Thread: Call for Abstracts SSA 2019 - Session, ”Machine Learning in Seismology”

Started: 2019-01-07 08:06:57
Last activity: 2019-01-07 08:06:57
Topics: SSA Meetings
Dear Colleagues,

We would like to draw your attention to our SSA session, “Machine Learning in Seismology.”

+. Submission Deadline: 11 January 2019 at 5 p.m. Pacific<x-apple-data-detectors://0>.
+. Submission Website: https://www.seismosoc.org/annual-meeting/program/
+. Session Title: Machine Learning in Seismology
+. Session Description:

Recent advances in computer science and data analytics have brought machine-learning (ML) techniques, including deep learning, to the forefront of seismological research. While ML methods continue to produce impressive successes in conventional artificial intelligence (AI) tasks, they also start to show powerful applicability in augmenting big data analysis in seismology by improving accuracy and efficiency compared to the traditional methods. Successful ML applications in seismology include seismic event detection, seismic signal classification, earthquake parameter estimation, signal denoising, ground motion prediction, subsurface tomography, aftershock pattern recognition and efficient visualization. However, challenges remain in terms of discovering new ML methods that can be applied to seismic and other geophysical data to learn about Earth’s subsurface structure and the underlying processes of Earth such as earthquakes. Furthermore, instead of considering ML models as “black boxes”, developing human-interpretable ML models and learning about their decision-making process also remain as grand challenges in ML field. The goal of this session is to highlight some of most recent ML results in our seismology community to motivate discussions of new ML research directions in seismology and beyond.

This session is jointly organized by the Seismological Society of Japan and SSA.

Please feel free to forward this announcement to colleagues you think may be interested.

Thank you in advance!

Session Conveners

Page built 00:45:15 | v.f0f5b0a4