Digital technologies have a transformative impact in the economy and wider society. New innovations in Information Communication Technology (ICT) such as the next generation '5G' internet, automation and robotics, and big data and Artificial Intelligence (AI) have the potential to make a profound societal impact and the pace of development is staggering. The same technologies can though have a negative impact on society, including significantly increasing the carbon emissions related to ICT and thus creating damaging impacts on our environment. Managing this duality between ICT's benefits and risks must be at the heart of future ICT design and innovation - ensuring ICT can continue to bring value to our society and the economy, while keeping ICT innovations from exceeding planetary boundaries. However, there is currently scarce consideration of systemic impacts within ICT innovation, and design processes today lack the information and tools required to embed environmental sustainability into ICT. This project, PARIS-DE, will ensure that the carbon emissions associated with the ICT sector are aligned with the Paris agreement: limiting temperature increases to 1.5 degrees Celsius. To do this, the PARIS-DE project will develop a digital sustainability framework that systemically considers ICT's impacts and ensures Paris-compliant design through two key concepts: i) an evidence base around carbon emissions in the digital economy, and ii) a responsible innovation approach that targets environmental sustainability, yet maintains key aspects of ICT design that enable societal thriving. Using a range of disciplinary perspectives including computer science, human-centred design, philosophy and ethics and environmental economics, PARIS-DE will develop digital tools that support ICT development within planetary boundaries, and will create, demonstrate and evaluate the digital sustainability framework through three case studies: 1) big data and AI, 2) autonomous systems, and 3) video streaming. These case studies, taken as representative of the digital economy, will allow for an evaluation of different underlying technologies that threaten rising emissions. The case studies will also involve working closely with key stakeholders in ICT innovation (e.g. designers and developers in the ICT sector), ensuring the framework is comprehensive and effective. PARIS-DE will ultimately allow the ICT sector to innovate technology more sustainably and in-line with climate change mitigation targets.

AI/ML systems are becoming an integral part of user products and applications as well as the main revenue driver for most organizations. This resulted in shifting the focus toward the Edge AI paradigm as edge devices possess the data necessary for training the models. Main Edge AI approaches either coordinate the training rounds and exchange model updates via a central server (i.e., Federated Learning), split the model training task between edge devices and a server (i.e., split Learning), or coordinate the model exchange among the edge devices via gossip protocols (i.e., decentralized training). Due to the highly heterogeneous learners, configurations, environment as well as significant synchronization challenges, these approaches are ill-suited for distributed edge learning at scale. They fail to scale with a large number of learners and produce models with low qualities at prolonged training times. It is imperative for modern applications to rely on a system providing timely and accurate models. This project addresses this gap by proposing a ground-up transformation to decentralized learning methods. Similar to Uber's delivery services, the goal of KUber is to build a novel distributed architecture to facilitate the exchange and delivery of acquired knowledge among the learning entities. In particular, we seize an opportunity to decouple the training task of a common model from the sharing task of learned knowledge. This is made possible by the advances in the AI/ML accelerators embedded in edge devices and the high-throughput and low-latency 5G/6G technologies. KUber will revolutionize the use of AI/ML methods in daily-life applications and open the door for flexible, scalable, and efficient collaborative learning between users, organizations, and governments.

AI holds great promise in addressing several grand societal challenges, including the development of a smarter, cleaner electricity grid, the seamless provision of convenient on-demand mobility services, and the ability to protect citizens through advice and informed deployment of medical, emergency and police resources to fight epidemics, deal with crises and prevent crime. However, these promises can only be realised if citizens trust AI systems. In this fellowship, I will develop the fundamental science needed to build trusted citizen-centric AI systems. These AI systems will put citizens at their heart, rather than view them as passive providers of data. They will make decisions that maximise the benefit for citizens, given their individual constraints and preferences. They will use incentives where appropriate to encourage positive behaviour change, but they will also be robust to strategic manipulation, in order to prevent individuals from exploiting the system at the expense of others. Importantly, citizen-centric AI systems will involve citizens and other stakeholders in a feedback loop that enables them to audit decisions and modify the system's behaviour to ensure that effective but also ethical decisions are taken. Achieving this vision of citizen-centric AI systems requires several novel advances in the area of artificial intelligence. First, to safeguard the privacy of individuals, new approaches to understanding the constraints and preferences of citizens are needed. These approaches will be distributed in nature - that is, they will not depend on collecting detailed data from individuals, but will allow citizens to manage and retain their own data. To achieve this, I will develop intelligent software agents that act on behalf of each citizen, that store personal data locally and only communicate limited information to others when necessary. Second, to incentivise positive behaviour modifications and to discourage exploitation, I will draw on the field of mechanism design to model how self-interested decision-makers behave in strategic settings and how their actions can be modified through appropriate incentives. A particular challenge will be to deal with limited information, uncertainty about preferences and a constantly changing environment that necessitates incentives to be dynamically adapted via appropriate learning mechanisms. Finally, to enable an inclusive feedback loop involving citizens and other stakeholders, new interaction mechanisms are needed that can provide explanations for actions as well as information about whether the system is making fair decisions. While there is a wealth of emerging work on explainability and fairness in AI, this typically deals with simple one-shot problems. In contrast, I will consider more realistic and complex sequential settings, where actions have long-term consequences (including on fairness) that may not be immediately apparent. As part of the fellowship, I will work with a range of partners to put the research into practice and generate real impact. With EA Technology and the Energy Systems Catapult, I will work on incentive-aware smart charging mechanisms for electric vehicles. With Dstl and UTU Technologies, I will develop disaster response applications that use crowdsourced intelligence from citizens to provide situational awareness, track the spread of infectious diseases or issue guidance to citizens. With Siemens, Jaguar Land Rover, Thales and the Connected Places Catapult, I will develop new approaches for trusted on-demand mobility. With Fawley Waterside, I will work on citizen-centric solutions to smart energy and transportation in the Southampton area. With Dstl and Thales, I will explore further applications to national security and policing. Finally, with IBM Research, I will develop new explainability and fairness tools, and integrate these with their existing open source frameworks (AI Fairness 360 and AI Explainability 360).

The past years have witnessed a rapidly growing number of wirelessly-connected devices such as smartphones and Internet-of-Things (IoT) equipment, which generate ever-increasing amounts of data driving key Artificial Intelligence (AI) applications. However, users are increasingly unwilling to allow their private data (such as media, location, or sensor data) to be uploaded to a central location (e.g., cloud datacentre) for training Machine Learning (ML) models, and data-protection laws such as the Data Protection Act 2018 are growing more restrictive towards data usage. Federated Learning (FL) is a game-changing technology conceived to address the growing data privacy concern by moving training from the datacentre to user devices at the network edge, allowing sensitive data to remain on the devices where it was generated. FL has enormous potential for real-world, privacy-sensitive applications such as autonomous driving, diagnostic healthcare, and predictive maintenance. The operating environment for FL at the edge is extremely challenging for a variety of reasons: 1) the data owned by FL clients is highly heterogeneous (in regard to data distribution, quality, and quantity) and dynamic (data distributions change over time); 2) the hardware devices have diverse computing and communication capabilities with stringent resource constraints (e.g., battery power); and 3) FL clients work under unreliable wireless edge network conditions. Hence, despite FL's huge promise, there are considerable barriers to its wider real-world adoption for mission-critical AI applications that need real-time, on-demand responses, caused by several grand challenges: Challenge 1) lack of FL algorithms delivering consistent performance for dynamic client data, diverse client hardware, and unreliable wireless connections simultaneously; Challenge 2) lack of rigorous theoretical analyses of real-time, real-world FL algorithms; Challenge 3) lack of optimised, energy-efficient, versatile hardware acceleration for real-time FL. To address these important challenges, this project will create revolutionary algorithm-hardware co-design approaches to make FL a real-time process with unparalleled speed, performance, and energy-efficiency at the wireless edge, capable of meeting the stringent requirements of mission-critical applications. This research will pioneer a set of original methods and innovative technologies including: 1) disruptive lightweight hardware-aware FL algorithms that significantly reduce communication, computing, and energy costs while achieving fast model updates; 2) rigorous mathematical analyses of the proposed algorithms to prove their convergence rates and offer theoretical insights into how they perform under various edge network conditions; 3) an automatic hardware-software co-optimisation framework integrating specialised training-acceleration and power-reduction methods to realise optimised, energy-efficient hardware acceleration; and 4) a unique prototype system that will integrate the designed FL hardware accelerator and real-time FL algorithms and be evaluated in a realistic wireless edge networking testbed. This project has the potential to transform FL from a lengthy and disjointed process to a continuous, real-time procedure with concurrent model training and deployment. The proposed research will contribute to the UK's digital transformation and green economy by creating ground-breaking technologies for creating innovative AI-empowered products with significantly improved performance and energy-efficiency while complying with strict data-privacy protection.

DRIVE-Health will train a minimum of 85 PhD health data scientists and engineers with the skills to deliver data-driven, personalised, sustainable healthcare for 2027 and beyond. Co-created with the NHS Trusts, healthcare providers, patients, healthtech, pharma, charities and health data stakeholders in the UK and internationally, it will build on the successes of its King's College London seed-funded and industry-leveraged pilot. Led by an established team, further growing the network of funding partners and collaborators built over the past four years, it will leverage an additional £1.45 of investment from King's and partners for every £1 invested by EPSRC. A CDT in data driven health is needed to deliver the EPSRC Priority for Transforming Health and Healthcare, EPSRC Health Technologies Strategy, and on challenges laid out in the UK Government's 2022 Plan for Digital Health and Social Care envisaging lifelong, joined-up health and care records, digitally-supported diagnoses and therapies, increasing access to NHS services through digital channels, and scaling up digital health self-help. This ambition is made possible by the increasing availability of real-world routine healthcare data (e.g. electronic health care record, prescriptions, scans) and non-healthcare sources (e.g. environmental, retail, insurance, consumer wearable devices) and the extraordinary advances in computational power and methods required to process it, which includes significant innovations in health informatics, data capture and curation, knowledge representation, machine learning and analytics. However, for these technological and data advances to deliver their full potential, we need to think imaginatively about how to re-engineer the processes, systems, and organisations that currently underpin the delivery of healthcare. We need to address challenges including transformation of the quality, speed and scale of multidisciplinary collaborations, and trusted systems that will facilitate adoption by people. This will require a new generation of scientists and engineers who combine technical knowledge with an understanding of how to design effective solutions and how to work with patients and professionals to deliver transformational change. DRIVE-Health's unique cohort-based doctoral research and training ecosystem, embedded across partner organisations, will equip students with specialist skills in five scientific themes co-produced with our partners and current students: (T1) Sustainable Healthcare Data Systems Engineering investigates methods and frameworks for developing scalable, integrated and secure data-driven software systems (T2) Multimodal Patient Data Streams will enable the vision of a highly heterogeneous data environment where device data from wearables, patient-generated content and structured/unstructured information from electronic health records can combine seamlessly (T3) Complex Simulations and Digital Twins focuses on the paradigm of building simulated environments, including healthcare settings or virtual patients, to make step-change advances in individual predictive models and to inform clinical and organisational decision-making. (T4) Trusted Next-Generation Clinical User Interfaces will place usability front and centre to ensure health data science applications are usable in clinical settings and are aligned with users' workflows (T5) Co-designing Impactful Healthcare Solutions, is a cross-cutting theme that ensures co-production and co-design in the context of health data science, engagement with stakeholders, evaluation techniques and achieving maximum impact. The theme training will be complemented with a cohort and programme-wide approach to personal, career, professional and leadership development. Students will be trained by an expert pool of 60+ supervisors from KCL and across partners, delivering outstanding supervision, student mentoring, opportunities, research quality and impact.