@article{fdi:010093784,
  title = {{D}eep inference of seabird dives from {GPS}-only records : performance and generalization properties},
  author = {{R}oy, {A}. and {B}ertrand, {S}ophie and {F}ablet, {R}.},
  editor = {},
  language = {{ENG}},
  abstract = {{A}t-sea behaviour of seabirds have received significant attention in ecology over the last decades as it is a key process in the ecology and fate of these populations. {I}t is also, through the position of top predator that these species often occupy, a relevant and integrative indicator of the dynamics of the marine ecosystems they rely on. {S}eabird trajectories are recorded through the deployment of {GPS}, and a variety of statistical approaches have been tested to infer probable behaviours from these location data. {R}ecently, deep learning tools have shown promising results for the segmentation and classification of animal behaviour from trajectory data. {Y}et, these approaches have not been widely used and investigation is still needed to identify optimal network architecture and to demonstrate their generalization properties. {F}rom a database of about 300 foraging trajectories derived from {GPS} data deployed simultaneously with pressure sensors for the identification of dives, this work has benchmarked deep neural network architectures trained in a supervised manner for the prediction of dives from trajectory data. {I}t first confirms that deep learning allows better dive prediction than usual methods such as {H}idden {M}arkov {M}odels. {I}t also demonstrates the generalization properties of the trained networks for inferring dives distribution for seabirds from other colonies and ecosystems. {I}n particular, convolutional networks trained on {P}eruvian boobies from a specific colony show great ability to predict dives of boobies from other colonies and from distinct ecosystems. {W}e further investigate accross-species generalization using a transfer learning strategy known as ?fine-tuning'. {S}tarting from a convolutional network pre-trained on {G}uanay cormorant data reduced by two the size of the dataset needed to accurately predict dives in a tropical booby from {B}razil. {W}e believe that the networks trained in this study will provide relevant starting point for future fine-tuning works for seabird trajectory segmentation.},
  keywords = {{PEROU} ; {PACIFIQUE} ; {PESCADORES} {ILES} ; {GUANAPE} {ILES} ; {FERNANDO} {DE} {NORONHA} {ARCHIPEL}},
  booktitle = {},
  journal = {{P}lo{S} {C}omputational {B}iology},
  volume = {18},
  numero = {3},
  pages = {e1009890 [18 ]},
  ISSN = {1553-7358},
  year = {2022},
  DOI = {10.1371/journal.pcbi.1009890},
  URL = {https://www.documentation.ird.fr/hor/fdi:010093784},
}