NVIDIA is cutting into a huge market called "medical" with weapons of deep learning

NVIDIA, the biggest graphic board, has a GPU with powerful parallel computing performanceMachine learningWe are developing new technologies in various fields such as artificial intelligence, automatic driving car, image recognition technology, etc. by applying it to the system. Meanwhile, NVIDIA has been using deep learning technology as a weapon in a huge market called "medical".

Nvidia's Deep-Learning Chips May Give Medicine a Shot in the Arm - MIT Technology Review

It was held in San Francisco, USAEmTech Digital conference"Kimberly Powell, led by NVIDIA's health management field, said," There is a surprisingly rapid increase in demand in medical imaging research. " NVIDIA has been promoting the development of image recognition technology utilizing the powerful parallel computing performance of GPU, but among them,Deep learningIt is said that the adoption of image recognition technology making use of is rapidly expanding.

Already, Google's deep learning research teamTechnology to detect diabetic patients from images of eyesThe scientific journal of deep learning technology which establishment of the diagnosis of skin cancer by the research team of Stanford UniversityAnnounced at NatureHowever, according to Mr. Powell, NVIDIA has dispatched engineers to many hospitals in the US, and joint development of applications using artificial intelligence It seems to be active actively.

NVIDIA 's GPU processor can process parallel operations required for deep learning at high speed, so it is used by academic researchers and researchers using deep learning in industry. Accelerator for deep learning training already "TESLA P 100", A supercomputer specialized for deep learning"DGX-1"For automatic operation"Drive PXSeries, etc. We are planning to expand products for deep learning in the future. "NVIDIA hardware can increase the reliability of doctor's diagnosis and can significantly increase the level of medical care in developing countries where expertise is insufficient," NVIDIA products in the medical field I point out the usefulness of.

Deep learning which gives great results to image recognition shows great power even in finding patterns that human medical doctors can not see. NVIDIA has successfully identified genetic factors related to brain diseases from brain images in cooperation with Mayo Clinic's neuroradiation specialist Bradley Ericsson. It shows the prospect that it will be increasingly adopted at hospitals and medical research centers in the future.

Professor Gary Marcus of New York University says, "A field where artificial intelligence may have the greatest impact is" medicine ", for example the possibility that algorithms can identify risk factors likely to become cancer There is ".

Revolutionary achievements that deep learning technology brings to medical care have been pointed out, but technological complexity has become a bottleneck. Creating a new way to visualize the behavior of a deep neural network to help doctors to leverage deep learning for medical care has become a big issue for NVIDIA selling deep learning GPU processors It seems to be.

in Software,   Hardware,   Science, Posted by darkhorse_log