You'd think that the abilities of seismologists to detect earthquakes are already quite advanced; however, they are still unable to detect smaller earthquakes, much less predict them. The idea of using artificial intelligence to amplify the sensitivity of seismographs was proposed by Thibaut Perol and his colleagues when they were investigating the cause of a spike in the number of earthquakes in Oklahoma. Their idea proved to be effective since it is capable of detecting 17 times more earthquakes than older methods in only a fraction of the time. This technique is similar to the voice detection software used by digital assistants, like Alexa or Siri, that ignores background noise to listen to voice commands. The seismograph cancels out the natural rumbling of Earth to spot small or far away earthquakes. Perol says that "with this method, we are able to detect earthquakes of magnitude zero or minus one, and these are signals you wouldn't be able to see with a human eye." This is a huge step for seismologists because it increases the possibility of using AI to predict earthquakes by looking for patterns in data.
https://www.theverge.com/2018/2/14/17011396/ai-earthquake-detection-oklahoma-neural-network
No comments:
Post a Comment