Sign In

Communications of the ACM

ACM TechNews

Artificial Networks Learn to Smell Like the Brain

View as: Print Mobile App Share:
Artist's conception of an electronic nose.

When asked to classify odors, artificial neural networks adopt a structure that closely resembles that of the brains olfactory circuitry.

Credit: Mark Hooper

A team of researchers from the Massachusetts Institute of Technology (MIT) and Columbia University found a machine learning model can train itself to smell by building an artificial neural network that mimics the brain's odor-processing olfactory circuits.

The researchers used the fruit fly's olfactory system as a template, building an artificial network comprised of an input layer, a compression layer, and an expansion layer; links between neurons would be rewired as the model learned to classify smells.

The network self-organized in minutes into a structure closely resembling the fly brain's olfactory network.

MIT's Guangyu Robert Yang said, "By showing that we can match the architecture [of the biological system] very precisely, I think that gives more confidence that these neural networks can continue to be useful tools for modeling the brain."

From MIT News
View Full Article


Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA


No entries found