Recovering Sub-threshold S-wave Arrivals in Deep Learning Phase Pickers via Shape-Aware Loss

1. 1Examine training dynamics to identify optimization traps
2. 2Analyze loss landscape geometry to understand amplitude suppression
3. 3Implement shape-aware loss functions to recover sub-threshold S-wave arrivals
4. 4Evaluate model performance on seismic data with varying S-wave amplitudes

Seismic data analysts and machine learning engineers can benefit from this research to improve the accuracy of phase picking models, particularly in cases where S-wave arrivals are ambiguous or have low amplitudes

💡 Optimization traps can cause deep learning phase pickers to miss sub-threshold S-wave arrivals, but shape-aware loss functions can help recover these arrivals

🔍 Improve seismic phase picking with shape-aware loss functions to recover sub-threshold S-wave arrivals