WGMLEARN Report 2019.pdf (424.88 kB)
Download file

Working Group on Machine Learning in Marine Science (WGMLEARN)

Download (424.88 kB)
posted on 2019-01-01, 00:00 authored by ICESICES

The working group on machine learning in marine science (WGMLEARN) is tasked with charting the current status and exploring the potential for the use of machine learning methods in the various fields of marine science. The group members’ presentations covered primarily computer vision for classification problems, derivation of new variables from remotely sensed data, and inference regarding species interactions. Those topics were complemented by three invited presentations: Tristan Cordier on genomics, Cedric Jamet on remote sensing, and Periklis Panagiotidis from the ICES Data Centre. In future meetings, we will strive to cover topics not covered this year (acoustics, fishing effort, etc.).

We started to assemble a comprehensive literature database to document all applications of machine learning in marine sciences, in particular in relation to the ecosystem approach to fisheries. It will serve as the basis for a review paper; both the database and the paper will help the numerous scientists interested in applying these relatively new techniques to their questions to get a broad and exhaustive overview of prior work. It will also highlight active topics and future research questions.

Approximately 500 papers were registered, covering various data types (acoustics, imaging, etc.), machine learning techniques (classic learning, deep learning, etc.), and topics (stock assessment, biogeochemistry, etc.). They are now in the process of being tagged according to these three categories, to facilitate searching the database. Topics for which the members present this year did not have sufficient expertise were identified and assigned to other group members, known to be interested and competent. An early outline for the review article was drafted, based on the distribution of topics for which papers were found.

A recurring theme was the need for training of marine scientists in the relatively new field of machine learning. For this purpose, possible new directions were discussed, including the creation and maintenance by members of the group of an online list of relevant conferences and training options (such as video lectures and MOOC courses) or the organisation of dedicated ICES training courses.


Published under the auspices of the following ICES Steering Group or Committee

  • EOSG

Published under the auspices of the following ICES Expert Group or Strategic Initiative



ICES Scientific Reports





Contributors (Editors)

Ketil Malde; Jean-Olivier Irisson

Contributors (Authors)

Alessandra Gomes; Bernhard Keuhn; Cédric Jamet; Edwin Van Helmond; Jean-Baptiste Romagnan; Jose Fernandes; Julien Simon; Laura Hoebeke; Laura Uusitalo; Linh Nguyen; Madiop Lo; Michiel Stock; Perikilis Panagiotidis; Rainer Kiko; Raphaelle Sauzede; Rubens Lopes; Tristan Cordier; Vincent Rossi; William Michaels



Recommended citation

ICES. 2019. Working Group on Machine Learning in Marine Science (WGMLEARN). ICES Scientific Reports. 1:45. 13 pp. http://doi.org/10.17895/ices.pub.5539