NVIDIA Launches Supercomputer to Aid UK Healthcare

NVIDIA’s new supercomputer, Cambridge-1, is now available to UK healthcare researchers to allow quicker solution of pharmaceutical challenges, including COVID-19. Major industry companies including AstraZeneca (AZ), GSK, Oxford Nanopore and Guy’s and St Thomas’ NHS Foundation Trust. 

Cambridge-1 is the 29th most powerful supercomputer in the world, able to manage over 400 petaflops of AI performance and eight petaflops of Linpack performance. NVIDIA was able to bring the project to completion after just 20 weeks, compared with an average time for such projects of several years, due to the modular nature of its architecture. 

Cambridge-1 will focus on four key areas within the UK healthcare and life science sector. These are joint industry research to solve large-scale healthcare problems and decrease industry costs; computing time donated to certain university-granted studies to seek cures; support and provide early-access tools to AI startups; and educate the next generation of AI practitioners. 

The computer acts as part of NVIDIA’s AI Center of Excellence in Cambridge, which intends to house further supercomputers in the future to support other industries around the UK. NVIDIA has stated it will be investing around £50 million in the supercomputer. 

Supercomputers have long been theorised to soon play an extremely important role in healthcare. As one of the most inefficient and expensive sectors in the world, healthcare and pharmaceutical companies are awash in data that cannot all be processed.

With their incredible processing capacities, supercomputers promise a solution to the vast amounts of data that the field is facing. Rapid processing and analysis capabilities would allow for swift data analysis, finding more biomarkers to predict patient disorders before they occur. Additionally, the computer aids doctors in making evidence-based diagnoses, running through data infinitely more fast than a manual check ever could.

GSK and AZ are two of the big pharma companies already set to make use of Cambridge-1 in the near future. The applications of such a computer are far-ranging, running from research and development through to late stages of drug development. Among other things, Cambridge-1 will be able to better identify targets and compounds, use Natural Language Processing to integrate and analyse data, and create breakthroughs into a number of health conditions. 

AZ intends to use the computer to create a transformer-based generative AI model for chemical structures, as well as a project using AI in digital pathology to analyse tissue sample slide images more quickly during drug response research. GSK has said it will use Cambridge-1 to create medicines twice as likely to go through to market than current products. 

Research institutions such as Guy’s and St Thomas’ NHS Foundation Trust are desiring to use the computer to revolutionise diagnosis and treatment of patients through the creation of large-scale models that can determine best course of action for patients using imaging scans and blood reports. 

One example given by the Trust related to creation of an AI model which would design synthetic brain images using machine learning to parse through thousands of real MRI scans. This model could generate unlimited images with specific disease characteristics. This in turn would let scientists learn more about such diseases, and thus diagnose and treat them much earlier. 

NVIDIA said it selected the UK for its site due to the country’s medical research prowess. This comes as part of a boom in the country’s healthcare research, with huge resources of data from the NHS and UK Biobank dwarfing the datasets of other companies. 

Joshua Neil, Editor
Proventa International