AI will help predict earthquakes. It's already very efficient

Artificial Intelligence (AI) developed by scientists at the University of Texas in Austin, over the course of seven months, was able to successfully predict 70% of earthquakes in the vicinity of China a week before they occurred - such information was conveyed by the "Bulletin of the Seismological Society of America".

A team of scientists from the University of Texas at Austin developed an algorithm that utilizes a set of statistical functions based on their understanding of the physics of earthquakes. The artificial intelligence that the algorithm is based on was trained using a five-year seismic data database. Then, the AI "listened" to signals received by seismographs to detect signs of impending earthquakes.

Thanks to this process, artificial intelligence was able to create weekly forecasts. As a result, AI successfully predicted 14 earthquakes within a radius of over approximately 200 miles, almost accurately predicting the strength of the earthquakes. One earthquake was overlooked by AI, and it sent false warnings eight times.

This experiment was part of an international competition organized in China, in which artificial intelligence developed by the University of Texas in Austin won first place out of 600 other projects.

Although it is not yet known whether the same approach will be effective in other locations, the authors are convinced that the obtained results are a significant step forward in earthquake forecasting research using artificial intelligence. Scientists from Texas hope that technology based on artificial intelligence can be used to minimize the impact of earthquakes on human life and the economy.

Experts believe that in places equipped with appropriate seismic monitoring networks, such as California, Italy, Japan, Greece, Turkey, and Texas, artificial intelligence can improve its effectiveness rate and narrow earthquake location predictions to a few dozen miles.

Sergey Fomel, a professor at the University of California and member of the research team, stated: "Predicting earthquakes is the Holy Grail. We are not yet close to predicting for any place in the world, but what we have achieved indicates that what we considered an impossible problem to solve can essentially be solved."

The next step in the research is to test the artificial intelligence in Texas, where there are many earthquakes of low and medium strength. The TexNet network has 300 seismic stations, and there are also over six years of continuous records, making Texas an ideal place to verify the method.