Conference Paper

Deep learning for ethoacoustical mapping: Application to a single Cachalot long term recording on joint observatories in Vancouver Island

Authors:
  • Université de Toulon ; UMR CNRS LIS
Conference Paper

Deep learning for ethoacoustical mapping: Application to a single Cachalot long term recording on joint observatories in Vancouver Island

If you want to read the PDF, try requesting it from the authors.

Abstract

During February and March, 2018, a lone sperm whale known as Yukusam was recorded first by Orcalab in Johnstone Strait and subsequently on multiple hydrophones within the Salish Sea [1]. We learn and denoise these multichannel clicks trains with AutoEncoders Convolutional Neural Net (CNN). Then, we build a map of the echolocations to elucidate variations in the acoustic behavior of this unique animal over time, in different environments and distinct levels of boat noise. If CNN approximates an optimal kernel decomposition, it requires large amounts of data. Via spline functionals we offer analytics kernels with learnable coefficients do reduce it. We [1-3] identify the analytical mother wavelet to represent the input signal to directly learn the wavelet support from scratch by gradient descend on the parameters of cubic splines [2]. Supplemental material http://sabiod.org/yukusam [1] Balestriero, Roger, Glotin, Baraniuk, Semi-Supervised Learning via New Deep Network Inversion, arXiv preprint arXiv:1711.04313, 2017 [2] Balestriero, Cosentino, Glotin, Baraniuk, WaveletNet : Spline Filters for End-to-End Deep Learning, Int. Conf. on MachineLearning, ICML, Stockholm, http://sabiod.org/bib, 2018 [3] Spong P., Symonds H., et al., Joint Observatories Following a Single male Cachalot during 12 weeks—The Yukusam story, ASA 2018.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... Given the large magnitude of data, a key step is to build appropriate data storage and processing infrastructure, including automated ML pipelines (maintainable and reusable across multiple data collecting devices) that will replace the annotation currently done largely by hand by marine biologists. ML-based methods are already being used for detection and classification among marine mammals (Gillespie et al., 2009;Shiu et al., 2020) and for sperm whale click detection and classification (Bermant et al., 2019;Ferrari et al., 2020;Glotin et al., 2018;Jiang et al., 2018); such methods are potentially scalable to large datasets containing years of recording that would otherwise be beyond reach with previous manual approaches. ...
Article
Full-text available
Machine learning has been advancing dramatically over the past decade. Most strides are human-based applications due to the availability of large-scale datasets, however, opportunities are ripe to apply this technology to more deeply understand non-human communication. We detail a scientific roadmap for advancing the understanding of communication of whales that can be built further upon as a template to decipher other forms of animal and non-human communication. Sperm whales, with their highly-developed neuroanatomical features, cognitive abilities, social structures, and discrete click-based encoding make for an excellent model for advanced tools that can be applied to other animals in the future. We outline the key elements required for the collection and processing of massive datasets, detecting basic communication units and language-like higher-level structures, and validating models through interactive playback experiments. The technological capabilities developed by such an undertaking hold potential for cross-applications in broader communities investigating non-human communication and behavioral research.
... 60]; Glotin et al. used machine learning to study sperm whale bioacoustics[61,62]; Al-Barazanchi et al. used CNN not only to classify plankton images, but also to extend to new classifications[63]. ...
Article
Full-text available
In recent years, machine learning has become a hot research method in various fields and has been applied to every aspect of our life, providing an intelligent solution to problems that could not be solved or difficult to be solved before. Machine learning is driven by data. It learns from a part of the input data and builds a model. The model is used to predict and analyze another part of the data to get the results people want. With the continuous advancement of ocean observation technology, the amount of ocean data and data dimensions are rising sharply. The use of traditional data analysis methods to analyze massive amounts of data has revealed many shortcomings. The development of machine learning has solved these shortcomings. Nowadays, the use of machine learning technology to analyze and apply ocean data becomes the focus of scientific research. This method has important practical and long-term significance for protecting the ocean environment, predicting ocean elements, exploring the unknown, and responding to extreme weather. This paper focuses on the analysis of the state of the art and specific practices of machine learning in ocean data, review the application examples of machine learning in various fields such as ocean sound source identification and positioning, ocean element prediction, ocean biodiversity monitoring, and deep-sea resource monitoring. We also point out some constraints that still exist in the research and put forward the future development direction and application prospects.
... Par ailleurs, l'exploration de notre corpus de données ne fait que commencer. L'apprentissage par réseau profond et apprentissage de forme d'onde, d'un détecteur automatique de "downsweep" (ou signal modulé descendant) de grands rorquals est en cours d'élaboration en collaboration avec Franck Malige, Hervé Glotin et Randall Balestriero [51]. Ce détecteur devrait permettre de mieux comprendre ces unités vocales. ...
Thesis
Full-text available
La surveillance par acoustique passive est un outil d'importance croissante pour l'étude des mammifèresmarins. Cette thèse pose des nouveaux modèles pour l'étude du plus grand d'entre eux, la baleine bleue(Balaenoptera musculus). Pour ce faire, nous avons enregistré un corpus inédit dans l'archipel de Humboldt aunord du Chili. Nos données révèlent un chant caractéristique de la baleine bleue 'chilienne', dont nous étudionsla structure pulsée et l'évolution au cours des dernières décennies. Le classement en signal tonal ou non-tonalnous permet, en nous focalisant sur la fréquence fondamentale mise à jour, de quantifier la baisse en fréquencedes chants et d'effectuer une comparaison au niveau mondial. Notre troisième contribution est une méthode delocalisation mono-hydrophone basée sur des simulations de propagation par éléments spectraux. C’est à notreconnaissance le premier modèle de ce type, implémenté en milliers d’heures de calcul haute performance.
ResearchGate has not been able to resolve any references for this publication.