In this new seminar sub-series, our new academic colleagues working in data science and AI-related areas introduce themselves to the IDSAI community with short talks on their current research.
On 7 March, Simon Rudkin and Xian Yang will talk about their latest research. This will be a hybrid event, taking place at The University of Manchester and online.
Simon Rudkin | Inference on Multi-Dimensional Data with Topological Data Analysis Ball Mapper
The goal of statistical modelling is to provide the best fit to an outcome surface which extends across a joint distribution of explanatory factors. Anscombe (1973)’s quartet of datasets demonstrate the pitfalls of applying a univariate linear model without visualising the behaviour of the outcome across the distribution of the explanatory factors. Topological Data Analysis Ball Mapper (TDABM) produces visualisations and metrics that represent the joint distribution of the characteristics of each data point. The space is covered by a set of balls which represent points with similar characteristics. Balls may then be coloured according to functions on the data points within the ball. For example, the function may be the average value of an outcome of interest. Consequently, it becomes possible to talk about joint-density across the space, identify subspaces where outcomes are observably different and conduct localised statistical analyses. This talk introduces the TDABM algorithm, presents examples and highlights the research agenda for deriving inference on multi-dimensional data with TDABM.
Bio: Simon Rudkin is a Senior Lecturer in Data Science based within the Social Statistics Department. His research focuses on the information which is held within data and the ability to use that information for societal benefit. Much of Simon’s research focuses on the development of Topological Data Analysis (TDA) for understanding data in the social sciences and humanities. His work has considered applications in the UK, China, Europe, and the USA. Topics covered include the health impacts of supermarkets, regional productivity, the digital economy, and finance. He welcomes applications for PhD research on any application where the improved use of statistical methodologies may answer research questions as yet not fully understood.
Xian Yang | Towards personalized healthcare: AI-based approaches
Precision medicine aims to provide personalized healthcare for individual patients based on their health conditions. In the field of precision medicine, electronic health records (EHRs) are becoming increasingly important in understanding patients' health conditions and making clinical decisions. However, analysing diverse, high-dimensional, and long-term EHRs can be challenging. This presentation will explore various AI models applied to EHR data for precision medicine. Important tasks within precision medicine will be discussed, including patient cohort selection, disease risk prediction, and drug recommendation. Models utilizing multi-modal learning, natural language processing, and graph neural networks will be presented as solutions for these precision medicine tasks.
Bio: Xian Yang is currently a lecturer in Data Science at Alliance Manchester Business School (AMBS), The University of Manchester. Prior to joining AMBS, she worked as an Assistant Professor at the Department of Computer Science at HKBU, a researcher at Microsoft Research Asia, and a research fellow at the Data Science Institute of Imperial College London. In 2016, Dr. Yang received her PhD from the Department of Computing at Imperial College London. Her research interests include artificial intelligence in healthcare, natural language processing, data mining, and computational epidemiology. She has published her work in several top-tier conferences and journals such as ICDM, WWW, FSE, COLING, and EMNLP.
This event has been organised by the Institute for data Science and AI.
IDSAI is one of The University of Manchester's Digital Futures network themes.