Published

A video game created by NASA scientists not only lets gamers embark on a virtual ocean expedition of a coral reef, as they play their responses will help train a NASA supercomputer to better recognize corals from any image. It's called NeMO-NET which stands for Neural Multi-Modal Observation and Training Network.

A person plays the video game in which they classify images of coral
A user playing the NeMO-Net game that helps train a NASA supercomputer how to classify coral. Credits: NASA/Ames Research Center/Ved Chirayath

As game players classify the coral on the expedition, their answers feed back to the Pleiades supercomputer at NASA's Ames Research Center in Silicon Valley, helping the computer "learn" what's a coral, and what's not. The more people who play NeMO-NET, the better the supercomputer's mapping abilities will become. This is a type of "machine learning," as the computer gets more information on what is and is not a coral, it's better able to make that determination on its own.

Ecological Forecasting program area lead, Woody Turner, says this innovative technique is something he’ll be looking for in future proposals to the NASA Earth Applied Sciences program. “This extraordinary project combines revolutionary remote sensing technology, cutting-edge artificial intelligence, and strong citizen science in a true trifecta for our world’s threatened coral reefs.”

Learn more at the NASA.gov story, NASA Calls on Gamers, Citizens Scientists to Help Map World's Corals.

 

Related Impact