Massive computational resources that can facilitate investigation everywhere are required for scientific researchers. Access to cutting-edge AI computer resources is essential for fostering creativity, whether researchers are working on ground-breaking pharmaceutical research, investigating alternative energy sources, or developing innovative techniques to stop financial crime. The next wave of innovation can be fueled by this new computer model, which can also address the problems with generative AI.
Some of the greatest healthcare experts in the nation have used Cambridge-1, a supercomputer NVIDIA launched in the U.K. during the pandemic, to power their research. As part of the NVIDIA DGX Cloud, the system is now being used to quicken the pace of scientific innovation and discovery in practically every sector.
As a cloud-based resource, it will increase researchers’ access to AI supercomputing in a variety of fields, including climate science, driverless vehicles, worker safety, and others. It will be delivered with the ease and speed of the cloud, and it is well situated for access in the U.K. and Europe.
Any business can now access cutting-edge supercomputing resources from a browser thanks to the multinode AI training service known as DGX Cloud. 80 NVIDIA DGX computers made up the initial Cambridge-1 infrastructure; now, it will integrate with DGX Cloud to give customers access to top-tier infrastructure.
Healthcare Insights’ Past
The Cambridge-1 supercomputing platform was leveraged by academia, entrepreneurs, and the UK’s major pharma ecosystem to speed up research and provide fresh methods for drug development, genomics, and medical imaging using generative AI. Some of these methods include:
With the help of NVIDIA, the Technical University of Munich Lab, and InstaDeep, a 2.5 billion-parameter LLM for genomics on Cambridge-1 was created. The goal of this study was to develop a model that could predict DNA sequence properties more precisely.
King’s College London used Cambridge-1 to produce 100,000 artificial brain images, making them freely available to researchers in the field of medicine. The researchers at King’s developed accurate, high-resolution 3D photographs of human brains using the open-source AI imaging platform MONAI, training the system in weeks as opposed to months.
Oxford Nanopore used Cambridge-1 to create models for base calling in DNA sequencing very quickly and with excellent efficiency. The business supported inference for the ORG using the supercomputer as well. In one initiative, severely endangered species’ DNA will be sequenced.
In order to better understand why some antibodies function effectively, Peptone, in partnership with a pharmaceutical partner, used Cambridge-1 to run physics-based simulations to assess the impact of mutations on protein dynamics. The discovery of biologics and antibody development may benefit from this research.
Relation Therapeutics created a sophisticated language model that can read DNA to better comprehend genes, an essential step in developing novel drugs. Their work advances our understanding of how specific diseases’ genes are controlled.
