The Advances in Data Science and AI seminar series showcases innovative research in data science and AI from across the world.
This event is hybrid.
Title: Learning from healthcare data across domains and modalities
Abstract : In healthcare, we have data across domains such as multiple sites, and across modalities including medical images, multiomics, and electronic health records. In practice, clinicians often make decisions using data from more than one modality and leveraging experience from related domains. However, it is challenging for machine learning models to generalise and learn from healthcare data across domains and modalities, mainly due to their inherent complexity and the limited availability of labelled data. In this talk, I will share how we can tackle these challenges via developing advanced machine learning models including domain adaptation, multimodal learning, and graph neural networks, while encouraging interpretability. I will show applications on brain disorder classification, drug-target interaction prediction, polypharmacy side effect prediction, and cancer diagnosis from multiomics. I will also introduce PyKale, our open-source machine learning software library for enabling and accelerating interdisciplinary research.
Bio: Haiping Lu is a Professor of Machine Learning, the Turing Network Development Award Lead, and Insigneo Research Director for Healthcare Data / AI at University of Sheffield. He is also the lead organiser of the Alan Turing Institute’s interest group on meta-learning for multimodal data. He received his BEng and MEng from Nanyang Technological University, Singapore, in 2001 and 2004, and his PhD from University of Toronto, Canada, in 2008.
His research focuses on developing translational AI technologies for better analysing multimodal data in healthcare and beyond, particularly multidimensional data and heterogeneous graphs in bioinformatics and medical imaging. He leads the development of the PyKale library (http://pykale.github.io/) to provide more accessible machine learning from multiple sources for interdisciplinary research, officially part of the PyTorch ecosystem.
He serves as an Associate Editor of IEEE Transactions on Neural Networks and Learning Systems, and IEEE Transactions on Cognitive and Developmental Systems. He leads the development of a course on An Introduction to Transparent Machine Learning (https://pykale.github.io/transparentML/), part of the Alan Turing Institute’s online learning courses in responsible AI. He was a recipient of a Turing Network Development Award, an Amazon Research Award, an AAAI Outstanding PC Member Award, a Hong Kong Research Grants Council Early Career Award, and an IEEE CIS Outstanding PhD Dissertation Award. He was also a joint-recipient of a Wellcome Trust Innovator Award and an NIHR AI in Health and Care Award.
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.
Please book via ADSAI Eventbrite link under More Information.
