Zeyu’s research focuses on developing efficient and accurate seismic imaging and inversion methods. The research topics include efficient reverse time migration (RTM) methods using compressed dataset, for instance plane wave and beam data, frequency dependent AVO analysis and full waveform inversion (FWI) techniques. Recently, he has been focusing on 1) developing a hybrid optimization workflow for FWI to tackle the local minima issues related to highly nonlinear inverse problems, making the inversion heuristically independent from the starting models. 2) developing efficient MCMC nonlinear sampling methods for large-scale FWI to quantify the uncertainties related to the ill-posed inverse problems.
exploration geophysics, seismic imaging, seismology, inverse problems, high performance computing