SPEAKER: Dr. Hongyu Sun (Caltech)

DATE: 4:00 pm, Friday, December 2nd

TITLE: Learning Seismic Waves for Subsurface Imaging

ABSTRACT: I will show applications of deep learning in solving challenges in seismic inversion and imaging. Deep neural networks are trained to generate seismic waves that weren’t originally recorded, but which are essential for subsurface imaging. I will discuss two examples with either active or passive seismic data. For active data, we develop deep-learning methods to extrapolate missing low-frequency waves from band-limited seismograms. The extrapolated low frequencies enable us to initiate full-waveform inversion from rough initial models. Moving from active to passive data, we develop deep-learning methods to extract accurate Green’s functions from realistic noise environments. The proposed method can potentially be used to overcome the temporal limitation of noise recording length for real-time monitoring and the spatial limitation of passive sources for the universal application of seismic interferometry.