View Full Version : Earthquake prediction using AI

Roger E. Moore
2018-Aug-30, 08:08 PM

Artificial intelligence nails predictions of earthquake aftershocks

Alexandra Witze
29 August 2018

A neural-network analysis outperforms the method scientists typically use to work out where these tremors will strike.

The scientists looked at more than 131,000 mainshock and aftershock earthquakes, including some of the most powerful tremors in recent history, such as the devastating magnitude-9.1 event that hit Japan in March 2011. The researchers used these data to train a neural network that modelled a grid of cells, 5 kilometres to a side, surrounding each main shock. They told the network that an earthquake had occurred, and fed it data on how the stress changed at the centre of each grid cell. Then the scientists asked it to provide the probability that each grid cell would generate one or more aftershocks. The network treated each cell as its own little isolated problem to solve, rather than calculating how stress rippled sequentially through the rocks.

When the researchers tested their system on 30,000 mainshock-aftershock events, the neural-network forecast predicted aftershock locations more accurately than did the usual stress-failure method. Perhaps more importantly, DeVries says, the neural network also hinted at some of the physical changes that might have been happening in the ground after the main shock. It pointed to certain parameters as potentially important — ones that describe stress changes in materials such as metals, but that researchers don’t often use to study earthquakes.

2018-Aug-31, 03:47 PM
I wonder how much better the neural-network analysis did versus conventional stress analysis, which, as the article states, isn't very reliable.

2018-Sep-01, 08:35 AM
Interesting article, but the thread title is misleading. This is aftershock prediction, not that of initial earthquakes.

The reference abstract linked at the end of the article says,

We show that a neural network trained on more than 131,000 mainshock–aftershock pairs can predict the locations of aftershocks in an independent test dataset of more than 30,000 mainshock–aftershock pairs more accurately (area under curve of 0.849) than can classic Coulomb failure stress change (area under curve of 0.583).


I don't know how to convert that to lay-speak "how much better than". Does anyone else?


2018-Sep-01, 09:12 AM
I also noticed it is a prediction of where aftershocks will happen, not when. Knowing both when and where would be useful.

Sent from my iPhone using Tapatalk

2018-Sep-03, 05:05 PM
Good read, I noticed it mentioed that even the AI was dropping hints that the trigger mechnism for quakes are hugely complex.