Today, the field of machine intelligence unfolds at exponential rates. As the amounts and variety of data continue to increase, corresponding computational workloads challenge the scalability and performance of the existing network infrastructure. The problem escalates further for streaming data, which may quickly expire and, thus, have to be processed in near-real-time. To handle that, machine learning (ML) services have started to gradually shift from the cloud closer to where the data are generated - to the edge and further on to densely deployed and mobile consumer devices. In reality, however, this distributed approach faces challenges associated with an extreme diversity of client hardware - ranging from wearable sensors to high-end XR gear - and, hence, with the heterogeneity of on-device resources and unbalanced data. The research community has paid immense attention to exploring the field of distributed ML - with particular emphasis on heterogeneity-aware and privacy-preserving methods developed within the recently shaped area of federated learning. In supporting federated and, generally, distributed learning, a fundamental role is played by communications and wireless connectivity. This talk discusses recent advances of distributed learning deployed over wireless networks and its new challenges that result from the extreme resource heterogeneity, the need to preserve user privacy, and the presence of constrained and dynamic wireless resources.
Dr. Olga Galinina received her Ph.D. degree from the Department of Electronics and Communications Engineering at Tampere University of Technology, Finland. She previously received her B.Sc. and M.Sc. degrees in Applied Mathematics from the Department of Applied Mathematics, Faculty of Mechanics and Physics, Saint-Petersburg State Polytechnical University, Russia. She has publications on mathematical problems in the novel telecommunication protocols in internationally recognized journals and high-level peer-reviewed conferences. Her research interests include queueing theory, stochastic processes, optimization theory and their applications; heterogeneous wireless networking, machine-to-machine, wearable and device-to-device communications.
Complex systems recently attracted worldwide attention due to the Nobel Prize in Physics being awarded to research in that area. We will discuss how large-scale future networks' automation can be realised via communication and computing functionalities, aided by complex systems- and AI-based design. Complex systems theory and tools can help us address the issue of scalability, while AI can be used to predict changes in such systems experiencing high dynamics, and whose optimization in a traditional way would take too much time due to the network scale. We will discuss general principles that can be applied to the analysis and design future networks in terms of: (i) the promise of prediction and the problem of scaling, (ii) how to distribute network tasks to address the scaling issue, and (iii) the role of time and the power of prediction. We will discuss such general principles relying on recent research spanning mm-wave systems, massive MIMO, in-network computing, and public health applications.
Nicola Marchetti is Associate Professor in Wireless Communications at Trinity College Dublin, Ireland. He performs his research under the Irish Research Centre for Future Networks and Communications (CONNECT), where he leads the Wireless Engineering and Complexity Science (WhyCOM) lab. He received the PhD in Wireless Communications from Aalborg University, Denmark in 2007, the M.Sc. in Electronic Engineering from University of Ferrara, Italy in 2003, and the M.Sc. in Mathematics from Aalborg University in 2010. His research interests include AI-enabled Networks, Communications for Biology, Complexity and Self-Organization, MAC Protocols and Resource Allocation, Signal Processing for Communications, and Quantum Communications. He has authored more than 150 journals and conference papers, 2 books and 8 book chapters, holds 4 patents, and received 4 best paper awards.
The MSCA Postdoctoral Fellowships (PFs) target researchers holding a PhD who wish to carry out their research activities abroad, acquire new skills and develop their careers. PFs help researchers gain experience in other countries, disciplines and non-academic sectors. However, most of the recent PhD graduates do not have strong experience in preparing project proposals for competitive programmes. This talk is devoted to introducing the MSCA-PFs as well as to providing some hints in preparing a strong proposal for this action.
Joaquín Torres-Sospedra holds a Bachelor’s Degree in Computer Science (2003) and a PhD with distinction (2011), both from Universitat Jaume I. His main background is on Artificial Neural Networks and Ensemble Learning. In 2013, he joined GEOTEC where he is currently leading the indoor positioning and navigation research line. His current interests are on Smartphone-based indoor positioning, Wi-Fi fingerprinting and Bluetooth Low Energy deployment. He is currently chairing the IPIN-ISC and he is really active on evaluating Indoor Positioning systems through competitions in real-life environments.
Time is tricky. And essential. And busy. Let's talk more about how to utilize time and discover ways how to work with it. Levi Poelmans, Kyndryl CZ 1st Line Manager, brings you insight in Time Management.
Levi Poelmans joined IBM in 2016 and started as a Customer Service Representative. During his 5.5 year career in IBM/Kyndryl, he fulfilled multiple roles in Account Management (Service Coordinator, Account Trainer, Delivery Leader), as well as the role of PHC Coordinator in Server Management. Currently, he has been working as a First Line Manager for SDDC Networking Team.
Problem solving is the process of identifying a problem, developing possible solution paths, and taking the appropriate course of action. How to do it? Martina Kökény, Kyndryl Client Center CZ EAM GCIC Brno Location Leader & Enterprise Asset Management Leader, will tell you more how to do it.
Martina has been in IBM since 2005 and started his career in Human Resources later on in Project Management. She became manager in 2009 of the newly created European team. She played critical role in 2 major transformations in her organization while implementing Global Delivery Framework and later Agile. She went trough various leadership roles both local and international. In current role, she has responsibility of 150 Asset Management Specialists who serve customers globally and France Asset Management Leader role.