Machine Learning

By utilizing the high-quality PmP dataset (10,192 manual picks by Li et al., 2022) in southern California, we further develop PmPNet (Ding et al., 2022), a deep-neural-network-based algorithm to automatically identify PmP waves efficiently. PmPNet applies similar techniques in the machine learning community to address the unbalancement of PmP datasets. The trained optimal PmPNet can efficiently achieve high precision and high recall simultaneously to automatically identify PmP waves from a massive seismic database.

PmPNet training flow: (i) one batch of data points is fed into PmPNet, and the loss between the PmPNet output and the true labels is computed; and (ii) the optimizer reads in the loss and update the trainable parameters of PmPNet. One epoch of training consists of continuing this iteration until the whole dataset has been tranversed. The training phase for PmPNet is complete when the pre-selected maximum number of epochs is reached.

Tip

Codes of PmPNet and PmP-traveltime-Net can be downloaded at Zenodo.

Example data can be downloaded at OSF.

• PmPNet Structure
• Loss Function of PmPNet
• Performance of PmPNet
• Use PmPNet to Identify PmP Phase
• Use PmP-traveltime-Net to Predict PmP Traveltime