Earthquake Location with GNN

    Earthquake Location is a very important problem in seismology, which is aimed to compute the location of seismic sources from the data recorded at seismic stations or geophones. However, traditional methods of earchquake location is costly, and requires manual work of picking seismic waves’ first arrivals and the velocity model of underground structures. In recent years, deep learning techniques are applied to earchquake location which partially solves problems mentioned above. But in these methods, the location information are not fully captured, which affects the accuracy of the output location. Therefore, we propose a new method which uses graph neural networks to do earchquake location.