Virtual Seminar by Yu Wang

Title: Learning Scheduling and Optimization in Federated Edge Learning

Time and Date: Tuesday, Nov. 14, 2023, 9:00AM US Eastern Time (New York Time)

Presenter: Dr. Yu Wang, Professor in the Department of Computer and Information Sciences, Temple University, Philadelphia, Pennsylvania 19122, USA

Venue: https://zoom.us/j/9172542706 (Password: 4Zn7xZ)

Abstract: Edge computing and federated learning (FL) have gained popularity since they provide a promising edge learning framework that mitigates the limitations of long latency, high cost, and privacy concerns in cloud-based centralized learning. While most existing works on federated edge learning focus on optimizing the training of the global model in edge systems, the concurrent training of multiple FL models from different applications in a shared edge cloud can lead to edge resource competition and affect the training performance of each model. Hence, in this talk, I will present our recent works in addressing this challenge by proposing optimization algorithms to jointly select FL participants and learning rates or topologies for each model, with an aim of minimizing the total training cost. Particularly, I will first introduce a multi-stage optimization framework that allows FL models to select their participants and learning rates or learning topologies. Then, I will describe a quantum assisted algorithm to tackle the joint participant selection and learning scheduling problem using both quantum and classical computing. Finally, I will summarize the talk with discussions about future directions in federated edge learning.

Bio: Dr. Yu Wang is currently a Professor in the Department of Computer and Information Sciences at Temple University. He holds a Ph.D. from Illinois Institute of Technology, an MEng and a BEng from Tsinghua University, all in Computer Science. His research interest includes wireless networks, smart sensing, and mobile computing. He has published over 200 papers in peer reviewed journals and conferences. He has served as general chair, program chair, program committee member, etc. for many international conferences (such as IEEE MASS, IEEE IPCCC, ACM MobiHoc, IEEE INFOCOM, IEEE GLOBECOM, IEEE ICC), and has served as Editorial Board Member for several international journals (including IEEE Transactions on Parallel and Distributed Systems and IEEE Transactions on Cloud Computing). He is a recipient of Ralph E. Powe Junior Faculty Enhancement Awards from Oak Ridge Associated Universities (2006), Outstanding Faculty Research Award from College of Computing and Informatics at the University of North Carolina at Charlotte (2008), Fellow of IEEE (2018), and ACM Distinguished Member (2020).

Virtual Seminar by Nan Cheng

Title: RingSFL: An Adaptive Split Federated Learning Towards Taming Client Heterogeneity Date and Time: Sept. 1, 2023, 9:00 am – 10:00 am US Eastern Time (New York Time) Zoom Link: https://us04web.zoom.us/j/76618091026?pwd=bSR2dziIyZUS0aWF2R666BJEl4pMS5.1 Abstract: Federated learning (FL) has gained increasing attention due to its ability to collaboratively train while protecting client data privacy. However, in most wireless networks, such as vehicular networks, client heterogeneity exists, as different vehicular users have different computation capabilities. Vanilla FL cannot handle client heterogeneity, leading to a degradation in training efficiency due to stragglers, and is still vulnerable to privacy leakage. To address these issues, we propose RingSFL, a novel distributed learning scheme that integrates FL with a model split mechanism to adapt to client heterogeneity while maintaining data privacy. In RingSFL, all clients form a ring topology. For each client, instead of training the model locally, the model is split and trained among all clients along the ring through a pre-defined direction. By properly setting the propagation lengths of heterogeneous clients, the straggler effect is mitigated, and the training efficiency of the system is significantly enhanced. Additionally, since the local models is blended, it is less likely for an eavesdropper to obtain the complete model and recover the raw data, thus improving data privacy. The experimental results on both simulation and prototype systems show that RingSFL can achieve better convergence performance than benchmark methods on independently identically distributed (IID) and non-IID datasets, while effectively preventing eavesdroppers from recovering training data. Bio: Nan Cheng received the Ph.D. degree from the Department of Electrical and Computer Engineering, University of Waterloo in 2016, and B.E. degree and the M.S. degree from the Department of Electronics and Information Engineering, Tongji University, Shanghai, China, in 2009 and 2012, respectively. He worked as a Post-doctoral fellow with the Department of Electrical and Computer Engineering, University of Toronto, from 2017 to 2019. He is currently a professor with State Key Lab. of ISN and with School of Telecommunications Engineering, Xidian University, Shaanxi, China. He has published over 90 journal papers in IEEE Transactions and other top journals. He serves as associate editors for IEEE Internet of Things Journal, IEEE Transactions on Vehicular Technology, IEEE Open Journal of the Communications Society, and Peer-to-Peer Networking and Applications, and serves/served as guest editors for several journals. His current research focuses on B5G/6G, vehicular networks, AI-driven future networks, and space-air-ground integrated network. About the Virtual Seminar Series: The seminar series is organized by the IEEE TCCN Special Interest Group on AI empowered Internet of Vehicles.

Virtual Seminar by Dusit Niyato

Title: Unleashing the Power of Mobile Edge-Cloud Generative AI Services and AIGC Networks

Time and Date: Monday, Aug. 21, 2023, 9:00am US Eastern Time (New York Time)

Presenter: Dr. Dusit Niyato, President’s Chair Professor in Computer Science and Engineering in the School of Computer Science and Engineering, Nanyang Technological University, Singapore

Venue: https://zoom.us/j/9172542706 (Password: 4Zn7xZ)

Abstract: Artificial Intelligence Generated Content (AIGC) is an automated method for generating, manipulating, and modifying valuable and diverse data using AI algorithms creatively. This talk aims to on the deployment of AIGC applications, e.g., ChatGPT, at mobile edge networks, i.e., mobile AIGC networks that provide personalized and customized AIGC services in real-time while maintaining user privacy. We begin by introducing the background and fundamentals of generative models and the lifecycle of AIGC services at mobile AIGC networks, which includes data collection, training, fine-tuning, inference, and product management. We then discuss the collaborative cloud-edge-mobile infrastructure and technologies required to support AIGC services and enable users to access AIGC at mobile edge networks. Furthermore, we explore AIGC-driven creative applications and use cases for mobile AIGC networks. We highlight some future research directions and open issues for the full realization of mobile AIGC networks.

Bio:Dusit Niyato is currently a President’s Chair Professor in Computer Science and Engineering in the School of Computer Science and Engineering, Nanyang Technological University, Singapore. He received B.E. from King Mongkuk’s Institute of Technology Ladkrabang (KMITL), Thailand in 1999 and Ph.D. in Electrical and Computer Engineering from the University of Manitoba, Canada in 2008. Dusit’s research interests are in the areas of distributed collaborative machine learning, Internet of Things (IoT), edge intelligent metaverse, mobile and distributed computing, and wireless networks. Dusit won the IEEE Communications Society (ComSoc) Best Survey Paper Award, IEEE Asia-Pacific Board (APB) Outstanding Paper Award, the IEEE Computer Society Middle Career Researcher Award for Excellence in Scalable Computing in and Distinguished Technical Achievement Recognition Award of IEEE ComSoc Technical Committee on Green Communications and Computing. Dusit also won a number of best paper awards including IEEE Wireless Communications and Networking Conference (WCNC), IEEE International Conference on Communications (ICC), IEEE ComSoc Communication Systems Integration and Modelling Technical Committee and IEEE ComSoc Signal Processing and Computing for Communications Technical Committee 2021. Currently, Dusit is serving as Editor-in-Chief of IEEE Communications Surveys and Tutorials, an area editor of IEEE Transactions on Vehicular Technology, editor of IEEE Transactions on Wireless Communications, associate editor of IEEE Internet of Things Journal, IEEE Transactions on Mobile Computing, IEEE Wireless Communications, IEEE Network, IEEE Transactions on Information Forensics and Security (TIFS), and ACM Computing Surveys. He was a guest editor of IEEE Journal on Selected Areas on Communications. He was a Distinguished Lecturer of the IEEE Communications Society for 2016-2017. He was named the 2017-2022 highly cited researcher in computer science. He is a Fellow of IEEE and a Fellow of IET.

Virtual Seminar by Jie Wu

Title: On Optimal Partitioning and Scheduling of DNNs in Mobile Edge/Cloud Computing

Time and Date: Friday, July 21, 2023, 9:00am US Eastern Time (New York Time)

Presenter: Dr. Jie Wu, Director of the Center for Networked Computing and Laura H. Carnell professor, Temple University, USA

Venue: https://zoom.us/j/9172542706 (Password: 4Zn7xZ)

Abstract: As Deep Neural Networks (DNNs) have been widely used in various applications, including computer vision on image segmentation and recognition, it is important to reduce the makespan of DNN inference computation, especially when running on mobile devices. Offloading is a viable solution that offloads computation from a slow mobile device to a fast, but remote edge/cloud. As DNN computation consists of a multiple-stage processing pipeline, it is critical to decide on what stage should offloading occur to minimize the makespan. Our observations show that the local computation time on a mobile device follows a linear increasing function, while the offloading time on a mobile device is monotonic decreasing and follows a convex curve as more DNN layers are computed in the mobile device. Based on this observation, we first study the optimal partition and scheduling for one line-structure DNN. Then, we extend the result to multiple line-structure DNNs. Heuristic results for general-structure DNNs, represented by Directed Acyclic Graphs (DAGs), are also elaborated based on a path-based scheduling policy. Extensions to DNN training are also discussed.

Bio: Jie Wu is the Director of the Center for Networked Computing and Laura H. Carnell professor at Temple University. He served as Chair of Department of Computer and Information Sciences from the summer of 2009 to the summer of 2016 and Associate Vice Provost for International Affairs from the fall of 2015 to the summer of 2017. Prior to joining Temple University, he was a program director at the National Science Foundation and was a distinguished professor at Florida Atlantic University where he received his PhD in 1989. His current research interests include mobile computing and wireless networks, routing protocols, network trust and security, distributed algorithms, applied machine learning, and cloud computing. Dr. Wu regularly publishes in scholarly journals, conference proceedings, and books. He serves on several editorial boards, including IEEE Transactions on Service Computing and Journal of Computer Science and Technology. Dr. Wu is/was general chair/co-chair for IEEE DCOSS’09, IEEE ICDCS’13, ICPP’16, IEEE CNS’16, WiOpt’21, ICDCN’22, IEEE IPDPS’23, and ACM MobiHoc’23 as well as program chair/cochair for IEEE MASS’04, IEEE INFOCOM’11, CCF CNCC’13, and ICCCN’20. He was an IEEE Computer Society Distinguished Visitor, ACM Distinguished Speaker, and Chair for the IEEE Technical Committee on Distributed Processing (TCDP). Dr. Wu is a Fellow of the AAAS and a Fellow of the IEEE. He is a Member of Academia Europaea (MAE).

Virtual Seminar by Nada Golmie

Title: On Communication and Sensing Measurements and Modeling for Next “G”

Time and Date: Thursday, June 15, 2023, 9:00am US Eastern Time (New York Time)

Presenter: Dr. Nada T. Golmie, National Institute of Standards and Technology, USA

Venue: https://zoom.us/j/9172542706 (Password: 4Zn7xZ)

Abstract: As the needs for sensing the physical world to support detection, tracking, AR/VR and slew of other applications increase, the use of communications waveforms for sensing is attracting a lot of attention and is likely to emerge as one of the main features in NextG development.
In this talk we discuss key challenges in measurement science and modeling approaches for advanced communications and sensing going forward. We will review the main building blocks for joint communications and sensing in terms of efficient spectrum use, higher frequency bands and the use machine learning. The millimeter-wave and terahertz bands hold the promise of significant bandwidth and speed due to large swaths of untapped spectrum. In addition, as massive data volumes are being collected, analyzed, and delivered, communications systems have become too complex to develop, manage, and operate. The insights that are “mined” from the data using Machine Learning (ML) techniques have become standard practice. In this talk, we discuss state -of -the art and key challenges in measurement and modeling techniques to expedite the development and pave the way for the next “G”.

Bio: Nada Golmie (nada@nist.gov) received her Ph.D. in computer science from the University of Maryland at College Park. Since 1993, she has been a research engineer at the National Institute of Standards and Technology. From 2014 to 2022, she served as the chief for the Wireless Networks Division in the Communications Technology Laboratory. She is an IEEE and a NIST Fellow. Her research in media access control and protocols for wireless networks led to over 200 technical papers presented at professional conferences, journals, and contributed to international standard organizations and industry led consortia. She is the author of “Coexistence in Wireless Networks: Challenges and System-level Solutions in the Unlicensed Bands,” published by Cambridge University Press (2006). She leads several projects related to the modeling and evaluation of future generation wireless systems and protocols and serves as the NextG Channel Model Alliance chair.

Virtual Seminar by Arup Bhuyan

Title: 5G and Future G Wireless Security

Date and Time: January 25, 2023 at 11 AM ET

Registration Process: Please register at https://albany.zoom.us/meeting/register/tJctceysqjsoHdEotovMh872sxFBObm4gOsD

Abstract: Monitoring and detecting abnormalities in the 5G and Future FG wireless networks are a necessity for their secure use. Spectrum sensing is also a key element in effective and secure use of the unlicensed and shared spectrum bands that are becoming available to fully harness the power of the transformative use cases that 5G is making possible. In this talk we discuss the primary attack surfaces in 5G and how machine learning can play a major role towards secure use of 5G and Future G technologies. We conclude with an innovative and distributed spectrum sharing solution that utilizes reinforcement learning.

Bio: Dr. Arupjyoti (Arup) Bhuyan is the Technical Director of the Wireless Security Institute in the Idaho National Laboratory. The focus of his research is on secure implementation of future generations of wireless communications with scientific exploration and engineering innovations across the fields of wireless technology, cybersecurity, and computational science. Specific goals are to lead and focus wireless security research efforts for 5G and beyond with national impact, to secure communications for a nationwide unmanned aerial system and for 5G spectrum sharing with distributed scheduling. Arup has extensive industry experience in wireless communications from his work before he joined INL in October 2015. He received his Ph.D. in Engineering and Applied Sciences from Yale University. He is a senior member of IEEE.

About the Monthly Virtual Seminar Series:

The IEEE TCCN Security Special Interest Group conducts a monthly virtual seminar series to highlight the challenges in securing the next generation (xG) wireless networks. The talks will feature cutting edge research addressing both technical and policy issues to advance the state-of-the-art in security techniques, architectures, and algorithms for wireless communications.

Virtual Seminar by Tim O’Shea

Title: Securing and Optimizing Wireless Systems with AI-Native & Data-Driven Wireless Signal Processing in the Physical Layer

Date and Time: December 13, 2022 at 10 AM ET

Registration Process: Please register at https://tinyurl.com/38s8vxuc

Abstract: This talk will provide an overview of ways in which machine learning and data-driven signal processing are impacting wireless sensing systems, spectrum anomaly-detection and analytic systems, and leading to powerful spectrum awareness capabilities which help to provide security, monitoring, and adaptation tools never previously possible at scale in wireless systems. We’ll talk about some of the work we’ve done at both DeepSig and VT to enable this, highlighting several works which illustrate how AI/ML based edge processing provide valuable security tools which fill significant gaps in the wireless threat surface which were previously prohibitive or highly manual and labor intensive to address. We’ll show some examples of how these approaches can be rapidly adopted from applied research into fieldable commercial systems, and considering their potential impact in securing and optimizing future wireless systems. Finally, we’ll also consider the impact of data driven baseband processing for wireless transmission and reception – The concept of an AI-Native Air Interface – often in a form resembling a channel autoencoder, presents a powerful tool for optimizing physical layer waveforms for hardening against a wide range of wireless threats and failure modes. We’ll highlight some of our work as well as others in this area, and consider where this can help optimize wireless security and resilience in communications systems in the future. Finally, we’ll consider future areas for research, industry adoption of techniques, and potentially transformative wireless security trends going forward.

Bio: Tim O’Shea is the CTO and Co-Founder at DeepSig Inc and a Research Assistant Professor at Virginia Tech in Arlington, VA. He is focused on building machine learning and AI-Native wireless baseband processing capabilities to enhance the spectral and energy performance of 5G, 5G-Advanced, and 6G wireless air interfaces, and leveraging AI-Driven spectral and spatial channel awareness and sensing to optimize multi-user and multi-access next generation wireless systems. Previously he worked with wireless startups Hawkeye 360 and Federated Wireless in seed stage and held engineering R&D positions with both the US DOD and with Cisco Systems. He is the author of over 50 peer reviewed works and patents in the machine learning for communications space, and is involved in IEEE COMSOC, IEEE MLC ETI, Next-G Alliance, and OpenRAN efforts to accelerate AI driven communications systems and their adoption within next generation RAN and Open-vRAN.

About the Monthly Virtual Seminar Series:

Virtual Seminar by Bhaskar Krishnamachari

Title: Blockchain Technology and its applications to the Internet of Things

Date and Time: November 11, 2022 at 11AM ET

Registration Process: Please register at https://tinyurl.com/y9ebze79

Abstract: Blockchain technology is bringing fundamental new capabilities pertaining to decentralized trust and enabling micropayments for data. I will present results from research at USC touching on both the core technology and its application to the internet of things. These include a new mobile-oriented blockchain protocol, a middleware to decentralize publish-subscribe brokers, smart contracts to enable cheat-proof peer-to-peer trading of digital goods, a streaming data payment protocol, a decentralized review mechanism suitable for data marketplaces, as well as applications of blockchain to network security.

Bio: Bhaskar Krishnamachari is a Professor of Electrical and Computer Engineering, and Director of the Center for Cyber-Physical Systems and the Internet of Things at the USC Viterbi School of Engineering. His research spans the design and evaluation of algorithms and protocols for wireless networks, distributed systems, and the internet of things. He is the co-author of more than 300 technical papers, and 3 books, that have been collectively cited more than 30,000 times. He has received best paper awards at IPSN (2004, 2010), Mobicom (2010), and VNC (2021). He is the recipient of the NSF Career Award, the ASEE Terman Award, and has been listed in MIT Technology Review’s TR-35 and Popular Science magazine’s “Brilliant 10”.

About the Monthly Virtual Seminar Series:

Virtual Seminar by Honggang Zhang

Title: Rethinking Modern Communication from Semantic Coding to Semantic Communication

Time and Date: Thursday, Dec. 1, 2022 at 9:00AM US Eastern Standard Time

Presenter: Professor Honggang Zhang, Zhejiang University, China

Venue: https://zoom.us/j/9172542706 (Password: 4Zn7xZ)

Abstract: Modern communications are usually designed to pursue a higher bit-level precision and fewer bits while transmitting a message. We give a rethinking on these two fundamental features and introduce the concept and advantage of semantics that take advantage of a new kind of semantics-aware communication framework, incorporating both the semantic encoding and the semantic communication issues. After analyzing the underlying defects of existing semantics-aware techniques, we establish a confidence-based distillation mechanism for the semantic coding problem and put forward a reinforcement learning (RL)-powered semantic communication solution, which endow a communications system the ability to convey the semantics instead of pursuing the bit level accuracy. On top of these technical contributions, this talk provides a new insight to understand how the semantics are processed and represented in a semantics-aware coding and communication system, and verifies the significant benefits of doing so. Targeting at the next-generation (6G) semantics-aware communication, a number of critical concerns and open challenges such as the information overhead, semantic security and implementation cost are also discussed and envisioned.

Bio: Dr. Honggang Zhang is the Chief Managing Editor of Intelligent Computing Journal, a Science Partner Journal jointly established by AAAS (American Association for the Advancement of Science) and Zhejiang Lab (https://spj.sciencemag.org/journals/icomputing/). He is a Full Professor of College of Information Science and Electronic Engineering, Zhejiang University, China. He was an International Chair Professor, CominLabs Excellence Center, Université Européenne de Bretagne (UEB) & Supélec, France (2012-2014). He was an Honorary Visiting Professor of the University of York, UK (2010-2018).

Virtual Seminar by Lingyang Song

Title: Holographic Radio: A New Paradigm for Ultra-Massive MIMO

Time and Date: July 22, 2022, at 9:00AM ET

Presenter: Professor Lingyang Song, Peking University, China

Venue: https://zoom.us/j/9172542706 (Password: 4Zn7xZ)

Abstract: To enable a ubiquitous intelligent information network, the forthcoming sixth generation (6G) wireless communications are expected to provide revolutionary mobile connectivity and high-throughput data services through ultra-massive multiple input multiple output (MIMO). Widely-utilized phased arrays relying on costly components make the implementation of ultra-massive MIMO in practice become prohibitive from both cost and power consumption perspectives. The recent developed reconfigurable holographic surfaces (RHSs) composing of densely packing sub-wavelength metamaterial elements can achieve holographic beamforming without costly hardware components. By leveraging the holographic principle, the RHS serves as an ultra-thin and lightweight surface antenna integrated with the transceiver, thereby providing a promising alternative to phased arrays. In this talk, we will first provide a basic introduction of RHSs. We then introduce the unique features of RHSs which enables ultra-massive MIMO for both communication and sensing, in a comprehensive way. Related design, analysis, optimization, and signal processing techniques will be presented along with typical RHS-based applications for wireless communications. The implementation issues along with our developed prototypes and experiments will also be discussed.

Bio: Lingyang Song (S’03-M’06-SM’12-F’19) received his PhD from the University of York, UK, in 2007, where he received the K. M. Stott Prize for excellent research. He worked as a research fellow at the University of Oslo, Norway until rejoining Philips Research UK in March 2008. In May 2009, he joined the School of Electronics Engineering and Computer Science, Peking University, and is now a Boya Distinguished Professor. His main research interests include wireless communications, mobile computing, and machine learning. Dr. Song is the co-author of many awards, including IEEE Leonard G. Abraham Prize in 2016, IEEE ICC 2014, IEEE ICC 2015, IEEE Globecom 2014, and the best demo award in the ACM Mobihoc 2015. He received National Science Fund for Distinguished Young Scholars in 2017, First Prize in Nature Science Award of Ministry of Education of China in 2017. Dr. Song has served as a IEEE ComSoc Distinguished Lecturer (2015-2018), an Area Editor of IEEE Transactions on Vehicular Technology (2019-), Co-chair of IEEE Communications Society Asia Pacific Board Technical Affairs Committee (2020-). He is a Clarivate Analytics Highly Cited Researcher.