When applied to data sets taken from the Bai Dai area, the algorithms detected significantly more earthquakes and made it easier to figure out how and where they started. And when applied to data from In the 2014 La Habra earthquake, also in California, the team observed more than four times the number of seismic findings in the “clarified” data than in the officially recorded numbers.
This is not the only work applying AI to earthquake hunting. Researchers from Penn State have been training deep learning algorithms to accurately predict how changes in measurement might occur. Indicate upcoming earthquakes—A quest that has puzzled experts for centuries. And members of the Stanford team have previously trained models to choose the phase or measure the arrival time of seismic waves in an earthquake signal, which can be used to estimate the location of an earthquake.
Paula Koelemeijer, a seismologist at London’s Royal Holloway University who was not involved in the study, says deep learning algorithms are particularly useful for earthquake monitoring because they can ease the burden for human seismologists.
In the past, seismologists would look at graphs produced by sensors that record ground movement during an earthquake, and they would identify patterns by eye. Deep learning can make that process faster and more accurate, by helping to slice through large volumes of data, Koelemeijer said.
“Shows [the algorithm] Working in noisy urban environments is very helpful, as noise in urban environments can be a nightmare and very challenging,” she says.
