AI software may help eggheads predict where aftershocks will strike

Neural networks do a better job at guessing, it seems

Neural networks can predict where aftershocks will occur after earthquakes to a higher accuracy than standard techniques, according to a new study published in Nature on Wednesday.

After a major earthquake aftershocks can endanger the survivors and those seeking to help them. Now a group of researchers from Harvard University, the University of Connecticut and Google, have tried to figure this out by training neural networks to predict if an aftershock will occur at a particular position.

A simple feedforward neural network was trained using data taken from more than 131,000 pairs of main shocks and their aftershocks. The seismic wave patterns are split into grid cells measuring a fixed volume. The neural network learns to predict if an aftershock will happen at the center of each grid cell based on the distribution of stress caused by the earthquake in each region of space.

“By discretizing the volume around each mainshock in this way, aftershock forecasting can be formulated as a large-scale binary classification problem, with the goal of accurately classifying each 5 km × 5 km × 5 km grid cell in the volume around each mainshock as either ‘containing aftershocks’ or ‘not containing aftershocks,” the paper said.

The neural network was tested on 30,000 earthquake events and managed to determine the possibility of aftershocks more accurately than the Coulomb failure stress change method that is more commonly used. The old method inspects the change in stress to surrounding material after an earthquake event.


Boffins want machine learning to predict earthquakes


A perfectly accurate classifier has a score of one and a score no better than chance is 0.5. The neural network had an classifier of 0.849 compared to the Coulomb failure stress change of 0.583.

“The Coulomb failure stress change criterion has been widely used to explain aftershock locations. We show that across our testing data set, the neural network can perform better at explaining aftershock locations than Coulomb failure stress change,” Phoebe DeVries, a postdoctoral fellow at Harvard University, told The Register.

The model is still a prototype and doesn’t take into account different types of physical stresses so it won’t be used practically any time soon, but development is continuing and it's hoped that it can be deployed in the field by seismologists and rescuers.

“We view this as a motivating first step. There are a lot of different physical phenomena that may affect aftershock behavior,” DeVries said.

"We focused on static elastic stress changes in this work, but going forward, machine learning could be a very powerful tool for linking all kinds of physical phenomena [such as] dynamic stress changes, poroelastic stress changes, existing geological structures, etc. to aftershock behavior." ®

Similar topics

Broader topics

Other stories you might like

  • Venezuelan cardiologist charged with designing and selling ransomware
    If his surgery was as bad as his opsec, this chap has caused a lot of trouble

    The US Attorney’s Office has charged a 55-year-old cardiologist with creating and selling ransomware and profiting from revenue-share agreements with criminals who deployed his product.

    A complaint [PDF] filed on May 16th in the US District Court, Eastern District of New York, alleges that Moises Luis Zagala Gonzalez – aka “Nosophoros,” “Aesculapius” and “Nebuchadnezzar” – created a ransomware builder known as “Thanos”, and ransomware named “Jigsaw v. 2”.

    The self-taught coder and qualified cardiologist advertised the ransomware in dark corners of the web, then licensed it ransomware to crooks for either $500 or $800 a month. He also ran an affiliate network that offered the chance to run Thanos to build custom ransomware, in return for a share of profits.

    Continue reading
  • China reveals its top five sources of online fraud
    'Brushing' tops the list, as quantity of forbidden content continue to rise

    China’s Ministry of Public Security has revealed the five most prevalent types of fraud perpetrated online or by phone.

    The e-commerce scam known as “brushing” topped the list and accounted for around a third of all internet fraud activity in China. Brushing sees victims lured into making payment for goods that may not be delivered, or are only delivered after buyers are asked to perform several other online tasks that may include downloading dodgy apps and/or establishing e-commerce profiles. Victims can find themselves being asked to pay more than the original price for goods, or denied promised rebates.

    Brushing has also seen e-commerce providers send victims small items they never ordered, using profiles victims did not create or control. Dodgy vendors use that tactic to then write themselves glowing product reviews that increase their visibility on marketplace platforms.

    Continue reading
  • Oracle really does owe HPE $3b after Supreme Court snub
    Appeal petition as doomed as the Itanic chips at the heart of decade-long drama

    The US Supreme Court on Monday declined to hear Oracle's appeal to overturn a ruling ordering the IT giant to pay $3 billion in damages for violating a decades-old contract agreement.

    In June 2011, back when HPE had not yet split from HP, the biz sued Oracle for refusing to add Itanium support to its database software. HP alleged Big Red had violated a contract agreement by not doing so, though Oracle claimed it explicitly refused requests to support Intel's Itanium processors at the time.

    A lengthy legal battle ensued. Oracle was ordered to cough up $3 billion in damages in a jury trial, and appealed the decision all the way to the highest judges in America. Now, the Supreme Court has declined its petition.

    Continue reading

Biting the hand that feeds IT © 1998–2022