Massive data sets play a central role in most areas of modern life. IT giants such as Amazon evaluate large amounts of customer data in order to provide instant buying recommendations, Facebook maintains a social network of billions of monthly active users, and Google collects data from billions of webpages every day in order to answer more than 70,000 search queries every second. Areas such as healthcare, education, and finance are heavily reliant on Big Data technology. However, extracting useful information from massive data sets is a challenging task. Since modern massive data sets far exceed the RAM (random access memory) of modern computers, algorithms for processing such data sets are never able to "see" the input in its entirety at any one moment. This is a fundamental restriction that is similar to how we humans process the world around us: Our senses provide us with an almost unlimited amount of information as we walk through life, however, our brains are merely able to store a small summary of our past and most information is lost. Data streaming algorithms are massive data set algorithms that mimic this behaviour: A data streaming algorithm scans a massive input data set once only and maintains a summary in the computer's RAM that is much smaller than the size of the input. The objective is to maintain summaries that are as small as possible but still represent the input data well enough in order to solve a specific task. This poses many interesting questions that have been the subject of research for multiple decades: Which problems allow the design of small-space streaming algorithms? Are there problems that cannot be solved with small space, and if this is the case, can these problems at least be solved approximately? This project addresses streaming algorithms for processing massive graphs. Graphs are central objects in computer science and have countless applications. They allow us to describe relations (called edges) between entities (called nodes). For example, a social network graph consists of friendship relations (edges) between pairs of users (nodes). Most research on data streaming algorithms for processing massive graphs assumes that graphs are static objects that never change in structure or size. However, this assumption can rarely be guaranteed in practice. For example, social network graphs change both in structure and size when users join or leave the network or when new friendships are established and existing ones are ended. This observation yields the central questions of this project: Processing dynamic graphs is at least as hard as processing static graphs, but how much harder is it? By how much do summaries have to increase in size? The latter question was first addressed in a seminal paper in 2012. To the surprise of many data streaming researchers, it was shown that for many important problems, the summaries required for the dynamic setting only need to be marginally larger than those for the static setting, while a substantial increase in size was expected. Today, a multitude of problems are known that behave in a similar way while only very few problems are known that require substantially larger summaries in the dynamic setting. The aim of this project is to shed light on the space requirements of streaming algorithms for processing massive dynamic graphs. While our current knowledge suggests that most problems do not become substantially harder in the dynamic setting, we believe that this picture is somewhat skewed and that a multitude of key problems are in fact much harder to solve in the dynamic case. To confirm our conjecture, we will design new streaming algorithms and prove impossibility results that show this to be the case. Where streaming dynamic graph processing provably requires too much space to be practically feasible, we will provide alternative models that allow for the design of streaming algorithms with small space.

'Autonomous systems' are machines with some form of decision-making ability, which allows them to act independently from a human controller. This kind of technology is already all around us, from traction control systems in cars, to the helpful assistant in mobile phones and computers (Siri, Alexa, Cortana). Some of these systems have more autonomy than others, meaning that some are very predictable and will only react in the way they are initially set up, whereas others have more freedom and can learn and react in ways that go beyond their initial setup. This can make them more useful, but also less predictable. Some autonomous systems have the potential to change what they do, and we call this 'evolving functionality'. This means that a system designed to do a certain task in a certain way, may 'evolve' over time to either do the same task a different way, or to do a different task. All without a human controller telling it what to do. These kinds of systems are being developed because they are potentially very useful, with a wide range of possible applications ranging from minimal down-time manufacturing through to emergency response and robotic surgery. The ability to evolve in functionality offers the potential for autonomous systems to move from conducting well defined tasks in predictable situations, to undertaking complex tasks in changing real-world environments. However, systems that can evolve in function lead to legitimate concerns about safety, responsibility and trust. We learn to trust technology because it is reliable, and when a technology is not reliable, we discard it because it cannot be trusted to function properly. But it may be difficult to learn to trust technology whose function is changing. We might also ask important questions about how functional evolutions are monitored, tested and regulated for safety in appropriate ways. For example, just because a robot with the ability to adapt to handle different shaped objects passes safety testing in a warehouse does not mean that it will necessarily be safe if it is used to do a similar task in a surgical setting. It is also unclear who, if anyone, bears the responsibility for the outcome of functional evolution - whether positive or negative. This research seeks to explore and address these issues, by asking how we can, or should, place trust in autonomous systems with evolving functionality. Our approach is to use three evolving technologies - swarm systems, soft robotics and unmanned air vehicles - which operate in fundamentally different ways, to allow our findings to be used across a wide range of different application areas. We will study these systems in real time to explore both how these systems are developed and how features can be built into the design process to increase trustworthiness, termed Design-for-Trustworthiness. This will support the development of autonomous systems with the ability to adapt, evolve and improve, but with the reassurance that these systems have been developed with methods that ensure they are safe, reliable, and trustworthy.

The LEAP Digital Health Hub is a partnership of the South West's leading Universities, more than 20 supporting companies nationally, many NHS Trusts & Health Boards, 4 social care organisations, the region's Local Authorities, the West of England Academic Health Science Network (AHSN), the award-winning business incubator SETsquared and Health Data Research UK (HDRUK). The 50+ partners that shaped this bid ranged from the research director for a provider of residential care homes, to a chief clinical information officer working in an intensive care unit; from the founder of a femtech startup to the head of the healthcare analytics team for a multinational consulting firm. In workshops through June and July 2022 they told us that Digital Health is as much about design and user experience as health data analysis; it is motivated by patient benefit but must also consider viable business models for industry. All Hub partners will have access to dedicated physical office space in central Bristol alongside the EPSRC Centre for Doctoral Training (CDT) in Digital Health and Care. There, they will train, network and research together across disciplines and sectors. They will engage with partners across the UK- and beyond. Recognising that UK breakthroughs in Digital Health may be equally (or more) impactful abroad, the Hub's new "Global Digital Health Network" links the Hub to Digital Health expertise from the US, China, India, Nigeria and Australia (sections B1.2, B5). The Hub's unique Skills and Knowledge Programme is designed to address the professional training needs of industry, health and social care providers and academia within the two Themes of Transforming Health & Care Beyond the Hospital and Optimising Disease Prediction, Diagnosis & Intervention. This is proposed to be the world's largest Digital Health taught programme. The Hub's Fellowship programme will comprise 5 different schemes to develop future leaders, within not only academia, industry and the health/care sector, but also within the community - as patients or informal carers. The Hub's Research programme focusses on pre-competitive research within the Hub's two thematic areas of Transforming Health and Care Beyond the Hospital and Optimising Disease Prediction, Diagnosis and Intervention. The Hub will add value by surfacing health priorities from its partner health and social care organisations, working with the West of England AHSN and also with Hub members such as Chief Nursing Information Officers, with charities, social care providers, patient and community groups.