Back to Calendar
Dr. Yunyue Elita Li, an Assistant Professor at the National U. of Singapore and Research Affiliate in the MIT Earth Resources Laboratory, presents “Waveform inversion with gradient sampling – new understanding of the spatial shifts” at ERL.
“Waveform inversion (WI) with high frequency seismic data is highly non-convex in both the nonlinear (aka. full waveform inversion) and the linearized (aka. least-squares reverse time migration) scenarios. To mitigate the nonconvexity, we explore a wider model space using a modified gradient sampling algorithm (GSA) introduced for non-smooth, non-convex optimization. We show that straightforward physical intuitions that velocity changes induce location changes in the wavefield can be formalized to efficiently approximate the sampled source and scattered wavefields. A space-time domain implementation of WI enables extensive velocity sampling by shifting wavefields with different space lags at each time step. The resulting algorithm represents an implicit model expansion followed by an explicit model reduction and hence is less prone to the nonconvexity while maintaining the same cost as the original WI algorithms. Through numerical examples, we demonstrate that the simple physical-based modifications significantly improve the results of full waveform inversion and least-squares reverse time migration.”
Dr. Elita Li joined the Department of Civil and Environmental Engineering at the National University of Singapore as an assistant professor in 2016. Before coming to Singapore, Dr. Li did her postdoctoral research at Massachusetts Institute of Technology, holding a joint position in the Earth Resources Laboratory and the Department of Mathematics. Dr. Li received her Ph.D. and M.S. degrees in Geophysics from Stanford University in 2014 and 2010, respectively. She obtained her B.S. degree in Information and Computational Science from China University of Petroleum, Beijing in 2008. Her research interests include, but not limited to, seismic wave simulation, seismic inversion for structural imaging, subsurface characterization with deterministic and stochastic methods, data integration with multi-physical data, seismic monitoring for reservoir and civil purposes, and digital signal and image processing. Dr. Li’s paper on integrated VTI model building received “Honorable Mention” for Best Paper in Geophysics in 2016. Her presentations were ranked among “the Top Presentations at the SEG Annual Meeting” in 2016 and 2017. Dr. Li was the recipient of the J. Clarence Karcher Award from SEG in 2018.