Offre en lien avec l’Action/le Réseau : – — –/– — –

Laboratoire/Entreprise : UBO/IUEM/LGO
Durée : 24
Contact : sara.bazin@univ-brest.fr
Date limite de publication : 2021-05-07

Contexte :
Nous avons soumis un sujet de post-doc Marie-Curie pour développer des algorithmes de reconnaissance automatique par apprentissage machine (ML) de signaux de données hydroacoustiques (séismes notamment). Nous disposons de 10 ans d’enregistrements du réseau OHASISBIO dans l’océan indien (https://www-iuem.univ-brest.fr/lgo/les-chantiers/ohasisbio/).
Lien vers l’annonce Euraxess : https://euraxess.ec.europa.eu/jobs/623070
Les candidat.e.s devront avoir passé au moins 12 mois à l’étranger lors des trois dernières années.

Sujet :
Mooring networks of autonomous hydrophones is an effective way for monitoring the ocean soundscape and its sources: undersea earthquakes and volcanic eruptions, marine mammals, iceberg cracks, sea-state, ship noise… For more than 10 years, our laboratory has been maintaining hydroacoustic networks in the open ocean, composed of few hydrophones moored in the sound channel, which acts as an acoustic waveguide, carrying acoustic waves over thousands of kilometers.
In the Indian Ocean, the OHASISBIO network comprises 7 to 9 distant hydrophones continuously recording low-frequency sounds (0-120Hz) since 2010. Its objective is to monitor the seismic activity of mid-ocean ridges, but also the presence and migration patterns of large whales, and the oceanic ambient noise in general. Indeed, mid-oceanic spreading centers generate a large number of earthquakes and thus acoustic waves, indicative of the intervening seafloor spreading processes. Moreover, large baleen whales produce many loud and distinctive calls and songs, which provides clues as to when and where species dwell and migrate. Other sounds of interest are cryogenic sounds produced by icebergs or man-made noises (ship traffic, seismic exploration).
Over the years, passive acoustic monitoring of the ocean results in very large data sets (e.g. 25G/yr/instrument x 10 instr. x 10 years). The preliminary but indispensable, and time consuming step in the data analysis consists in identifying the different types of acoustic events. To achieve a more complete and efficient analysis, we wish to develop a deep learning application for event detection and signal discrimination in our acoustic database.
The fellow will hence develop an automatic detection and classification tool for acoustic signals recorded in the ocean, based on machine learning techniques. Among the wide range of approaches for intelligent classification, we seek for the implementation that would best extract information from our growing acoustic database.
Supervised learning consists in teaching a model how to make classification predictions, here: earthquake, icequake, seismic-shot or whale-call. Parts of the OHASISBIO dataset have already been manually processed and classified, and will serve for training the model.
Once events are detected and classified on several hydrophones, their source can be localized based on their arrival times, the geometry of the network, and the sound-speed in the ocean. Ultimately, resulting seismic catalogs will depict the spatial and temporal seismicity that will help understanding the dynamics of seafloor spreading. Bioacoustic catalogs will be used to establish statistics on the presence of marine mammals and its evolution over the years, a key to developing conservation measures. Ocean noise pollution by man-made noise is becoming a major issue and its evolution has yet to be characterized in the long term. These are among the outcomes expected from a thorough, systematic and enhanced analysis of continuous acoustic recordings in the open ocean, through machine learning techniques.

Profil du candidat :
Applicants must have a maximum of 8 years full-time equivalent experience in research, measured from the date applicants were in possession of a doctoral degree. Years of experience outside research and career breaks (e.g. due to parental leave), will not be taken into account.
Nationality & Mobility rules: Applicants can be of any nationality but must not have resided more than 12 months in France in the 36 months immediately prior to the MSCA-PF call deadline (September 15th, 2021)

Formation et compétences requises :
– Skills in Machine Learning algorithms and their implementation
– Skills in large dataset analysis and signal processing
– Post-doctoral publication(s) in peer reviewed journals, related to these fields
– Experience in acoustics or geoscience will be a plus
– Required Language: English (French is not required)

Adresse d’emploi :
Institut Universitaire Européen de le Mer (Université de Brest), Plouzané, France