The role of facial resemblance in kin selection in mandrills:
contribution of artificial intelligence
Please feel free to contact me if you have any questions about my work
Places: CNRS-CEFE, Office 210 B / LIRMM, Bldg 5, Office 03/061
Supervisors: Julien Renoult (CEFE-E3CO), William Puech (LIRMM-ICAR) Marie Charpentier (ISEM-Evolutionary Anthropology)
Keywords: Behavioural ecology, Artificial intelligence, Mandrills (Mandrillus sphinx), Kin selection, Generative adversarial networks (GANs), Facial similarity.
Coming from a background in bioinformatics, my doctoral work focuses on the use of artificial intelligence (AI) , and in particular image generative techniques, to study the behavior of Mandrills (Mandrillus sphinx) via the Mandrillus Project (http://www.projetmandrillus.com/).
The idea is to understand and experimentally verify the role of facial similarity in mate selection. In order to control the confounding factors inherent to correlative studies, I will set up experimental studies presenting several artificially generated images (with GANs - Generative Adversarial Networks) varying according to a same variable to a population of Mandrills in Gabon, to test several hypotheses.
In addition to that, I am working on more fundamental aspects of cognitive science concerning the links between neural fluency and beauty perception in the brain.
Generally speaking, I am as much interested in the applications of AI to solve scientific problems as in the more fundamental and innovative aspects of its functioning.
This PhD is half of the time spent in the ICAR team of LIRMM, a computer science team specialized in image processing (http://www.lirmm.fr/icar/)
More informations on theses.fr: https://www.theses.fr/s311973
More information on Linkedin: https://www.linkedin.com/in/nicolas-dibot-203385178/
Erwan HarscouetCNRS - CEFE1919 Route de Mende34293 Montpellier France
The aim of my PhD project is to understand how birds nests evolve. The nest’s primary function is to provide protection for developing offspring. However nests can also be very conspicuous, decorated, or associated with the presence of individuals performing remarkable displays, when we could expect that camouflaged nests should be favored by natural selection. Thus it is probable that other reasons excluding offspring protection may act upon nest’s evolution! This project is essentially trying to answer this simple question: why do bird’s nests look the way they do?
To answer this question we study the Ploceidae family and we will try to determine :
1) Which factors contribute to the evolution and diversification of the nest?
2) Which evolutionary paths the nests follow when various species live in the same area?
3) Which are the links between the nest and acoustic and visual signalling?
Nicolas SilvaCNRS - CEFE1919 Route de Mende34293 Montpellier France
Mate choice in a monomorphic dull bird social species: a role for behavior ?
The objective of my thesis is to understand the mate choice in a monomorphic dull bird social species (Sociable weavers, Philetairus socius), with no extra pair copulation and no divorce. This research is supervised by Claire Doutrelant and Rita Covas (CIBIO, Porto, Portugal). The PhD is focusing on the potential role for sexual selection on the nest building behavior as well as the plumage. I analyse a large long term data rpoject dataset and also do a lot of field observations of the bird's behavior. My objectives are:
The ecological problematics are not my only interest, as I a part of my research focuses on artificial intelligence (Deep learning) methods en ecology science. During my PhD, I will implement several neural networks in order to, first automate the data extraction on pictures of phenotypic traits (bib and scaly feathers) and second partly analyses the nest building behavior.
Photos: Nicolas Silva
On going and soon hopefully ...
| Annuaire | Contacts | Intranet | Plan d'accès | Plan du site | CNRS | Crédits | Mentions légales |