Machine learning offers great promise in helping us solve problems by automatically learning solutions from data, without us having to specify all details of the solution as in earlier computational approaches. However, we still need to tell machine learning systems what problems we want them to solve, and this is currently undertaken by specifying desired outcomes and designing objective functions and rewards. Formulating the rewards for a new problem is not easy for us as humans, and is particularly difficult when we only partially know the goal, as is the case at the beginning of scientific research. In this programme we develop ways for machine learning systems to help humans to steer them in the process of collecting more information by designing experiments, interpreting what the results mean, and deciding what to measure next, to finally reach a conclusion and a trustworthy solution to the problem. The machine learning techniques will be developed first for three practically important problems and then generalized to be broadly applicable. The first is diagnosis and treatment decision making in personalized medicine, the second steering of scientific experiments in synthetic biology and drug design, and the third design and use of digital twins in designing physical systems and processes. An AI centre of excellence will be established at the University of Manchester, in collaboration with the Turing Institute and a number of partners from the industry and healthcare sector, and with strong connections to the networks of best national and international AI researchers.

Addressing the UK's nuclear legacy is the largest, most important environmental remediation programme in Europe, with estimated expenditure of £115 billion over the next 120 years. A significant proportion of this cost is associated with decommissioning and management of high and intermediate level radioactive waste; material that is too radioactive for direct human handling. There is therefore a need for remotely operated, waste characterisation technologies to enable monitoring of such wasteforms in their interim and final storage locations. Due to the extreme radiation fields present, retrospectively fitting sensors that rely upon cables for power and data transmission is not feasible and hence alternative technologies for powering sensors are required. Our project will seek to address this challenge by developing a solution using advanced diamond materials to harvest energy from radioactive decay to power small, portable devices containing multiple sensors that pass data over wireless networks. There are clear benefits for the technology including: less wiring, less maintenance, less dose to operators and an extended lifespan of sensors or mobile platforms. The sensors powered by such devices would be able to provide information for long periods of time that would otherwise be challenging to gather but none the less very important for long term safety cases. Therefore, this technology could represent a significant financial saving for UK plc. By the end of the project we would aim to demonstrate this technology by: (i) deployment in active plant at Sellafield; and (ii) deployment in a reactor core at Kyoto University Research Reactor Institute, Japan.

When you talk to people about what makes humans different from other animals, one of the features that they will rapidly identify is the human hand. Indeed they are very likely to identify the 'opposable thumb' as a uniquely human characteristic. Whilst it can be argued that this is not strictly true, it is certainly the case that there are no other animals that have anything like the degree of precise control of their hands that humans have. We take for granted the fine movements of individual fingers that allow us to play the piano or tie our shoelaces, and these are activities that are impossible for non-humans to achieve. It is likely that the evolutionary history of our species is very closely linked to the evolutionary history of our hands and this is therefore an important area for scientific study. We currently know a great deal about how the shape of our hands has changed from those of closely related species but we actually know very little about how these shape changes are linked to how the hand functions. The story of our divergence from the common ancestor of chimpanzees and modern humans includes walking on our hindlimbs, the creation of stone tools, the increase in our intelligence, and living in extended social groups. The change in function of the human hand by losing its locomotor role and allowing its specialisation for manipulating and sensing the world, and extending its role in communication becomes a compelling narrative. In particular the idea that the evolution of our hands is closely linked to our adoption of increasingly sophisticated tools seems extremely plausible. Thus the aim of this research project is to explore the changes in functional capabilities of the human hand and to use this information to evaluate the evolutionary history of the hand and its relationship to tool use and manufacture. To achieve this goal we need to collect information about how the individual parts of the hand are used in humans. This needs to be done in a controlled fashion so that we can make objective comparisons of the mechanical requirements of different actions that we can link to specific artifacts in the achaeological record. We therefore propose to collect movement and force information from humans whilst performing such a range of tasks. We will use a range of exciting new technologies developed for virtual reality and movie special effects where hand and finger movements can be recorded automatically using specially instrumented gloves and by attaching reflective markers to the fingers. In addition simply recording this information is insufficient to fully understand a mechanism as complex as a hand. We will also construct 3D computer simulations of these hand and arm movements using information from medical imaging and dissections. We will then use a variety of sophisticated mechanical engineering techniques to evaluate how the individual bones and muscles function within the hand. We also need to evaluate how human hand function has changed over time and this means that we need to investigate the hands of fossil primates as well as their living relatives. To do this we will create equivalent computer simulations for these extinct species reconstructed from the fossil bones. The computer models will allow us to predict the capabilities of these species and we will be able to directly evaluate the changes in locomotor, foraging and tool use capabilities of the hands of our closest ancestors over time.

Fullerenes are football-shaped cages of carbon atoms, for the discovery of which the British scientist Harry Kroto won the Nobel prize in 1996. Inside the cage is an empty space. Chemists and physicists have found many ingenious ways of trapping atoms or molecules inside the tiny fullerene cages. These encapsulated compounds are called endofullerenes. A remarkable method was pioneered by the Japanese scientists Komatsu and Murata, one of whom is a project partner on the current proposal. They performed "molecular surgery". First, a series of chemical reactions was used to open a hole in the fullerene cages. A small molecule such as water (H2O) was then inserted into each fullerene cage by using high temperature and pressure. Finally, a further series of chemical reactions was used to "sew" the holes back up again. The result was the remarkable chemical compound called water-endofullerene, denoted H2O@C60. Our team has succeeded in developing a new synthetic route which requires milder conditions and has improved yield for the production of H2O@C60. In addition we will encapsulate other small molecules in the fullerene cage, including ammonia (NH3) and methane (CH4). Molecules of ordinary water have two forms, which are called ortho and para-water, which are distinguished by the way the magnetic hydrogen nuclei point: in opposite sense for para-water, and in the same sense for ortho-water. In ordinary conditions, these two forms interconvert rapidly, and cannot be isolated. However, by trapping water molecules inside fullerene cages, the two forms are isolated and may be studied separately. We recently observed that these two forms of water have different electrical properties. At low temperatures, the two forms interconvert over a period of tens of hours. We will study the interconversion of the two forms of water, and develop a theory of why this conversion changes the electrical properties. In order to understand how these molecules behave, we will use several techniques. These methods include nuclear magnetic resonance (which involves a strong magnet and radiowaves), neutron scattering (in which the material is bombarded with neutrons from a nuclear reactor) and infrared spectroscopy (which involves the absorption of low-energy light waves). By combining the information from these different techniques, we will build up a complete picture of the quantum-mechanical behaviour of the trapped molecules. Since ortho and para-water have different electrical properties, we expect to distinguish between single H2O@C60 molecules in the ortho and para states, by measuring the electrical response of single molecules. This will be done scanning over a surface loaded with the fullerenes, using a very sharp tip. In this way, we hope to observe the ortho to para transition of single molecules - something that has never been done before. Although most of this project concerns basic science, this project could lead to technological and even medical advances in the future. For example, the ortho and para states of the individual H2O@C60 molecules could allow the storage of one bit of information inside a single molecule, without damaging it in any way. This might lead to a new form of very dense data storage. Since a single gram of H2O@C60 contains about 10^19 molecules, this single gram could in principle store 1 million terabytes of information, sufficient to store the DNA sequences of everyone on the planet (although it will be very difficult to store and retrieve this information). In addition, the quantum behaviour of the encapsulated molecules is expected to give rise to greatly enhanced magnetic resonance signals, leading to the possibility of greatly enhanced MRI images, with considerable medical benefits.