https://sputnikglobe.com/20260705/russian-chinese-ai-algorithm-aims-to-improve-earthquake-forecasting-1124409654.html
Russian-Chinese AI Algorithm Aims to Improve Earthquake Forecasting
Russian-Chinese AI Algorithm Aims to Improve Earthquake Forecasting
Sputnik International
Researchers at Russia's Far Eastern Federal University (FEFU), with partners from China, have developed a new algorithm that could improve earthquake forecasting.
2026-07-05T12:41+0000
2026-07-05T12:41+0000
2026-07-07T12:41+0000
beyond politics
science & tech
russia
china
scientists
scientific research
https://cdn1.img.sputnikglobe.com/img/07e4/0a/1e/1080928854_0:395:1200:1070_1920x0_80_0_0_f644f8c13631f22523351c5c16f25058.jpg
The findings were presented at the Solar-Terrestrial Relations and Physics of Earthquake Precursors conference.The main method of spotting earthquake signs is based on monitoring sounds from deep within the Earth.Increased tectonic activity is accompanied by distinctive geoacoustic emissions—acoustic signals generated as underground rock formations build up huge tectonic stress, FEFU researchers explained."Normally, intense man-made noise and vibrations interfere with seismic observations," said Sergey Shevkun, Associate Professor at FEFU's Polytechnic Institute and one of the developers of the new program."Changes in that structure may indicate increasing seismic activity and could serve as one of the precursors of an earthquake," Shevkun said.The research team developed software that processes seismic observation data using Physics-Informed Neural Networks (PINNs) together with the Kolmogorov-Arnold Networks (KAN) neural network architecture."Traditionally, improving accuracy requires increasing the number of observations," Shevkun said.According to the researchers, the new approach reduces the computational resources needed to process acoustic signals while improving image resolution.That makes it possible to analyze larger subsurface volumes and generate more accurate 2D and 3D maps of geological structures.The research is still in the theoretical stage and has been validated through numerical simulations, which produced encouraging results.In the future, the technology could find applications in seismic exploration for oil, gas, coal and mineral deposits, as well as in assessing ground stability before the construction of major infrastructure projects.The study was supported by the Russian Ministry of Science and Higher Education.
https://sputnikglobe.com/20260629/new-earthquake-recorded-on-monday-in-venezuelas-la-guaira---seismologists-1124374677.html
russia
china
Sputnik International
feedback@sputniknews.com
+74956456601
MIA „Rossiya Segodnya“
2026
Sputnik International
feedback@sputniknews.com
+74956456601
MIA „Rossiya Segodnya“
News
en_EN
Sputnik International
feedback@sputniknews.com
+74956456601
MIA „Rossiya Segodnya“
Sputnik International
feedback@sputniknews.com
+74956456601
MIA „Rossiya Segodnya“
science & tech, russia, china, scientists, scientific research
science & tech, russia, china, scientists, scientific research
Russian-Chinese AI Algorithm Aims to Improve Earthquake Forecasting
12:41 GMT 05.07.2026 (Updated: 12:41 GMT 07.07.2026) Researchers at Russia's Far Eastern Federal University (FEFU), with partners from China, have developed a new algorithm that could improve earthquake forecasting. The computer program enables more accurate mapping of the Earth's crust while using much less computing power.
The findings were
presented at the Solar-Terrestrial Relations and Physics of Earthquake Precursors conference.
The main method of spotting earthquake signs is based on monitoring sounds from deep within the Earth.
Increased tectonic activity is accompanied by distinctive geoacoustic emissions—acoustic signals generated as underground rock formations build up huge tectonic stress, FEFU researchers explained.
"Normally, intense man-made noise and vibrations interfere with seismic observations," said Sergey Shevkun, Associate Professor at FEFU's Polytechnic Institute and one of the developers of the new program.
"We propose using an anthropogenic source to create additional seismic 'illumination' of a section of the Earth's crust, allowing us to reveal its internal structure," he added.
"Changes in that structure may indicate increasing seismic activity and could serve as one of the precursors of an earthquake," Shevkun said.
The research team developed software that processes seismic observation data using Physics-Informed Neural Networks (PINNs) together with the Kolmogorov-Arnold Networks (KAN) neural network architecture.
"Traditionally, improving accuracy requires increasing the number of observations," Shevkun said.
"Our method overcomes this limitation by combining observational data with established physical laws governing processes within the Earth's crust."
According to the researchers, the new approach reduces the computational resources needed to process acoustic signals while improving image resolution.
That makes it possible to analyze larger subsurface volumes and generate more accurate 2D and 3D maps of geological structures.
The research is still in the theoretical stage and has been validated through numerical simulations, which produced encouraging results.
In the future, the technology could find applications in seismic exploration for oil, gas, coal and mineral deposits, as well as in assessing ground stability before the construction of major infrastructure projects.
The study was supported by the Russian Ministry of Science and Higher Education.
