The Dell Technologies HPC Community Interviews: From Narrow to General AI – Decoding the Brain to Train Neural Networks

Print Friendly, PDF & Email

HPC veteran Luke Wilson lives at the forefront AI research. In this interview he tells us about research whose goal is to move computer intelligence from “narrow” (one task at a time) to “general” (more than one task simultaneously), a key to which is “context neuron switching.” One research strategy involves brain decoding by reverse mapping brain activity using “functional MRI activation maps,” said Wilson, who is chief data scientist and distinguished engineer at Dell’s HPC & AI Innovation Lab. Of the research conducted with McGill University, the Montreal Neurological Institute and Intel (using Dell’s Intel Xeon-powered Zenith cluster), he said: “What we’re trying to do is take that image of an activated brain and infer, using a neural network, what the patient was being asked to do.”