Preprints
https://doi.org/10.5194/nhess-2024-102
https://doi.org/10.5194/nhess-2024-102
17 Jun 2024
 | 17 Jun 2024
Status: this preprint is currently under review for the journal NHESS.

How can seismo-volcanic catalogues be improved or created using robust neural networks through weakly supervised approaches?

Manuel Titos, Carmen Benítez, Milad Kowsari, and Jesús M. Ibáñez

Abstract. Real-time monitoring of volcano-seismic signals is complex. Typically, automatic systems are built by learning from large seismic catalogues, where each instance has a label indicating its source mechanism. However, building complete catalogues is difficult owing to the high cost of data-labelling. Current machine learning techniques have achieved great success in constructing predictive monitoring tools; however, catalogue-based learning can introduce bias into the system. Here, we show that while monitoring systems recognize almost 90 % of events annotated in seismic catalogues, other information describing volcanic behavior is not considered. We found that weakly supervised learning approaches have the remarkable capability of simultaneously identifying unannotated seismic traces in the catalogue and correcting mis-annotated seismic traces. When a system trained with a master dataset and catalogue is used as a pseudo-labeller within the framework of weakly supervised learning, information related to volcanic dynamics can be revealed and updated. Our results offer the potential for developing more sophisticated semi-supervised models to increase the reliability of monitoring tools. For example, the use of more sophisticated pseudo-labelling techniques involving data from several catalogues could be tested. Ultimately, there is potential to develop universal monitoring tools able to consider unforeseen temporal changes in monitored signals at any volcano.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.
Manuel Titos, Carmen Benítez, Milad Kowsari, and Jesús M. Ibáñez

Status: open (until 19 Aug 2024)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
Manuel Titos, Carmen Benítez, Milad Kowsari, and Jesús M. Ibáñez
Manuel Titos, Carmen Benítez, Milad Kowsari, and Jesús M. Ibáñez

Viewed

Total article views: 272 (including HTML, PDF, and XML)
HTML PDF XML Total Supplement BibTeX EndNote
220 33 19 272 27 7 7
  • HTML: 220
  • PDF: 33
  • XML: 19
  • Total: 272
  • Supplement: 27
  • BibTeX: 7
  • EndNote: 7
Views and downloads (calculated since 17 Jun 2024)
Cumulative views and downloads (calculated since 17 Jun 2024)

Viewed (geographical distribution)

Total article views: 256 (including HTML, PDF, and XML) Thereof 256 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 20 Jul 2024
Download
Short summary
Developing seismo-volcanic monitoring tools is crucial for Volcanic Observatories. Our study reviews current methods using Transfer Learning techniques and finds that while these systems identify nearly 90 % of seismic events, they miss other important volcanic data due to the catalogue-learning bias. We propose a weakly supervised technique to reduce bias and uncover new volcanic information. This method can improve existing databases and create new ones efficiently using machine learning.
Altmetrics