Project description
Realising the dream of driverless vehicles via better distributed algorithmic integration
Federated learning (FL) is a machine learning (ML) technique that collaboratively trains an algorithm across multiple decentralised devices. The computational demands of driverless vehicles require better FL integration with cellular networks so as to realise real-time responses. Funded by the Marie Skłodowska-Curie Actions programme, the FLASH project will work towards this aim. It will develop a model for FL algorithms to handle the constraints of distributed data sets. It will also develop algorithms capable of allocating cellular network resources to inference tasks determined by ML. The application of FLASH outcomes in (semi-) autonomous vehicles will improve safety and reliability, in addition to reducing fossil fuel usage and emissions.
Objective
The FLASH project aims to establish the theoretical foundations of machine learning and wireless communications that will enable the vision of assisted and self-driving systems. Unfortunately, current systems cannot provide safe and reliable driving because they lack distributed and real-time learning algorithms meeting the critical latency and reliability requirements in highly dynamic and fast-varying wireless channels. Although the fifth generation of cellular systems supports the communication demands for assisted and self-driving, and machine learning proposes federated learning for distributed scenarios, the wireless communications and machine learning domains are not sufficiently integrated for real-time critical applications. Yet, it is only by their integration that the vision of assisted and self-driving will become real. To this end, we will establish a theoretical and algorithmic integration of federated learning and cellular networks that serve vehicles, which we refer to as federated learning supporting efficient and reliable inference over vehicular networks (FLASH). FLASH builds on the co-design of a fundamentally new ecosystem in which federated learning algorithms address critical constraints from vehicular applications, while resource allocation algorithms adapt wireless communication resources to the inference tasks. The project will implement FLASH by establishing and validating theoretical and algorithmic foundations for assisted and self-driving systems. Thus, we not only expect to have an academic impact but also a great societal impact by enabling the fulfilment of sustainable development goals through reduced fuel consumption, traffic emissions, and fatalities. Ultimately, the project provides outstanding training for a talented young researcher, Dr. Mairton Barros, at Princeton University for 24 months with Prof. H. Vincent Poor, and KTH Royal Institute of Technology for 12 months with Prof. Carlo Fischione.
Fields of science
- engineering and technologyelectrical engineering, electronic engineering, information engineeringinformation engineeringtelecommunicationstelecommunications networksmobile network
- natural sciencescomputer and information sciencescomputer securitynetwork security
- engineering and technologyelectrical engineering, electronic engineering, information engineeringinformation engineeringtelecommunicationsradio technology
- natural sciencescomputer and information sciencesartificial intelligencemachine learning
Keywords
Programme(s)
- HORIZON.1.2 - Marie Skłodowska-Curie Actions (MSCA) Main Programme
Funding Scheme
HORIZON-TMA-MSCA-PF-GF - HORIZON TMA MSCA Postdoctoral Fellowships - Global FellowshipsCoordinator
100 44 Stockholm
Sweden