Xinming Wu

Xinming Wu

MSc	Geophysics
PhD	Geophysics
MSc university	Tongji University
PhD university	Colorado School of Mines

Xinming Wu joined the USTC (University of Science and Technology of China) as a professor in 2019, where he started the Computational Interpretation Group (CIG). Xinming received an engineering degree (2009) in geophysics from Central South University, an M.Sc. (2012) in geophysics from Tongji University, and a Ph.D. (2016) in Geophysics from the Colorado School of Mines where he was a member working with Dave Hale at the Center for Wave Phenomena. He interned twice at Transform Software and Services/DrillingInfo during the summer and winter of 2014. From 2016 to 2019, he was a postdoctoral fellow working with Sergey Fomel at Bureau of Economic Geology, The University of Texas at Austin.

Xinming received the J. Clarence Karcher Award from the Society of Exploration Geophysics (SEG) in 2020 and was selected to be the 2020 SEG Honorary Lecturer, South and East Asia. Xinming also received the SEG's awards for:

• Best Paper in Geophysics Award with David Hale in 2016, 3D seismic image processing for faults,
• Honorable Mentions (Best Paper, Annual Meeting) with Sergey Fomel in 2018,
• Best Student Poster Paper with Sean Bader and Sergey Fomel in 2017.

He also received the Shanghai excellent master thesis award in 2013 (Generating 3D seismic Wheeler volumes: methods and applications)..

Xinming writes a lot of Java packages for his research on seismic structural and stratigraphic interpretation, deep learning (e.g., FaultSeg), subsurface modeling, joint seismic and well-log interpretation, and geophysical inversion with geologic constraints.

Biography Citation for the J. Clarence Karcher Award 2020

Contributed by David Hale

I first met Xinming Wu in a remote video call eight years ago as he was applying for admission to the PhD program in geophysics at Colorado School of Mines. As we reviewed his undergraduate transcript, I asked him to explain low grades in two English courses. I learned that he had grown up in a farming community in China where English writing and speaking was not well taught. Those university courses were mostly his first ones involving these English skills.

Five years later in 2016, Xinming received SEG's award for the Best Paper in Geophysics. I contributed to the research and was fortunate to be a coauthor, but the inspiration and prose for that paper were all Xinming's.

Even before graduating from Mines in 2016, Xinming began working with others, first with Dean Witte at Transform, then with Guillaume Caumon of the RING consortium, but mostly with Sergey Fomel and geologists at the Texas Bureau of Economic Geology and at the University of Texas. There, Xinming began to develop applications of machine learning to seismic processing and interpretation.

A prerequisite for machine learning is a large collection of so-called “labeled data” — the inputs and desired outputs used to train the machines. Xinming was among the first to demonstrate that we can train them well using only synthetic data designed specifically for this purpose.

When he was still a student at Mines, Xinming once told me that he planned to submit two abstracts for the next SEG Annual Meeting. I remember thinking that submissions and rejection rates were both going up, and Xinming was not a well-known speaker. So I suggested that he instead choose the abstract he thought best and submit only that one; otherwise the program committee might choose only the other one. He ignored my advice, successfully, and in years since has continued to present multiple papers at SEG meetings. In 2020, as an Honorary Lecturer for SEG, Xinming speaks to audiences worldwide.

As a young professor at the University of Science and Technology in Hefei, China, Xinming now leads the new Computational Interpretation Group of graduate students and others who share research interests and computers with some serious GPUs. Let's watch what they do next!

2020 SEG Honorary Lecturer, South and East Asia

Deep learning for seismic processing and interpretation

Seismic interpretation involves detecting and extracting structural information, stratigraphic features, and geobodies from seismic images. Although numerous automatic methods have been proposed, seismic interpretation today remains a highly time-consuming task which still requires significant human efforts. The conventional seismic interpretation methods or workflows are not automated or intelligent enough to efficiently or accurately interpret the rapidly increasing seismic data sets, which leaves significantly more data uninterpreted than interpreted.

We improve automatic seismic interpretation by using CNNs (convolutional neural networks) which recently have shown the best performance in detecting and extracting useful image features and objects. One main limitation of applying CNNs in seismic interpretation is the preparation of many training data sets and especially the corresponding geologic labels. Manually labeling geologic features in a seismic image is highly time-consuming and subjective, which often results in incompletely or inaccurately labeled training images. To solve this problem, we propose a workflow to automatically build diverse geologic models with geologically realistic features. Based on these models with known geologic information, we further automatically create numerous synthetic seismic images and the corresponding ground truth of geologic labels to train CNNs for geologic interpretation in field seismic images. Accurate interpretation results in multiple field seismic images show that the proposed workflow simulates realistic and generalized geologic models from which the CNNs effectively learn to recognize real geologic features in field images.

In this lecture, I would like the share you with our research experience on the following topics:

1. Automatic preparation of training data sets and labels;
2. CNN for fault detection, fault orientation estimation, and fault surface construction;
3. CNN for relative geologic time and seismic horizons;
4. CNN for seismic geobody tracking;
5. CNN-based multitask learning in seismic interpretation.

Additional Resources

A recording of the lecture is available.

Listen to Xinming discuss his lecture in The present and future of seismic interpretation, Episode 72 of Seismic Soundoff, in-depth conversations in applied geophysics.

Honorable Mentions (Best Paper, Annual Meeting) 2018

Xinming Wu shares the 2018 Honorable Mentions (Best Paper, Annual Meeting) with Dr. Sergey Fomel

Best Poster Paper Presented by a Student at the Annual Meeting 2017

Xinming Wu shares the 2017 Best Poster Paper Presented by a Student at the Annual Meeting with Sean Bader and Dr. Sergey Fomel.

Best Paper in Geophysics 2016

Xinming Wu shares the 2016 Best Paper in Geophysics Award with David Hale.

External links

find literature about Xinming Wu