Multimedia Gallery
The Discovery FilesArtificial intelligence may help predict earthquakes and volcanoes
Researchers using deep learning have found subtle signals of an impending event were there in the overwhelming amount of data before a deadly Greenland landslide. With funding from the National Science Foundation, the team found small quake clusters can't hide from AI.
Credit: National Science Foundation/Karson Productions
Ready to rumble.
I'm Bob Karson with the Discovery Files, from the National Science Foundation.
(Sound effect: filtered voice) Greenland. June 17, 2017. 11:39 pm. (Sound effect: large rumbling sound) A massive landslide creates a tsunami (Sound effect: heavy surf) heading across a bay and devastating a small fishing village. Four are killed. Buildings and homes washed into the sea. Were there clues this was about to happen? Yes. Could they have been pieced together in time to make a prediction? Not by humans.
Researchers at Rice University believe the solution may be artificial intelligence. Their new study shows how we may someday be able to use deep learning -- AI -- to accurately predict seismic events like earthquakes (Sound effect: rumbling) or volcanic eruptions (Sound effect: volcanic eruption).
To test their algorithm, the team used data that had been collected by a nearby seismic station prior to the catastrophic event. The analysis revealed previously unseen, weak but repetitive rumblings that increased and accelerated starting nine hours before the slide. The kind of subtle signals AI can quickly identify in an overwhelming amount of data.
The team says although promising, the method is not ready to predict earthquakes just yet but marks the beginnings of a new approach -- applying machine learning to the study of geophysics.
I'd consider that (Sound effect: loud rumble) an 'earth-shaking' breakthrough.
"The discovery files" covers projects funded by the government's National Science Foundation. Federally sponsored research -- brought to you, by you! Learn more at nsf.gov or on our podcast.
Images and other media in the National Science Foundation Multimedia Gallery are available for use in print and electronic material by NSF employees, members of the media, university staff, teachers and the general public. All media in the gallery are intended for personal, educational and nonprofit/non-commercial use only.
Images credited to the National Science Foundation, a federal agency, are in the public domain. The images were created by employees of the United States Government as part of their official duties or prepared by contractors as "works for hire" for NSF. You may freely use NSF-credited images and, at your discretion, credit NSF with a "Courtesy: National Science Foundation" notation.
Additional information about general usage can be found in Conditions.