Waveform inversion is highly sought after for its potential to produce a high-resolution elastic model of the Earth. In spite of this potential, waveform inversion faces many challenges, chief among them are robustness and cost. These challenges are directly related to the size of our data and the high nonlinearity in the relation between the model and such data. The machine has proved its ability to learn how to deal with large data and most importantly how to infer dependencies between inputs to outputs even for nonlinear problems. In this presentation, I will show examples in which deep learning helped improve waveform inversion’s ability to produce a convergent model, short of trying to use it to predict the physics of wave propagation. Thus, we will still use the wave equation to generate predicted data to compare with the observed ones. However, within the framework of waveform inversion, we use an ML-misfit as a metric to measure the distance between the observed and predicted data to avoid cycle skipping, we use an ML-decent to help converge faster to an accurate solution, and we use an ML-regularization to help mitigate the NULL space. If the machine, for now, cannot directly find the model that corresponds to particular observed data, it can, at least, help us find it using waveform inversion. I will specifically share our experiences in what we managed to teach the machine? What could we not teach it? And more importantly, what it taught us?
Speaker Bio
Tariq Alkhalifah
Tariq A. Alkhalifah is a professor of geophysics in the division of Physical Sciences and Engineering at King Abdullah University for Science and Technology (KAUST). He assumed his duties there in June 2009. Prior to joining KAUST, Tariq was a research professor and director of the Oil and Gas Research Institute at King Abdulaziz City for Science & Technology (KACST). He has also been associate research professor, assistant research professor and research assistant at KACST. From 1996 to 1998, Tariq served as a postdoctoral researcher for the Stanford Exploration Project at Stanford University, USA. He received the J. Clarence Karcher Award from the Society of Exploration Geophysicists (SEG) in 1998 and the Conrad Schlumberger Award from the European Association for Geoscientists and Engineers (EAGE) in 2003. He is a member of SEG and EAGE. Tariq received his doctoral degree in geophysics (1997) and master’s degree (1993) in geophysical engineering from the Colorado School of Mines, USA. He holds a bachelor’s degree (1988) in geophysics from King Fahd University of Petroleum and Minerals, Saudi Arabia. Tariq’s research interests are in imaging, inversion and velocity model building for exploration seismic data with special emphasis on media that exhibit anisotropic behavior.