Clustering earthquake signals and background noises in continuous seismic data with unsupervised deep learning

• Published in: Nature Communications
• Main Authors: Seydoux, Léonard, Balestriero, Randall, Poli, Piero, Hoop, Maarten De, Campillo, Michel, Baraniuk, Richard
• Other Authors: Institut des Sciences de la Terre (ISTerre), Institut de Recherche pour le Développement (IRD)-Institut national des sciences de l'Univers (INSU - CNRS)-Université Savoie Mont Blanc (USMB Université de Savoie Université de Chambéry )-Centre National de la Recherche Scientifique (CNRS)-Université Gustave Eiffel-Observatoire des Sciences de l'Univers de Grenoble (Fédération OSUG)-Université Grenoble Alpes (UGA), Electrical and Computer Engineering - Rice University, Rice University Houston, ANR-19-P3IA-0003,MIAI,MIAI @ Grenoble Alpes(2019), European Project: 789742335,F-IMAGE
• Format: Article in Journal/Newspaper
• Language: English
• Published: HAL CCSD 2020
• Subjects: [SDU.STU.GP]Sciences of the Universe [physics]/Earth Sciences/Geophysics [physics.geo-ph]; [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing; [STAT.ML]Statistics [stat]/Machine Learning [stat.ML]; [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]; Greenland; Nuugaatsiaq
• Online Access: https://hal.univ-grenoble-alpes.fr/hal-03411505; https://hal.univ-grenoble-alpes.fr/hal-03411505v2/document; https://hal.univ-grenoble-alpes.fr/hal-03411505v2/file/Preprint.pdf; https://doi.org/10.1038/s41467-020-17841-x

International audience The continuously growing amount of seismic data collected worldwide is outpacing our abilities for analysis, since to date, such datasets have been analyzed in a human-expert-intensive, supervised fashion. Moreover, analyses that are conducted can be strongly biased by the standard models employed by seismologists. In response to both of these challenges, we develop a new unsupervised machine learning framework for detecting and clustering seismic signals in continuous seismic records. Our approach combines a deep scattering network and a Gaussian mixture model to cluster seismic signal segments and detect novel structures. To illustrate the power of the framework, we analyze seismic data acquired during the June 2017 Nuugaatsiaq, Greenland landslide. We demonstrate the blind detection and recovery of the repeating precursory seismicity that was recorded before the main landslide rupture, which suggests that our approach could lead to more informative forecasting of the seismic activity in seismogenic areas.