[Comp-neuro] Postdoc position in Deep Learning and Animal Behavior modellig

david robbe david.robbe at inserm.fr
Tue Jan 23 09:44:48 CET 2018


Dear Colleagues,

The Turing Centre for Living Systems in Marseille seeks applicants for a 
postdoc position in computational
neuroscience/robotics under the joint supervision of David Robbe and 
Christophe Eloy.

Project abstract – The goal of this project is to understand, at an 
algorithmic level, how animals learn
to develop adaptative behavior by trial-and-error interaction with their 
environment. We propose an
interdisciplinary approach combining behavioral experiments in rats and 
the development of an «artificial»
agent facing challenges similar to those experienced by the rats. The 
experimental data (already available)
show that rats challenged in a time estimation task converged 
progressively toward a conserved embodied
strategy (Rueda-Orozco and Robbe, Nature Neuroscience, 2015). Key 
determinants of that strategy have
been isolated experimentally, by altering task rules and environment and 
quantifying how these alterations
affect animal performance. The core of the artificial agent will be a 
reinforcement-learning algorithm, which
will be trained through virtual experiments simulating the time 
estimation task. A first objective will be to
parametrize algorithms such that they solve this time estimation task 
and to compare artificial and natural
(rat) strategies. The artificial double will eventually be complemented 
by a robot embodying the learning
algorithm. Assessment of algorithmic predictions and possible biological 
implementation will be tested in
new experiments in which the impact of alterations of key task 
parameters and brain activity on animals
performance will be measured. This back and forth interaction between 
experiments and theory will provide
us with new understandings of the mechanisms underlying adaptative motor 
learning.

Expected profile – The main task of the project will be to develop 
reinforcement learning algorithms and
therefore we expect the candidates to have a solid background in machine 
learning. Since one of the
objective is to build simple robots, knowledge in robotics would be an 
asset. The successful applicant will
work in close collaboration with experimentalists and robotic engineers. 
We are therefore looking for an
enthusiastic applicant with excellent communication skills and a track 
record in interdisciplinary collaborative
environments.

Supervisors
David Robbe - Inmed, UMR 1249 - The neural bases of sensorimotor 
learning 
<http://www.inmed.fr/en/en-avenir-dynamiques-neuronales-et-fonctions-des-ganglions-de-la-base>
Christophe Eloy - IRPHE, UMR 7342 - Biomechanics, aerodynamics, soft 
matter and biophysics <https://www.irphe.fr/%7Eeloy/>


Deadline for application: 28th February
Information and submission: 
http://centuri-livingsystems.org/recruitment-pd2018-10/




-- 
     David Robbe, PhD
     Neuronal dynamics and functions of the basal ganglia
     Institut de Neurobiologie de la Méditerranée (INMED) INSERM-UMR901
     163, avenue de Luminy. BP13
     13273 Marseille Cedex 9
     France

     http://www.inmed.fr/avenir-dynamiques-neuronales-et-fonctions-des-ganglions-de-la-base

     Tel: +33 4 91 82 81 99
     Fax: +33 4 91 82 81 05


"It is disturbingly common for studies to include behavior as simply a hasty add-on in papers
  that are otherwise crammed full of multiple techniques, types of results, and even species.
It is as if every paper needs to be a methodological decathlon in order to be considered important."
Krakauer et al.,  Neuron 2017

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.tnb.ua.ac.be/pipermail/comp-neuro/attachments/20180123/5c36ee77/attachment.html>


More information about the Comp-neuro mailing list