Nathalie PEYRARD

Révision datée du 1 février 2019 à 09:55 par Nduboispeyr (discussion | contributions) (Research)
Aller à : navigation, rechercher


Journées Francophones sur les Réseaux Bayésiens et les Modèles Graphiques Probabilistes (JFRB 2018), 31 mai-1 juin 2018 à l'INRA Toulouse.

Professional situation

DR2, MIA (Applied Mathematics and Computer Science) division at INRA.

Photo CV NP.jpg

Contact Informations

INRA Toulouse,  Unité de Mathématique et Informatique Appliquées de Toulouse

Chemin de Borde-Rouge BP 52627


Ph: +33(0)

Fax: +33(0)

Mail: nathalie.peyrard(-at-)


I work in the field of Computational Statistics. I develop frameworks and algorithms for inference and management in graphical models, with applications in epidemiology, agronomy and ecology.


  • graphical models : Markov random field, DBN, HMM, contact process on graph, factored Markov decision process (MDP).
  • Approximate inference, estimation
  • Variational methods
  • network inference
  • Application to ecological networks, meapopulation and metacommunity dynamics

PhD students:

  • Etienne Auclair (2015-2018). "Methods for ecological network inference and management". co-advised with Régis Sabbadin (MIAT-INRA-Toulouse).
  • Sebastian Le Coz (2015-2018). "A hidden Markov model for weed species dynamics in an agricultural landscape". co-advised with Pierre-Olivier Cheptou (CEFE-CNRS-Montpellier).
  • Julia Radoszycki (2012-2015). "Methods for optimization-based design of management strategies of spatio-temporal processes: application ti the management of weed species". co-advised with R. Sabbadin and S. Gaba
  • Mathieu Bonneau (2009-2012). "Adaptive spatial sampling for reconstruction of weeds maps", co-advised with R. Sabbadin and S. Gaba

Current Projects

  • MANAE: Managing Network in AgroEcology, joint INRA-CSIRO-University of Queensland projet (2013- ...)
  • ANR project AgroBiose: Biodiveristé et services ecosystémiques en agro-écosystèmes céréaliers intensifs (2014-2018)
  • EcoPhyto 2018 project VESPA: Valeur et optimisation des dispositifs dépidémiosurveillance dans une stratégie durable de protection des cultures (2013-2016)
  •   AIGM (ex MSTGA) network: Algorithmic Issues for Inference in Graphical Models (2006 - ?; funded by INRA MIA and RNSC/ISC Paris). This network gathers researchers from different disciplines (Statistics, Statistical Mechanics, Probability, Artificial Intellicence, Theoretical Epidemiology, Image Analysis), who share interest in stochastic processes on graphs and in particular in the study of variational methods as a mean for simplifying the spatial complexity in these models.


SEMMS: Spatial EM for Markovian Segmentation

GMDP Toolbox : tools for approximate resolution of graph-based Markov decision processes


[list available here]

Génotoul BioInfo
Outils personnels