AtomNet Netexplo 100 2018

Artificial intelligence for drug discovery
Developed by the start-up Atomwise, AtomNet is a drug discovery algorithm that uses a deep neural network. It tests interactions between thousands of potential drugs and a pathology (like cancer) to identify effective substances by simulating molecular interactions. Molecules identified as potentially effective by AtomNet can then be tested as a priority by research laboratories (because the algorithm does not assure their non-toxicity). A promising result was obtained on the Ebola virus.

The artificial intelligence here helps human researchers or competes with them. In a human intelligence-digital intelligence collaboration, this innovation presents a real benefit since it could accelerate research for treatments and enable researchers to focus on testing phases.
The score is 3