Garrett Kenyon’s Research Highlighted in Discover Magazine


February 2018, Discover Magazine featured an article on Garrett Kenyon and his team’s neuromorphic computing research titled, Computers Learn to Imagine the Future. Garrett Kenyon, a Los Alamos National Laboratory (LANL) and New Mexico Consortium (NMC) scientist does research on creating computers that can process things and learn in the same way the way as the human cerebral cortex. Other members of this project include Boram Yoon of the Applied Computer Science group at LANL, and Peter Schultz of the NMC.

In working to create artificial intelligence, scientists have their work cut out for them. So far, no computer can match the computing power of the human brain. The simplest tasks, such as distinguishing one object from another, or being able to tell a moving car from a static background and predict where the car will be in the next half second, are complex challenges for a computer.

Kenyon and his team’s research are changing all this. By using supercomputers, which possess a staggering amount of computational power, Kenyon has been able to simulate biological neural networks, and the result is that these machines are now able to learn about their surroundings, interpret data and make predictions in much the same way as the human brain.

In testing the ability to model neural processing, these researchers have created a “sparse prediction machine” that operates a neural network on the Trinity supercomputer. This machine is designed to work like a human brain. Researchers expose it to data, or video clips, and the machine learns about the visual world simply by watching thousands of video sequences, similar to how a human child learns about their world. Eventually, the sparse prediction machine was able to predict what would happen next in each video. The computer could imagine the future.

To see the full article:

Discover Magazine: Computers Learn to Imagine the Future

© 2018 New Mexico Consortium