Resources

AI-Medical is based in the Leeds Institute for Data Analytics adjacent to the Faculty of Medicine and Health, where  many of our clinical supervisors have their University base. We also plan to have a presence in the new Sir William Henry Bragg Building in the Faculty of Engineering and Physical Sciences, home to most of our AI supervisors.

 

Our researchers work closely with clinicians from hospitals that are part of the Leeds Teaching Hospitals Trust. St James’ hospital is home to the Leeds Cancer Centre, one of the largest cancer centres in the UK. The Leeds General Infirmary adjacent to the University campus is currently undergoing a major development of two new ‘Hospitals of the Future’ for adults, and for children and young people.

 

Our researchers use many important medical datasets, including historical data from regional health services. There are major developments generating new datasets that will become available in the near future. For example, the Northern Pathology Imaging Cooperative is installing facilities for digital pathology imaging across the  region and beyond. This will provide a vast repository of histopathology images to support the development of new AI methods for diagnosis and treatment

 

To undertake research at the forefront of machine learning, AI-Medical researchers have access to a range of High-Performance Computing (HPC) facilities. The University’s own HPC installation has over 6000 cores and a growing number of GPUs. We also use N8’s Bede machine consisting of 32 IBM Power 9 dual-CPU nodes, each with 4 NVIDIA V100 GPUs and high performance interconnect, and the national JADE facility consisting of 63 NVIDIA MAXQ Deep Learning Systems. Where our research uses particularly sensitive data, the LIDA LASER service provides a secure computing environment with ISO27001 and NHS Data Security and Protection Toolkit compliance.

 

The University has one of the largest libraries in the country, providing access to subscription journals and supporting the curation of research data.