#medical devices#Health technology

Deep learning in healthcare: Deep 6 AI and MIT demo uses

The subset of AI known as deep learning, which uses artificial neural networks to teach machines, has myriad uses. From image recognition used in autono...

William Smith
|Feb 25|magazine5 min read

The subset of AI known as deep learning, which uses artificial neural networks to teach machines, has myriad uses. From image recognition used in autonomous vehicles to the natural language processing used to turn human speech into machine instructions, the technology has the potential to change our lives. Nowhere is this more evident than in the healthcare sector.

A newly released paper in the journal Cell outlines the discovery of a powerful antibiotic known as Halicin via the deployment of deep learning. The scientists involved trained a neural network to predict molecules with “antibacterial activity”, eventually finding Halicin, which differs in structure from conventional antibiotics. A further eight antibacterial compounds were also discovered. 

A separate but related development comes from Deep 6 AI, a member of the Nvidia Inception accelerator program for AI startups, making use of Nvidia’s GPUs, which are frequently employed for machine learning owing to their highly parallel structure.

Deep 6 AI specifically are involved in finding new methods of screening patients for clinical trials as medicine becomes increasingly targeted. “In the age of precision medicine, clinical trial criteria are getting more challenging,” CEO Wout Brusselaers said. “When developing a drug that is targeting patients with a rare genetic mutation, you have to be able to find those specific patients.”

Deep 6’s method involves processing medical records to identify and then label whatever criteria are relevant for the trial. Using natural language processing to gather relevant mentions, Deep 6 can also analyse unstructured data for additional information before creating a graph for doctors to peruse. The process has led to over 100,000 matches so far.