The Earth Science Department and the MSc in Exploration and Applied Geophysics organize the seminar titled

"Can we use and store geophysical knowledge in Neural networks?"

Lecturer: Tariq Alkhalifah

Date: May 23, 10:00, Room C

All are invited. For the students of the Master of Science in Exploration and Applied Geophysics: attending the seminar will contribute to obtain ECTS for the "Other Activities".

 Download the flyer here.

Short biography:

Tariq Alkhalifah is professor in the Physical Sciences and Engineering Division. He joined KAUST in June 2009. Before he joined KAUST, Alkhalifah was research professor and director at King Abdulaziz City for Science & Technology (KACST). Previously, he held the positions of associate research professor, assistant research professor, and research assistant at KACST. From 1996 to 1998, Alkhalifah served as a postdoctoral researcher for the Stanford Exploration Project at Stanford University, United States. He received the J. Clarence Karcher Award from the Society of Exploration Geophysicists (SEG) and the Conrad Schlumberger Award from the European Association for Geoscientists & Engineers (EAGE) in 2003, and honorary lecturer for the SEG in 2011. Alkhalifah received his doctoral degree in geophysics and his master’s degree in geophysical engineering, both from the Colorado School of Mines, United States. He holds a bachelor’s degree in geophysics from King Fahd University of Petroleum and Minerals, Saudi Arabia.

Short abstract:

Neural networks, and the process behind training them, have gained a lot of attention, even in the science community in recent years, resulting in the rise of a new breed of scientists, the “Data scientist”, with their motto: it is all in the data. Some of those data scientists have even “hypothesized” that they will not need our scientific theories and knowledge to solve physical problems any more. My task in this presentation is to hopefully convince you that training neural networks, which relies on a mix of statistics and inverse theory, can benefit from our geo/physical laws and a priori knowledge to guide us through the maze of degrees of freedom the neural networks may offer, and specifically to alleviate the many weaknesses/gaps/biases in data. This can be accomplished by instilling our physical laws and their corresponding characteristics into the neural network.