- Earthquakes can be devastating and ever since modern age, it was the hope to have a system able to accurately predict such disasters.
- A Chinese company by the name of Acoustic Electromagnetic To AI‘. may have found a solution
- In 2020, the top 10 teams achieved an accuracy rate of over 70% for the YES/NO hit-rate, high location accuracy, and magnitude.
Recently, a research team from Peking University began working on a system designed to forecast earthquakes days before they occur with promisingly accurate results. People have begun moving into the digital world of harnessing big data and training AI to greatly assist mankind.
The research team has named this project AETA, which stands for ‘Acoustic Electromagnetic To AI‘. The team embarked on this mission from 2010, after two devastating earthquakes hit Sichuan and Qinghai, affecting the lives of over 400,000 people.
Over the past 4 years, the AETA team has deployed over 300+ 3-part sensory systems, used to collect data of the acoustics and electromagnetic fields in earthquake-affected areas predominantly in the Sichuan region, presently over 40TB of data has been collected.
With this data, the team has been able to train their algorithms to sort through past data leading up to, during, and after an earthquake, teaching the algorithm to forecast future earthquakes using real-time data.
In 2020 the AETA team organized a 9 month competition, inviting Chinese universities, research centers, and students to participate. The AETA team shared all data gathered over the past 4 years, with a sheet of times earthquakes were detected. They then gave the teams access to live data and had the competitors submit their results.
The algorithm’s accuracy from each team are determined upon 3 key factors: Firstly, a YES/NO rate to whether an earthquake will take place, secondly, the epicenter of the earthquake, and thirdly, the earthquake’s magnitude. These 3 measurements determine a team’s success rate.
In 2020, the top 10 teams achieved an accuracy rate of over 70% for the YES/NO hit-rate, high location accuracy, and magnitude.
Currently, the AETA team launched a new competition for 2021, inviting the international community to register and participate. The 2021 competition registration is open and will last until the 31st of March.
The AETA’s hardware sensory system was developed by SVV, an innovation-focused hardware development, and manufacturing company. Furthermore, the AETA project has drawn the attention of CSDN, Capgemini, and numerous other institutions.
The AETA team, and partners, are adamant that we will solve the mystery behind forecasting earthquakes, and begin expanding this solution across the globe, saving millions of lives in the future.
China is a country with frequent earthquakes and widely distributed fault zones. Earthquakes, especially large earthquakes, can cause incalculable damage to people’s lives and property once they occur in densely populated areas without people’s awareness. It is very challenging and of great scientific value and social significance to carry out the research work of precursor observation, correlation analysis, precursor mechanism research and earthquake three-element prediction model around the solution of earthquake prediction and forecast problem.
The Earthquake Monitoring and Prediction Technology Research Center of Peking University Shenzhen Graduate School has developed a broadband electromagnetic disturbance and geo-acoustic monitoring system, named multi-component earthquake monitoring and prediction system AETA.
AETA, short for Acoustic & Electromagnetism To AI, the system includes:
- one acoustic sensor probe: to collect geo-acoustic data
- one electromagnetic sensor probe: to collect electromagnetic disturbance data
- one terminal device: connects to the two sensors by cable, for data process, temporary storage and uploading (via cable, wifi or 3/4G network)
- data storage: currently using AliCloud
From 2016, 300 sets have been deployed in some of the most earthquake-active areas in China, of which 240 sets in Sichuan/Yunnan and neighbouring provinces, and 60 sets in other areas. More systems will be deployed as soon as additional funding becomes available. Presently, 38TB of data has been collected, and 20GB of data is being collected daily. We discovered some signal characteristics related to earthquakes with imminent earthquake characteristics. Based on these findings, the imminent prediction of strong earthquakes was carried out. Although some progress has been made, the solution to the problem of distance earthquake prediction and prediction requires more analysis and research.
“AETA Earthquake Prediction AI Algorithm Competition” aims at mining the correlation between precursor observation data and earthquake three elements through innovative algorithms, discovering abnormal signals and features related to impending earthquakes, and building earthquake prediction models based on historical observation data and earthquake catalogue, in the hope of promoting the solution of scientific problems of earthquake prediction and forecast. At the same time, we also hope that through this competition, more attention and participation of people from all walks of life are involved and more novel technology and methods will be applied in earthquake prediction and forecast.
Problem and Data
Earthquake prediction for the coming week on each Sunday based on historic data from AETA network in Sichuan and Yunnan region. The magnitude of target earthquake should be equal or larger than 3.5. The target region is 22°N -34°N, 98°E -107°E. For an earthquake with a magnitude of 3.5 or larger in the target area, it will not be counted if there is no AETA station within 100km.
Training data for model construction
91 kinds of feature data of electromagnetic disturbance and geo-acoustic will be given to all teams. The time interval of each data is 10 minutes which is marked by timestamp. The specification of 91 kinds feature data will be specified in a read me file. The time period of data is from Oct. 1th 2016 to Dec. 31th 2020. In addition, an earthquake catalogue of ≥3.5 seismic events in the target region is also provided. Earthquake catalogue is from China Earthquake Network Center (CENC, http://news.ceic.ac.cn)
Real-time data for prediction
From Jan. 1th 2021, 91 kinds of feature data of electromagnetic disturbance and geo-acoustic will be updated each week. Teams can download data by week. There are two ways to download data. One is to download data from the website manually. The other one is to download data automatically through login the data server by an executable program which will be supplied by the host. The prediction of each week can also be submitted through two ways just like data download