Contributions to hidden Markov models and applications to plant structure analysis
Speciality : Mathématiques Appliquées
19/10/2020 - 14:30 Jean-Baptiste Durand (Grenoble INP) Amphithéâtre - Maison Jean Kuntzmann
Keywords :
- Probabilistic graphical models
- Statistical analysis of tree-structured data
- Applications to plant structure analysis
This presentation will focus on diverse contributions to hidden Markov modelling on graphs, with a particular emphasis being placed on tree graphs. I will introduce various families of hidden Markov tree models, motivated by the analysis and prediction of plant growth and architecture. Connections with the theory of probabilistic graphical models and multivariate counts will be exposed, as well as dedicated methods to quantify hidden state uncertainty. Lastly, an overview of current research projects on spatial statistics will be provided, together with concluding remarks and general perspectives.
President:
Florence Forbes (DR Inria - Inria Rhône-Alpes)Raporteurs:
- Gerhard Buck-Sorlin (Professeur - IRHS, AgroCampus Ouest )
- Hajo Holzmann (Professeur - Philipps-Universität, Marburg, Allemagne )
- Geoffrey J. McLachlan (Professeur - University of Queensland, Brisbane, Australia )
Examinators:
- Nathalie Peyrard (DR INRAE - INRAE, Unité MIA, Castanet-Tolosan )
- Evelyne Costes (DR INRAE - INRAE, UMR Agap, Montpellier )
- Olivier François (Professeur - Grenoble INP / TIMC, Grenoble )