ABSTRACT: The convergence of artificial intelligence (AI) and smart sensor technologies is revolutionizing healthcare by enabling real-time monitoring, personalized health interventions, and ambient assisted living. In this talk, I will present a comprehensive overview of AI-enabled smart systems across the domains of mobility analysis, smart medical homes, wearable telehealth platforms, and lifestyle management tools. I will then discuss a holistic view of a smart medical home ecosystem driven by sensor fusion, edge computing, and AI analytics that form the core of next-generation aging-in-place and daily activity monitoring. Central to this is the concept of a "brain" or and autonomic decision-making system that orchestrates data flow, interprets contextual information, and delivers intelligent outputs. A key focus of the talk will be our mobility and walking pattern analyzer, a multi-sensor AI-based system that assesses gait and mobility patterns in real-time, empowering early gait diagnostics and rehabilitation tracking. I will also introduce the smart living diary, a multi-modal interface aggregating activity, nutrition, and sleep data to promote healthy lifestyle. Next, I will highlight the growing role of wearable telehealth devices and how AI augments their predictive and diagnostic capabilities. Trends in ubiquitous healthcare and sensor performance will be discussed, alongside ethical and interoperability challenges. The presentation will conclude with a forward-looking perspective on the future of AI-enabled healthcare, emphasizing the need for standardized frameworks, inclusive design, robust privacy-preserving mechanisms and future pathways for research and implementation in smart healthcare systems.

BIO: Dr. M. Jamal Deen is Distinguished University Professor and Director of the Micro- and Nano-Systems Laboratory at McMaster University. His current research interests are nanoelectronics, optoelectronics, nanotechnology, data analytics and their emerging applications to health and environmental sciences. As an educator, he won the Ham Education Medal from IEEE Canada, the McMaster University President’s Award for Excellence in Graduate Supervision, and the MSU Macademics’ Lifetime Achievement Award (highest award at McMaster University voted by the students) for his exceptional dedication to teaching and significant contribution to student life, academia, and the community at large. Recently (2024), he was the inaugural winner of the SM Sze Education Award from the IEEE Electron Devices Society “For impact leadership and global dissemination of biosensor education in underprivileged regions.”

As an undergraduate student at the University of Guyana, Dr. Deen was the top ranked mathematics and physics student and the second ranked student at the university, winning the Chancellor’s gold medal and the Irving Adler prize. As a graduate student, he was a Fulbright-Laspau Scholar and an American Vacuum Society Scholar. His awards and honors include the Callinan Award as well as the Electronics and Photonics Award from the Electrochemical Society; a Humboldt Research Award from the Alexander von Humboldt Foundation; the Eadie Medal from the Royal Society of Canada; and the McNaughton Gold Medal, the Fessenden Medal and the Gotlieb Computer Medal, all from IEEE Canada. In addition, he was awarded the five honorary doctorate degrees in recognition of his exceptional research, scholarly and education accomplishments, exemplary professionalism and valued services.

Dr. Deen has been elected by his peers as Fellow/Academician of thirteen national academies and professional societies including The Royal Society of Canada - The Academies of Arts, Humanities and Sciences (the highest honor for academics, scholars and artists in Canada), the Chinese Academy of Sciences (China’s highest national honor in the area of science and technology and highest academic title), , the National Academy of Sciences India, the Canadian Academy of Engineering, IEEE, APS (American Physical Society) and ECS (Electrochemical Society). He served as the elected President of the Academy of Science, The Royal Society of Canada in 2015-2017. Recently, he was elected the inaugural Vice President (North) of The World Academy of Sciences, representing the developed countries. He was also elected to the Order of Canada, the highest civilian honor awarded by the Government of Canada.

ABSTRACT: Edge computing systems must offer low latency at low cost and low power consumption for sensors and other applications, including the IoT, smart vehicles, smart homes, and 6G. Thus, substantial research has been conducted to identify optimum task allocation schemes in this context using non-linear optimization, machine learning, and market-based algorithms. Prior work has mainly focused on two methodologies: (i) formulating non-linear optimizations that lead to NP-hard problems, which are processed via heuristicss, and (ii) using AI-based formulations, such as reinforcement learning, that are then tested with simulations. These prior approaches have two shortcomings: (a) there is no guarantee that optimum solutions are achieved, and (b) they do not provide an explicit formula for the fraction of tasks that should be allocated to the different servers to achieve a specified optimum. In this talk we will present a different mathematically based principled method that explicitly computes the optimum fraction of jobs that should be allocated to the different servers to (1) minimize the average latency (delay) of the jobs that are allocated to the edge servers and (2) minimize the average energy consumption of these jobs at the set of edge servers. This approach has low computational cost and provides simple linear complexity formulas that achieve minimum latency and minimum energy consumption. We will also show our experiments where reinforcement learning with low computational complexity is used to reduce energy consumption against a small increase in average response time for the jobs.

BIO: Erol Gelenbe (FIEEE, FACM) received the B.S.E.E. degree from METU, Turkey, the M.S.E.E. and Ph.D. degrees from the Polytechnic Institute of NYU, and the D.Sc. degree in Mathematical Sciences from Sorbonne University. He pioneered system performance evaluation methods and invented the random neural network and G-Networks. He built research teams, helped to develop commercial tools for system modeling and simulation such as the queueing network analysis package QNAP and the manufacturing simulator FLEXSIM, designed system architectures such as the many-to-many communication switch Sycomore for Thomson CSF, and graduated over 90 Ph.D. students, including 24 women. He is currently a Professor at the Institute of Theoretical and Applied Informatics, Polish Academy of Sciences, a visiting faculty at King’s College London, and an associated researcher at CNRS I3S, Université Côte d’Azur (Nice). Previously, he held chaired professorships at Imperial College, the University of Central Florida, Duke University, Université Paris-Descartes, Paris-Saclay University, and Université de Liège. He was a Principal Investigator of numerous European Union research projects and a Coordinator of FP7 NEMESYS and H2020 SerIoT, and is currently funded by the H2020 DOSS project. His research was funded in the U.S. by NSF, ONR, ARO and industry, and in the UK by UKRI, MoD and also industry. He is also a Fellow of the RSS, AAIA and IET, and an Elected Fellow of the French National Academy of Technologies, the Science Academies of Poland, Turkey, the Royal Academy of Belgium, and an Honorary Fellow of the Hungarian Academy of Sciences and the Islamic World Academy of Sciences (Amman, Jordan). He is also a Member of Academia Europaea and the National Academy of AI (USA). His prizes include the Parlar Foundation Science Award (Turkey), the Grand Prix France Télécom, the ACM SIGMETRICS Life-Time Achievement Award, the IET Oliver Lodge Medal, and the Mustafa Prize. France awarded him the honors of Chevalier de la Lègion d’Honneur, Chevalier des Palmes Académiques, and Commandeur de l’Ordre du Mérite. He was also awarded Commander of the Order of the Crown of Belgium, Commendatore al Merito and Grande Ufficiale dell’Ordine della Stella by Italy, and Officer of the Order of Merit of Poland. He is ranked in the top 7% of the Stanford 2%. ScholarGPS ranks him No. 1 in Poland for all of Engineering and Computer Science, No.2 in Poland over all fields, and No. 2 in Electrical and Computer Engineering when Poland and the UK are grouped together.

ABSTRACT: While the success of deep learning hinges on its ability to process vast amounts of data, computing systems struggle to keep up with the unprecedented demand of ever-increasing data, leading researchers back to the notorious memory-wall problem. Unfortunately, the old paradigm or new technology we have, such as parallel computing, multicore, or dataflow, are all designed for speedup computing. They are not designed to address the memory-wall problem. In fact, they have put even more pressure on the already overstressed memory systems. Data access has become the number one performance killer of computing. A paradigm shift is needed for computing from a data-centric point of view. In this talk, we introduce the concept of dataflow under the von Neumann machine to address this issue. We begin by presenting the C-AMAT model, which quantifies the benefits of concurrent data access and reveals the relationship between data locality and concurrency. Next, we introduce the LPM (Layer Performance Matching) framework to optimize memory system performance and formally introduce the concept of dataflow under the von Neumann machine. We then discuss our recent work in I/O systems, focusing on the Hermes multi-tiered I/O buffering system. Hermes optimizes data movement based on the LPM framework and has been a significant success. Finally, we will address some fundamental issues and present forward-thinking computer system designs for AI and big data applications. We will discuss the critical role of collaborative research infrastructures, like the envisioned StoreHub, in accelerating progress.

BIO: Dr. Xian-He Sun is a University Distinguished Professor, the Ron Hochsprung Endowed Chair of Computer Science, and the director of the Gnosis Research Center for accelerating data-driven discovery at the Illinois Institute of Technology (Illinois Tech). Before joining Illinois Tech, he worked at DoE Ames National Laboratory, at ICASE, NASA Langley Research Center, at Louisiana State University, Baton Rouge, and was an ASEE fellow at Navy Research Laboratories. Dr. Sun is an IEEE fellow and is known for his memory-bounded speedup model, also called Sun-Ni’s Law, for scalable computing. His research interests include high-performance data processing, memory and I/O systems, and performance evaluation and optimization. He has over 350 publications and 7 patents in these areas and is currently leading multiple large software development projects in high performance I/O systems. Dr. Sun is the Editor-in-Chief of the IEEE Transactions on Parallel and Distributed Systems, and a former department chair of the Computer Science Department at Illinois Tech. He received the Golden Core award from IEEE CS society in 2017, the ACM Karsten Schwan Best Paper Award from ACM HPDC in 2019, and the first prize best paper award from ACM/IEEE CCGrid in 2021. More information about Dr. Sun can be found on his web site www.cs.iit.edu/~sun/.

ABSTRACT: Modern distributed systems deal with uncertain scenarios, where environments, infrastructures, and applications are widely diverse. In the scope of IoT-Edge-Fog-Cloud computing, leveraging these neuroscience-inspired principles and mechanisms could aid in building more flexible solutions able to generalize over different environments. A captivating set of hypotheses from the field of neuroscience suggests that human and animal brain mechanisms result from a few powerful principles. If proved to be accurate, these assumptions could open a deep understanding of the way humans and animals manage to cope with the unpredictability of events and imagination. In this talk, we will explore how Active Inference mechanisms can be utilized for Distributed Intelligence in the Computing Continuum.

BIO: Schahram Dustdar is a Full Professor of Computer Science at the TU Wien, heading the Research Division of Distributed Systems, Austria and part-time ICREA research Professor at UPF Barcelona. He holds several honorary positions: University of California (USC) Los Angeles; Monash University in Melbourne, Shanghai University, Macquarie University in Sydney, and University Pompeu Fabra, Barcelona, Spain. From Dec 2016 until Jan 2017 he was a Visiting Professor at the University of Sevilla, Spain and from January until June 2017 he was a Visiting Professor at UC Berkeley, USA.

From 1999 – 2007, he worked as the co-founder and chief scientist of Caramba Labs Software AG in Vienna (acquired by ProjectNetWorld AG), a venture capital co-funded software company focused on software for collaborative processes in teams. He is the co-founder and chief scientist of Coovally.ai, an AI infrastructure company based in Barcelona.

He serves as Editor-in-Chief of Computing (Springer). Dustdar is the recipient of multiple awards: IEEE TCSVC Outstanding Leadership Award (2018), IEEE TCSC Award for Excellence in Scalable Computing (2019), ACM Distinguished Scientist (2009), ACM Distinguished Speaker (2021), IBM Faculty Award (2012). He is an elected member of the Academia Europaea: The Academy of Europe, as well as an IEEE Fellow (2016) and an Asia-Pacific Artificial Intelligence Association (AAIA) Fellow (2021) and was AAIA president (from 2020-2021).

ABSTRACT: The most kind of traditional DP (Differential Privacy) mechanisms (e.g. Laplace, Gaussian, etc.) have unlimited output range. In real scenarios, most datasets have bounded output range, e.g. age [0-150]. Users would then need to use post-processing or truncated mechanisms to forcibly bound output distribution. However, these mechanisms would incur bias problem which has been a long-known DP challenge, resulting in various unfairness issues in subsequent applications. A tremendous amount of research has been done on analyzing this bias problem and its consequences, but no solutions can solve it fully.

As the world first solution to solve this long-known DP bias problem, this talk will present a new innovative DP mechanism named Composite DP. It will first illustrate this long-known bias problem, and then detail the rational of the new mechanism and its example noise functions as well as their implementation algorithms. All source codes are publicly available on Github for any deployment or verification.

BIO: Dr Jinjun Chen is a Professor from Swinburne University of Technology, Australia. He holds a PhD in Information Technology from Swinburne University of Technology, Australia. His research interests include data privacy and security, cloud computing, scalable data processing, data systems and related various research topics. His research results have been published in more than 300 papers in international journals and conferences. He received various awards such as IEEE TCSC Award for Excellence in Scalable Computing and Australia’s Top Researchers. He has served as an Associate Editor for various journals such as ACM Computing Surveys, IEEE TC, TCC and TSUSC. He is a MAE (Academia Europea) and IEEE Fellow (IEEE Computer Society). He is Chair for IEEE TCSC (Technical Community for Scalable Computing).

ABSTRACT: The computational power required for training deep learning models has been skyrocketing in the past decade as models scale with big data, and has become a very expensive and scarce resource. Therefore, distributed training, which can leverage distributed available computational power, is vital for efficient large-scale model training. However, most previous distributed training frameworks like DDP and DeepSpeed are primarily designed for co-located clusters under homogeneous computing and communication conditions, and hence cannot account for geo-distributed clusters with both computing and communication heterogeneity. In this talk, I will introduce a new data parallel based distributed training framework called Co-Adaptive Data Parallelism (C-ADP). Specifically, we consider a data owner and parameter server that distributes data to and coordinates the collaborative learning across all the computing devices. We employ local training and delayed parameter synchronization to reduce communication costs. Moreover, I will formulate a data parallel scheduling optimization problem to minimize the training time by optimizing data distribution. To solve this problem, I will devise an efficient algorithm and formally prove that the obtained solution is optimal in the asymptotic sense. At the end, I will present experiment results demonstrating the efficiency and efficacy of C-ADP.

BIO: Dr. Pan Li is currently the Academic Vice President and Lixing Chair Professor at Hangzhou Dianzi University, Hangzhou, China. He received his Ph.D. degree in Electrical and Computer Engineering from University of Florida in August 2009. After that, he became a faculty member with Mississippi State University and then with Case Western Reserve University for twenty years. His research focuses on artificial intelligence, cybersecurity, and the interplay between them. He has served as an Editor for nine internationally renowned journals such as TMC, TBD, TWC, TNSE, and on the organizing committee and technical program committee for flagship conferences like AAAI, IJCAI, USENIX Security. His research has been supported by both federal agencies and industry companies. He was a recipient of NSF CAREER Award and an IEEE/AAIA/AIIA Fellow.

ABSTRACT: Green computing is essential for keeping our planet sustainable. Cyber-Physical Systems are prevalent in driving innovation across various sectors. Effective policies that reduce the energy consumption of Cyber-Physical Systems can significantly benefit our environment. This talk presents machine learning techniques for intelligent power management in Cyber-Physical Systems, focusing on designing methods to derive adaptable, energy-efficient power management policies for Linux-based systems. Key topics include DQN-based DVFS governor for systems with periodic workloads that have deadline constraints, interleaving learning for systems with limited computational resources, profiling and software-based power estimation, and adaptation to Quality of Experience (QoE).

BIO: Dr. Man Lin received her B.E. degree in Computer Science and Technology from Tsinghua University, Beijing, China, and her Ph.D. degree from Linköping University, Sweden. She is currently a professor and the chair of the Department of Computer Science, as well as a W.F. James Research Chair in Pure and Applied Science, at St. Francis Xavier University in Canada. Her research interests include Low-Power Computing, Cyber-physical Systems, Real-Time Scheduling, Machine Learning, System Analysis and Optimization Algorithms. Dr. Man Lin’s research has been funded by NSERC (National Science and Engineering Research Council of Canada) and CFI (Canadian Foundation of Innovation). She served as the chair of the IEEE CIS Smart World Technical Committee (2022-2022). She has been involved in several IEEE conference organizations and currently serves as an associate editor for IEEE Transactions on Sustainable Computing and the Elsevier Journal of Systems Architecture.