Honorary Professor and Teaching Fellow at the Department of Computer Science, University College London
Abstract:
AI has finally come of age and is astonishing in its capabilities. But even today, people overestimate AI and make use of it in ways that are foolish or even dangerous. In this talk Peter Bentley argues that the purpose of AI should never be to automate intelligence. AI should be augmented intelligence - helping us achieve our goals better by providing us with intelligent assistance. He provides examples of different types of augmented intelligence from his lab at University College London: using AI to augment the ability of a genetic algorithm to solve constraint problems and using AI to help designers search a wider space of generative architecture.
Biography:Peter J. Bentley Ph.D. B.Sc. (Hons) is an Honorary Professor and Teaching Fellow at the Department of Computer Science, University College London (UCL), Visiting Professor at Autodesk, Collaborating Professor at the Korean Advanced Institute for Science and Technology (KAIST), co-founder of Kazoova Ltd (Hubbub), a consultant on AI and a freelance writer. Previously he was a Visting Fellow at SIMTech, A*STAR, Singapore, Visiting Research Fellow at Goldsmiths College, London, Visiting Research Fellow at the University of Kent, Canterbury, CTO of two AI companies, and a contributing editor for WIRED UK. He achieved a B.Sc. (Hon's) degree in Computer Science (Artificial Intelligence) in 1993 and a Ph.D. in evolutionary computation applied to design in 1996, at the age of 24. Peter runs the Digital Biology Interest Group at UCL. His research investigates evolutionary algorithms, computational development, neural networks, artificial immune systems, swarming systems and other complex systems, applied to diverse applications including sustainability, design, control, novel robotics, nanotechnology, fraud detection, mobile wireless devices, security, art and music composition. He is also author of the number one bestselling iPhone app iStethoscope Pro. Peter helped create the UK's Net Zero Carbon Buildings Standard. He was nominated for the $100,000 Edge of Computation Prize in 2005, and was a finalist for the AAAS 2010 Science Books & Films Prize. Through his research and his books he often gives public lectures, takes part in debates, and appears on radio and television; he was the host of the monthly Royal Institution's Cafe Scientifique, and a Science Media Expert for the RI Science Media Centre. He is a regular science and technology writer for BBC Focus magazine. He regularly gives plenary speeches at international scientific conferences and is a consultant, convenor, chair and reviewer for workshops, conferences, journals and books in the field of evolutionary computation and complex systems. He has published over 300 scientific papers and is editor of the books "Evolutionary Design by Computers", "Creative Evolutionary Systems" and "On Growth, Form and Computers", and author of "The PhD Application Handbook" and the popular science books "Digital Biology", "The Book of Numbers", "The Undercover Scientist", "Digitized", "10 Short Lessons in Artificial Intelligence and Robotics" and "Artificial Intelligence in Byte-sized Chunks."
Full professor of software engineering, The Free University of Bozen-Bolzano, Italy
Abstract:
5G network technologies with higher performance and reliability have started to support a wide range of applications. Autonomous cars pose a number of specific challenges. In this talk, we review a number of use cases such as in-car infotainment or coordinated maneuvers. We look at the supporting technologies to manage the emerging data processing this intelligently in an IoT and cloud-edge continuum supported by 5G communication. under the given latency and reliability constraints. Large volumes of discrete data is produced and communicated by cars, but also navigation or entertainment content such as videos are streamed into the cars. Both application and system data is collected in order to allow autonomous decisions to be made and actuation to be carried out both in the edge and the cars themselves. We investigate the use of machine learning for autonomous self-adaptation.
Biography:Claus Pahl is a full professor of software engineering at the Free University of Bozen-Bolzano, Italy He received a M.Sc. degree in Computer Science from the Technical University of Braunschweig and a PhD in Computer Science from the University of Dortmund. He has held academic positions at Dublin City University, University College Cork, Trinity College Dublin and the Technical University of Denmark, before taking up his position in Bolzano. He has been a visiting researcher and guest professor at universities in Oldenburg (Germany), Edinburgh (UK) and Shenyang (China). He served as Dean and Vice-Dean of Research of the Faculty of Computer Science at the Free University of Bozen-Bolzano. He has also been Principal Investigator and Cloud Architecture Area Leader of the Irish Centre for Cloud Computing and Commerce IC4, a National Technology Centre that is operated between 3 Universities and works together with more than 40 industry members. He has been a member of the Executive Board for Lero, the Irish Software Research Centre (a National Research Centre with more than 150 researchers working across 7 universities) and acted as Director of the CloudCORE Cloud Computing Research Centre. His research interests lie in the software engineering field, specifically focusing on software architecture. Service engineering and cloud/IoT architectures have served as a specific application context for his architecture research – looking into migration, architecture specification, dynamic quality and performance engineering. He has published more than 300 journal and conference papers including all top cloud publications, has a h-index of 53 (according to Google Scholar) with more than 11500 citations, has chaired many international conferences in the software engineering and cloud/edge technologies context, such as IEEE ECOWS, ICSOC, SOFSEM or CLOSER, has been on seven journal editorial boards. He has been awarded more than 5.5m Euro in research funding from national and international sources, involving both industry and academia. He has participated in the conference organisation of more than 200 events and has reviewed for more than 35 journals. He has been an evaluator for research and innovation projects in 9 countries across Europe, North America and Asia and has reviewed many FP7, H2020 and HE projects for more than a decade. He has been an invited speaker and panelist at many service, cloud and edge events.
Flower Labs GmbH, Germany
Abstract:
As AI reaches new heights, traditional centralized learning approaches face escalating challenges related to data privacy, security, and accessibility. Federated Learning (FL) has emerged as a transformative paradigm that enables the training of machine learning models across decentralized data sources without compromising individual privacy. By allowing models to learn from vast and diverse datasets distributed across devices and organizations, FL overcomes the limitations of data silos and regulatory constraints. This talk will delve into how Federated Learning stands as one of the most significant opportunities to enable the next AI breakthrough. We will explore the core principles of FL, compare it with conventional methods, and discuss real-world applications that demonstrate its potential to revolutionize industries such as healthcare, finance, and telecommunications. Through our work on Flower (https://flower.ai), an end-to-end framework that enables both FL research and FL production deployments in enterprise settings, we aim to illustrate a clear pathway for integrating Federated Learning into mainstream AI development. The convergence of FL with emerging technologies positions it not just as an incremental improvement but as a catalyst for unprecedented progress in AI.
Biography:Daniel is a key contributor to the field of federated learning and one of the creators of Flower (https://flower.ai), an open-source framework for training AI on distributed data. Flower has become a leading federated AI ecosystem, used by Fortune 500 companies and top universities worldwide. Daniel's work has advanced the field significantly, resulting in numerous scientific publications and collaborations. He is currently a PhD candidate in Computer Science at the University of Cambridge and co-founder of Flower Labs. At this conference, Daniel will present a comprehensive 3-hour tutorial on federated learning in the context of IoT.