Society. He has also served on the Executive Editorial Board of the IEEE Transactions on Information Theory and on the IEEE Information Theory Society Nominations and Appointments Committee.
Designing fast and programmable routers
Speaker:Anirudh Sivaraman, New York University
Time: 11:00 am - 12:00 pm Sep 21, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
Historically, the evolution of network routers was driven primarily by performance. Recently, owing to the need for better control over network operations and the constant demand for new features, programmability of routers has become as important as performance. However, today's fastest routers use fixed-function hardware, which cannot be modified after deployment. I will describe two router primitives we have developed to build fast and programmable routers. The first is a programmable packet scheduler. The second is a way to execute stateful packet-processing algorithms to manage network resources such as packet buffers and link capacity. Together, these primitives allow us to program several packet-processing functions at speeds approaching today's fastest routers. These functions include in-network congestion control, active queue management, data-plane load balancing, network measurement, and packet scheduling.
This talk is based on joint work with collaborators at MIT, Barefoot Networks, Cisco Systems, Microsoft Research, Stanford University, and the University of Washington.
About the Speaker: Anirudh Sivaraman is an assistant professor at NYU's Computer Science Department. His recent research work has focused on hardware and software for programmable high-speed routers. He has also been actively involved in the design and evolution of the P4 language for programmable network devices. His past research includes work on congestion control, network emulation, improving Web performance, and network measurement. He received the MIT EECS department's Frederick C. Hennie III Teaching Award in 2012 and shared the Applied Networking Research Prize in 2014 and the ACM SIGCOMM Best Paper Award in 2017.
Codes for speed: large-scale computation, signal recovery, and learning
Speaker:Kannan Ramchandran, University of California, Berkeley
Time: 10:00 am - 12:00 pm Oct 5, 2017
Location: 5 MetroTech Center, LC400, Brooklyn, NY
Abstract: Seven decades after Claude Shannon's groundbreaking work, codes are now an indispensable part of modern communications and storage systems. But do they have a role in today's information age that is witness to exponential data deluge? Can codes help address the challenge of scale in computation, inference, and learning by exploiting underlying structure such as sparsity? In this talk, we will explore how coding theory can go well beyond traditional communications applications, and can indeed offer an unconventional and valuable playground for some of these problems, with an emphasis on speed.
Specifically, we will view a diverse class of problems through the lens of sparse-graph codes. They form the core of a unified architecture featuring a divide-and-conquer strategy built on simple guess-and-check primitives, and fast peeling-based decoding. This allows for real-time sparse-structure recovery involving large datasets, in contrast to popular convex relaxation based optimization methods that can be computationally difficult to scale. We will illustrate our approach to computational tasks such as massive-scale sparse Fourier and Walsh transforms, sparse polynomial learning, support recovery in compressed sensing, phase-retrieval, and group testing, while unveiling insightful connections between sampling theory and coding theory. The application space is broad, encompassing MRI and optical imaging, hyper-graph sketching, fast neighbor discovery in IoT, spectrum sensing for cognitive radio, and learning mixtures of sparse linear regressions. Time-permitting, we will also highlight how codes can speed up machine learning in today's distributed cloud computing systems by rendering them robust to system noise in the form of `straggling' compute nodes.
Bio: Kannan Ramchandran has been a Professor of Electrical Engineering and Computer Science at UC Berkeley since 1999. He was on the faculty at the University of Illinois from 1993 to 1999. Prof. Ramchandran is a recipient of the 2017 IEEE Kobayashi Computers and Communications award for his contributions to the theory and practice of distributed storage coding and distributed compression. He is a Fellow of the IEEE, has published extensively in his field, and holds over a dozen patents. He has received several awards for his research and teaching including an IEEE Information Theory Society and Communication Society Joint Best Paper award for 2012, an IEEE Communication Society Data Storage Best Paper award in 2010, two Best Paper awards from the IEEE Signal Processing Society in 1993 and 1999, an Okawa Foundation Prize for outstanding research at Berkeley in 2001, and an Outstanding Teaching Award at Berkeley in 2009. His research interests lie at the broad intersection of signal processing, machine learning, coding and information theory, and peer-to-peer networking.
Data compression and learning with logarithmic loss
Speaker:Yanina Shkel, Princeton University
Time: 11:00 am - 12:00 pm Oct 12, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
Abstract: In this talk we discuss data compression, as well as learning from data, with a specific performance criterion: the logarithmic loss (log-loss). The log-loss measures distortion in settings when the reconstructed information is soft, that is, a distribution over possible values is provided. It is natural to construct data processing modules which interface by processing such soft information: consider, for example, the belief propagation algorithm that works by passing probabilities between nodes in order to perform statistical inference on graphical models. Conceptually, log-loss is an important performance criterion as it provides a bridge between the theory of data compression and sequential prediction, and has nice mathematical properties which allow for strong bounds for both problems. The focus of our talk will be on the non-asymptotic and universal fundamental limits of lossy compression where we obtain simple and elegant results that generalize those in lossless compression. A particular emphasis will be made on new connections between the problems of data compression and sequential prediction with logarithmic loss.
Bio: Yanina Shkel is a postdoctoral scholar in the department of Electrical Engineering at Princeton University; before this she was a postdoctoral fellow with the NSF Center for Science of Information where she worked with collaborators at Princeton University and University of Illinois at Urbana-Champaign. Yanina has B.S. degrees in Mathematics and in Computer Science, as well as a Ph.D. degree in Electrical and Computer Engineering from University of Wisconsin-Madison. Before attending graduate school she worked as a developer for Morningstar, Inc. where she administered databases containing and processing large amounts of financial data. More recently, she was an intern at 3M Corporate Research Labs where she had a unique opportunity to utilize her background in computation and information sciences for materials and product driven needs of 3M.
Yanina is broadly interested in identifying laws which govern the behavior of information in both engineered and naturally occurring systems, and using these laws to better understand the capabilities of such systems. Her most recent research focuses on points of connection between information theory, statistics, and machine learning.
Securing the Internet of Things (IoT): Need for a New Paradigm
Speaker:Tao Zhang, Cisco
Time: 11:00 am - 12:00 pm Oct 19, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
Abstract: The IoT will interconnect a significantly larger number and broader range of things than today’s Internet. These will range from simpler things (e.g., sensors and wearables) to complex things (e.g., vehicles, drones, and robots) to large systems (e.g., industrial control systems, connected transportation systems, smart buildings, and smart grids). As IoT emerges, security attacks are also moving from the average desktop software to the real world brought online by IoT—machines, sensors, and cars. Protecting IoT systems imposes new and unique challenges. In this talk, I will discuss some of these new challenges and why the current security paradigm cannot adequately address them. I will dive deeper into one specific challenge – how to automatically determine whether an IoT system behaves normally – and show why existing technologies cannot adequately address this challenge for IoT systems and describe one potential solution.
Bio: For over 25 years, Dr. Tao Zhang has been in various technical and executive positions leading research and development, which have created disruptive new technologies and products in fog computing, IoT, vehicular networks, all-IP cellular networks (3G/4G), and fiber optic networks.
He is Distinguished Engineer / Senior Director in Cisco Corporate Strategic Innovation Group. He joined Cisco in 2012 as the Chief Scientist / CTO for Smart Connected Vehicles, and has since also been driving strategies and new technology development for fog computing and IoT security. He cofounded the OpenFog Consortium, serves as its Board Director, and has been leading fog computing into a global industry trend and a vibrant research field. Inside Cisco, he guides fog computing strategies and architectures, and heads co-innovation programs on fog computing, IoT security, and connected vehicles.
Prior to Cisco, Tao was Chief Scientist and Director of Mobile and Vehicular Networking at Telcordia Technologies. He managed several R&D groups, and led teams to win many contracts from government and industry customers to build new R&D programs. His leadership and work helped Telcordia break into new industries such as connected vehicles and mobile networks, established Telcordia as a leader in these industries, and created new technologies that have been incorporated into international standards and led to first-in-the-industry commercial products. He was a cofounder and founding Board Director of the Connected Vehicle Trade Association (CVTA) and helped make it a highly influential champion in connected vehicles.
Dr. Zhang, an IEEE Fellow, holds 50+ patents and coauthored two books “Vehicle Safety Communications: Protocols, Security, and Privacy” (2012) and “IP-Based Next Generation Wireless Networks” (2004), one book chapter “Securing the Internet of Things: Need for a New Paradigm and Fog Computing” (2016), and 70+ technical papers. He is a Chair Professor at National Chiao Tung University, the CIO and a Board Governor of the IEEE Communications Society, and a Distinguished Lecturer of the IEEE Vehicular Technology Society. He has served on advisory boards for multiple organizations. He cofounded and served on leadership roles for multiple international conferences, and has been an editor or guest editor for many technical journals.
Communication-efficient and differentially-private distributed learning
Speaker:Theertha Suresh, Google Research, New York
Time: 11:00 am - 12:00 pm Oct 26, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
Abstract: Motivated by the need for distributed learning and optimization algorithms with low communication cost and privacy guarantees, we study communication efficient and differentially-private algorithms for distributed mean estimation. Unlike previous works, we make no probabilistic assumptions on the data. We propose three quantization schemes one of which is optimal up to a constant in the minimax sense i.e., it achieves the best mean square error for a given communication cost. Furthermore, we show that a modified version of the quantization schemes achieves differential privacy in addition to communication efficiency. We finally demonstrate the practicality of our algorithms by applying them to distributed gradient descent for neural networks, Lloyd's algorithm for k-means, and power iteration for PCA.
About the Speaker: Ananda Theertha Suresh received the B.Tech. degree from IIT Madras in 2006, and M.S. and Ph.D. degrees from University of California at San Diego in 2012 and 2016, respectively. He is currently a Research Scientist at Google, New York. His research interests lie in the intersection of statistics, machine learning, and information theory. He is a recipient of the 2015 Neural Information Processing Systems (NIPS) Best Paper Award and 2017 Marconi Society Paul Baran Young Scholar Award.
Fog-Aided Wireless Networks: An Information-Theoretic View
Speaker:Osvaldo Simeone, King's College London, United Kingdom
Time: 11:00 am - 12:00 pm Nov 2, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
Abstract: Fog-aided wireless networks are an emerging class of wireless systems that leverage the synergy and complementarity of cloudification and edge processing, two key technologies in the evolution towards 5G systems and beyond. The operation of fog-aided wireless networks poses novel fundamental research problems pertaining to the optimal management of the communication, caching and computing resources at the cloud and at the edge, as well as to the transmission on the fronthaul network connecting cloud and edge. In this talk, it will be argued via specific examples concerning the problem of content delivery that network information theory provides a principled framework to develop fundamental theoretical insights and algorithmic guidelines on the optimal design of fog-aided networks. This work has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 725731).
About the Speaker: Osvaldo Simeone is a Professor of Information Engineering with the Centre for Telecommunications Research at the Department of Informatics of King's College London. He received an M.Sc. degree (with honors) and a Ph.D. degree in information engineering from Politecnico di Milano, Milan, Italy, in 2001 and 2005, respectively. He was previously a Professor with the Center for Wireless Information Processing (CWiP) at New Jersey Institute of Technology (NJIT). His research interests include wireless communications, information theory, optimization and machine learning. Dr Simeone is a co-recipient of the 2017 JCN Best Paper Award, the 2015 IEEE Communication Society Best Tutorial Paper Award and of the Best Paper Awards of IEEE SPAWC 2007 and IEEE WRECOM 2007. He was awarded a Consolidator grant by the European Research Council (ERC) in 2016. His research has been supported by the U.S. NSF, the ERC, the Vienna Science and Technology Fund, as well by a number of industrial collaborations. He is currently a Distinguished Lecturer of the IEEE Information Theory Society. Dr Simeone is a co-author of a monograph, an edited book published by Cambridge University Press and more than one hundred research journal papers. He is a Fellow of the IEEE.
Adaptive Algorithms for Caching Networks with Optimality Guarantees
Speaker:Edmund Yeh, Northeastern University
Time: 11:00 am - 12:00 pm Nov 3, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
Abstract: We study the problem of optimal content placement over a network of caches. This problem arises naturally in several contexts, including information-centric networks (ICNs), content delivery networks (CDNs), and peer-to-peer (P2P) networks. Under a set of demands and routes that requests follow, we wish to determine the content placement across the network that maximizes the expected caching gain, i.e., the reduction of routing costs due to intermediate caching. This objective is not convex, and the offline version of this problem is NP-hard. We seek adaptive algorithms for assigning content to caches that operate without prior knowledge of the demand. We propose a distributed, adaptive algorithm that performs stochastic gradient ascent on a concave relaxation of the caching gain, and constructs a probabilistic content placement within a 1-1/e approximation factor from the optimal allocation, in expectation. Motivated by our analysis, we also propose a simpler adaptive heuristic, and show through numerical evaluations that both algorithms significantly outperform traditional algorithms like LRU and LFU over a broad array of network topologies. Next, we focus on the more general problem of jointly optimizing caching and routing decisions to minimize routing costs over an arbitrary network topology. We show that our concave relaxation approach extends to this more general setting, and produce distributed, adaptive algorithms with the same approximation guarantees. The resulting algorithms are shown to outperform the state of the art by several orders of magnitude. Joint work with Stratis Ioannidis
About the Speaker: Edmund Yeh received his B.S. in Electrical Engineering with Distinction and Phi Beta Kappa from Stanford University in 1994. He then studied at Cambridge University on the Winston Churchill Scholarship, obtaining his M.Phil in Engineering in 1995. He received his Ph.D. in Electrical Engineering and Computer Science from MIT under Professor Robert Gallager in 2001. He is currently Professor of Electrical and Computer Engineering at Northeastern University. He was previously Assistant and Associate Professor of Electrical Engineering, Computer Science, and Statistics at Yale University. He has held visiting positions at MIT, Stanford, Princeton, UC Berkeley, EPFL, and TU Munich. Professor Yeh was one of the PIs on the original NSF-funded FIA Named Data Networking project. He will serve as General Co-Chair for ACM Conference on Information Centric Networking (ICN) 2018 in Boston. He is the recipient of the Alexander von Humboldt Research Fellowship, the Army Research Office Young Investigator Award, the Winston Churchill Scholarship, the National Science Foundation and Office of Naval Research Graduate Fellowships, the Barry M. Goldwater Scholarship, the Frederick Emmons Terman Engineering Scholastic Award, and the President's Award for Academic Excellence (Stanford University). Professor Yeh has served as the Secretary of the Board of Governors of the IEEE Information Theory Society, as well as Associate Editor for IEEE Transactions on Networking, IEEE Transactions on Mobile Computing, and IEEE Transactions on Network Science and Engineering. He received the Best Paper Award at the 2017 ACM Conference on Information-centric Networking (ICN), and at the 2015 IEEE International Conference on Communications (ICC) Communication Theory Symposium.
OTFS - next generation modulation scheme addressing the challenges of 5G
Speaker:Ronny Hadani, University of Texas at Austin
Time: 11:00 am - 12:00 pm Nov 16, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
Abstract: In this talk I will introduce a novel modulation scheme referred to as OTFS (Orthogonal Time Frequency and Space) which is based on multiplexing information QAM symbols on localized pulses in the delay-Doppler signal representation. I will explain the mathematical foundations of OTFS, emphasizing how the underlying structure establishes a conceptual link between communication and Radar theory. I will show how OTFS naturally generalizes conventional time and frequency modulations such as TDM and FDM. I will also discuss the unique manner in which OTFS waveforms couple with the wireless channel which allows the coherent combining of all the time and frequency diversity modes of the channel to maximize the received energy. Finally, I will explain the intrinsic potential advantages of OTFS over multicarrier modulation schemes in the context of various 5G use cases such as MU-MIMO equalization and precoding, communication under high Doppler conditions (V2X and V2V) and communication under power constraints (such as in IoT).
About the Speaker: Ronny Hadani is an associate professor in the Mathematics Department of the University of Texas at Austin. Before that, he was a Dickson postdoctoral fellow in the Mathematics Department of the University of Chicago. He holds a PhD in pure mathematics from Tel-Aviv University under the direction of Prof. Joseph Bernstein and a Master degree in applied mathematics from The Weizmann Institute of Science. He also serves as the Chief Technology Officer at Cohere Technologies.
Distributed algorithms for Cyberphysical Systems
Speaker:Nick Freris, New York University Abu Dhabi
Time: 11:00 am - 12:00 pm Dec 14, 2017
Location: 2 MetroTech Center, 10th floor, Room 10.099, Brooklyn, NY
Abstract:Cyberphysical Systems (CPS) are very large networks in which collaborating agents possessing sensing, communication and computation capabilities are interconnected for controlling physical entities. Applications are ubiquitous in sensor networks, robotics, transportation, and smart grids. In this talk, I will present distributed, asynchronous and real-time algorithms for CPS, and illustrate applications in transportation, robotics and cyber security of CPS. In specific:
a) Distributed optimization: We propose and analyze novel algorithms for multi-agent optimization over networks that feature favorable properties in terms of convergence rate, scalability and robustness. We devise two classes of general-purpose methods : a) distributed asynchronous linear system solvers and b) a new block-coordinate operator splitting method that can handle a wide range of problems in multi-agent systems, signal processing and machine learning. We exhibit our methods in a variety of applications in Network Utility Maximization, Data Visualization, Distributed Model Predictive Control and GPS-free multi-agent localization.
b) Cyber Security: We establish fundamental asymptotic bounds on the security of distributed protocols to collusion attacks. Our analysis enacts an encouraging result, in that the number of attackers that can be tolerated in large-scale CPS is ‘almost linear’ in the number of benign agents. Furthermore, we propose a theme for performing computations directly on encrypted data in a distributed fashion, and discuss its implications in the realm of secure cloud computing.
c) Travel time estimation: We propose and analyze a method for performing compressed sensing on an infinite data stream. Our protocol involves a) encoding, via compressively sampling sliding windows of the data stream, and b) decoding, by means of solving LASSO using a newly developed quasi-Newton proximal method with accelerated convergence rates. We apply our framework to the problem of sparse kernel density estimation, and delineate its advantages for adaptively learning travel time distributions in transportation networks in real-time.
Keywords
Cyberphysical systems, Big data, Distributed optimization, Clock synchronization, Localization, Compressed Sensing, Cyber Security, Cloud Computing
Research areas
Optimization, Control, Machine Learning, Signal Processing, Cyber Security
Applications
Data Mining / Machine Learning, Transportation, Robotics, Sensor Networks
About the Speaker: Nick Freris is an assistant professor of Electrical and Computer Engineering at New York University Abu Dhabi (NYUAD), and a Global Network Assistant Professor at New York University Tandon School of Engineering. He is the director of Cyberphysical Systems Laboratory (CPSLab) at NYUAD, and a member of the Center for Cyber Security (CCS).
He received the Diploma in Electrical and Computer Engineering from the National Technical University of Athens (NTUA), Greece in 2005, and the M.S. degree in Electrical and Computer Engineering, the M.S. degree in Mathematics, and the Ph.D. degree in Electrical and Computer Engineering all from the University of Illinois at Urbana-Champaign in 2007, 2008, and 2010, respectively. Dr. Freris’s research interests lie in the area of cyberphysical systems: distributed estimation, optimization and control, data mining/machine learning, cyber security, and applications in transportation, sensor networks and robotics.
His research is sponsored by the National Science Foundation and was recognized with the 2014 IBM High Value Patent award, two IBM invention achievement awards and the Gerondelis foundation award. Dr. Freris served as plenary speaker at the 4th International Conference on Big Data Analysis and Data Mining, Sep. 2017, Paris. Previously, Dr. Freris was a senior researcher in the School of Computer and Communication Sciences at École Polytechnique Fédérale de Lausanne (EPFL), Switzerland, from 2012-2014, and a postdoctoral researcher in IBM Research – Zurich, Switzerland, from 2010-2012. Dr. Freris is a senior member of IEEE, and a member of SIAM and ACM.