The Scripps Research Translational Institute has agreed to work with graphics processing unit-maker NVIDIA to support the development of AI applications. The partners plan to forge AI and deep learning best practices, tools and infrastructure tailored to supporting the AI application development process.
In collaboration with NVIDIA, Scripps will establish a center of excellence for artificial intelligence in genomics and digital sensors. According to Dr. Eric Topol, the Institute’s founder and director, AI should eventually improve accuracy, efficiency, and workflow in medical practices. This is especially true of the data inputs from sensors and sequencing, he said in an NVIDIA blog item on the subject.
Scripps is already a member of a unique data-driven effort known as the “All of Us Research Program,” which is led by the National Institutes of Health. This program, which collects data on more than 1 million US participants, looks at the intersection of biology, genetics, environment, data science, and computation. If successful, this research will expand the range of conditions that can be treated using precision medicine techniques.
NVIDIA, for its part, is positioned to play an important part in the initial wave of AI application rollouts. The company is a leader in producing performance chipsets popular with those who play high-end, processor-intensive gaming which it has recently applied to other processor intensive projects like blockchain. It now hopes its technology will form the core of systems designed to crunch the high volumes of data used in AI projects.
If NVIDIA can provide hardware that makes high-volume number-crunching less expensive and more efficient, it could establish an early lead in what is likely to be a very lucrative market. Given its focus on graphics processing, the hardware giant could be especially well-suited to dominate rapidly-emerging radiology AI applications.
We can certainly expect to see more partnerships like this file into place over the next year or two. Few if any IT vendors have enough scientific expertise in-house to make important gains in biotech AI, and few providers have enough excess IT talent available to leverage discoveries and data in this arena.
It will be interesting to see what AI applications development approaches emerge from such partnerships. Right now, much AI development and integration is being done on a one-off basis, but it’s likely these projects will become more systematized soon.