Novel and innovative tools and techniques are required to ensure that whilst mining is carried out in order to achieve some of the United Nations Sustainable Development Goals, it is not at the detriment of others. Yet, whilst extensive work and research is carried out to address a number of concerns for the mining industry, the tools developed are rarely implemented - even where the science is excellent. There are a number of reasons for this, but we believe one of the most important is due to the complex relationships present within the raw materials sector. The mining industry itself is very complex, with a broad and diverse range of stakeholders. Innovation is often enabled by ACADEMICS, but not always in collaboration with MINING COMPANIES, hence the tools developed may not service them well. Where the two do work together, barriers may include FINANCE, as INVESTORS may not fully understand the advantages of the innovation. Perhaps the best way to integrate novel approaches may be through POLICY and REGULATORS, yet they may not have the subject specific knowledge to enforce this. Both the PROMT and PAMANA project are developing new scientific methods to improve aspects of mining in the Philippines. A Knowledge Exchange (KE) programme, carried out by Fellows who understand these nuances, will address the barriers outlined above. KE will improve understanding of stakeholder perceptions, challenges, and benefits. Then, tools and techniques can be embedded into stakeholder organisations. We propose a KE Fellowship to capitalise on diverse skills, and experience of building relationships across communities, private sector, and public sector; grounded in technical expertise in ecosystem science and economic geology. Our ambition is to build lasting networks that allow PROMT and PAMANA partners a vehicle to embed their research into stakeholder relationships and facilitate partnerships for further collaboration and project follow-on.

The main aim of this research project is to investigate how reinforcement learning, when used by robots in human-robot interactions, can perpetuate unfairness between people. Some of the key research questions that need to be answered to begin to understand the intersection between human-robot interaction, fairness and reinforcement learning are: 1. Where is reinforcement learning being most used in social robotics - are there common types of tasks or themes? 2. Do different types of reinforcement learning algorithms have different fairness issues when applied to social robotics? 3. Is there a unique feature of using reinforcement learning in robotics that creates unfairness? 4. How can algorithms, when using reinforcement learning to solve parts of a robot task, be audited to check for fairness issues? 5. Can any algorithmic mitigations or guidelines be created to help tackle these issues? To achieve these objectives, I will attempt to gain an understanding of the types of robot tasks that industry and research are currently applying reinforcement learning to. I will do this by formulating scenarios where robots interact with people via reinforcement learning, either as proof of concept tasks or from current research papers. These scenarios can be simulated; I will audit these simulations under different definitions of unfairness and bias. I will also seek to understand how reinforcement learning might contribute to this unfairness in ways that are different to data-driven AI, looking to see if the common ways of formulating the task (reward function, actions, and state space) can lead to fairness issues. I aim to come up with adjustments to these algorithms which may decrease unfairness within the results. I will also create some guidelines that future reinforcement learning researchers and engineers can follow to help mitigate unfairness in future work. The expected novel contributions are: 1. New methods for mitigating unfairness in Reinforcement Learning, with a specific attention to robots (or other automated agents, such as self-driving cars). 2. Guidelines for designing reinforcement learning algorithms for robots that will be used around and interacting with people. This PhD is relevant to the following EPSRC research areas: Artificial Intelligence and Robotics Theme, Artificial Intelligence Technologies, Human Computer Interaction, Robots for a safer world, Robotics.

This core facility at King’s College London (KCL) will form the hub of a multi-institution consortium of leading UK molecular imaging researchers and their clinical and biomedical collaborators. It will provide radiochemists and radiobiologists with a unique, comprehensive range of radioanalytical equipment for development and characterisation of the next generation of radiopharmaceuticals for whole-body molecular imaging, cell tracking and targeted radionuclide therapy. The objectives are to: 1. enable chemical synthesis and analysis, and biological evaluation, of new radiochemical platforms and radiopharmaceuticals, and develop radiosynthetic protocols for their clinical translation; 2. develop a new generation of radiotracers, building on recent innovations to help develop and support the clinical use of new advanced therapies: novel molecular, cell-based, nano and radio therapies; 3. use radionuclide molecular imaging to understand disease environment, pathology and resistance to treatment. The facility will enable a consortium of chemists, biologists, pharmacists and clinicians to pursue these objectives across cancer, immunology, neurology, cardiovascular disease, infection, regenerative medicine and transplant therapies, through collaboration with the applicants and through direct access. The requested equipment will complement existing KCL research facilities including whole-body preclinical imaging PET/CT and SPECT/CT scanners funded by a previous Wellcome Multi-User Equipment Grant (2008 – 2013). Scientists have recently discovered how cells such as stem cells and immune cells can cure disease, leading to new therapies based on cells rather than drugs. For this knowledge to benefit patients, it is important to understand how these cells behave in the body: Where do they go? Do they survive? Scientists have also discovered that particular disease cell “markers” can be used to predict whether or not a disease will respond to therapy. Detecting these markers by imaging the whole body with radiotracers will enable better disease monitoring and treatment. We propose to establish a research facility that will house equipment to enable doctors, chemists and biologists to develop new radiotracers which can bind to therapeutic cells or disease “markers” to map their location in living subjects across the whole body, enabling doctors to use new cell-based therapies and cell marker diagnostics safely and efficaciously.

The overarching goal of this proposal is to elucidate the functional diversity of the pancreatic mesenchymal lineage. In embryonic tissues, epithelial progenitors receive paracrine signals from the surrounding mesenchymal niche, which can modulate their ability to proliferate and differentiate. The pancreas consists of a variety of specialized epithelial cells, including endocrine and acinar cells, surrounded by a poorly defined heterogeneous mesenchyme. We hypothesise that different mesenchymal lineages define local instructive microenvironments, including cell–cell crosstalk, ECM and signalling molecules, which eventually trigger distinct differentiation programmes from pancreatic progenitors. Sc-RNA-sequencing has generated a transcriptional map of the pancreatic mesenchyme in the mouse embryo. Here, we will unravel the spatial architecture of the identified mesenchymal cell states, linking their position to emerging pancreatic cell identities. Next, we will assess if mesenchymal lineages with a distinct spatial address underlie unique niche regulatory functions, promoting acinar or β-cell differentiation. Finally, we will study the organisation and function of the identified niche microenvironment(s) in human tissue and pluripotent stem cells. The proposed programme will yield novel insight into pancreas biology and will set the stage for manipulating combinatorial pancreatic niches - an important step towards engineering functional β-cells for regenerative medicine applications. The pancreas is a complex organ comprised of a mixture of various cell types, including epithelial and mesenchymal cells. Pancreatic epithelial cells control vital functions, such as food digestion and regulation of blood sugar levels through insulin-producing cells. The pancreas is the target of incurable diseases: diabetes and pancreatic cancer. The crosstalk between pancreatic epithelial cells and the surrounding mesenchymal cells is fundamental for the organ formation and its functions. We previously discovered a high degree of heterogeneity within pancreatic mesenchymal cells. This proposal aims at answering the question whether this heterogeneity corresponds to functional diversity. We will address if mesenchymal cells at distinct locations create unique microenvironments that promote the formation and function of the adjacent pancreatic cell type. All findings obtained in the mouse will be validated and transposed onto human models. This will pave the way towards new possibilities of engineering functional pancreatic cells for biomedical applications.

he prevalence of mental health problems among adolescents in England is 14% [1]. In adults, there are clear and unexplained differences in the types of mental health problems experienced by white and black people; for instance, rates of psychosis are higher among black people [2]. Among young people, though, research suggests similarities in mental health problems between black and white youths, despite, on average, higher prevalence of risk factors among black adolescents [3]. Understanding ethnic differences in the development of mental health problems during adolescence and into adulthood is a key focus of REACH (Resilience, Ethnicity and AdolesCent Mental Health), the research group I've been working in for the last 2 years. Negatively stigmatising people from an early age has adverse impacts and we know this to be the case even if the label is unfounded. As an example, from educational psychology, "teachers and parents are more likely to perceive disabilities in and hold lower educational expectations for labelled adolescents than for similarly achieving and behaving adolescents not labelled with disabilities" [4]. My proposal is to produce is a novel perspective on how the process of labelling young people as a potential threat to society affects their mental health, and whether these effects differ by ethnic groups and among migrants. The theoretical framework behind labelling theory involves an individual receiving a label that they did not choose for themselves and assumes that, although an outcome can stem from other causes, once an individual has been defined as deviant, problematic or dangerous "they face new problems that stem from the reactions of self and others to negative stigma that are attached to that label" [5, 6, 7]. There is a plethora of research in adults on the criminogenic processes triggered by labelling [5] and, separately, the stigma associated with mental health problems [8]. But there is very little research, locally and globally, on the mental health implications of being labelled a threat to society at an early age.