Pierre Latouche

Professor of Statistics and Machine Learning
Laboratoire MAP5
Université Paris Cité

Ecole Polytechnique

I am Professor of Applied Mathematics at the MAP5 laboratory of the University of Paris. My group is involved in problems related to learning and statistics. Until 2018, I was Assistant Professor and then Associate Professor at Université Paris 1 Panthéon-Sorbonne.  I both studied in France and in the UK. I am a reviewer for journals and conferences including JASA, PNAS, Biometrika, NeurIPS, ICML, JRSS-B, JRSS-C. I serve as an associate editor for the Journal of the Royal Statistical Society and for the Bayesian Analysis Journal. I am an expert for the European research council. I co-invented and developed the Linkage and Topix softwares. I teach statistics and machine learning at the University of Paris and Ecole Polytechnique where I also have a (part time) professor position. In particular, I am involved in the master Artificial Intelligence & Advanced Visual Computing. Moreover, I am the director of a bachelor in data science at the University of Paris. I am also responsible along with P. A. Mattei (INRIA) of the (deep) probabilistic graphical models course of the master MVA of ENS Paris Saclay.


I am a member of the international working group on model based clustering created by Adrian Raftery from the University of Washington. Meetings are hold regularly in north America and in Europe.  I am also a member of the European cooperation for statistics of network data science action which brings together European researchers of the field.


I am also deeply involved in the knowledge transfer of the algorithms / softwares I develop. I have a patent in the US as well as one pending in Europe and I am involved in entrepreneurial projects.


My research interests include: 

* Statistical learning on networks, texts, and heterogenous data

* Statistical learning in high dimensions

* Statistical learning with processes

* Tests and proofs

* Deep graphical modelling / Deep learning

List of former and current postdoc and engineers:

* Carlos Ocanto Dávila

* Stéphane Petiot

* Damien Marié
* Laurent Bergé 

.

News

The very latest news regarding research and teaching

Linkage

A US patent is available for this technology we develop

Topix 

  

Publications


Recent papers

A. Leroy, P. Latouche, B. Guedj, and S. Gey. "MAGMA: Inference and prediction with multi-task Gaussian processes (2022). In: Machine Learning (p. 1821-1849) [web].

S. Ouadah, P. Latouche, S. Robin. "Motif-based tests for bipartite networks". In: Electronic Journal of Statistics (2021), p. 293-330 [web].

D. Liang, M. Corneli, C. Bouveyron, P. Latouche. "DeepLTRS: A deep latent recommender system based on user ratings and reviews". In: Pattern Recognition Letters (2021), p. 267-274 [web].

N. Jouvin, C. Bouveyron, P. Latouche. "A Bayesian Fisher-EM Algorithm for Discriminative Gaussian Subspace Clustering". In: Statistics and Computing (2020), p. 1-20 [web].

E. Côme, P. Latouche, N. Jouvin, and C. Bouveyron. "Hierarchical clustering with discrete latent variable models and the integrated classification likelihood". In: Journal of Advances in Data Analysis and Classification (2020), p. 957-98 [web].

N. Jouvin, P. Latouche, C. Bouveyron, G. Bataillon, and A. Livartowski. "Greedy clustering of count data through a mixture of multinomial PCA". In: Computational Statistics (2020), p. 1-33 [web]. 


Media

A. Mestre, "Eric Zemmour, nouveau président de la fachosphère ?". In: LeMonde (2022), p1. and p. 16-17 [web].

S. Laurent, "Comment la gauche sociale-démocrate a perdu la bataille des réseaux sociaux". In: LeMonde (2022), p. 16-17 [web]. 

S. Auffret, "Brigitte Macron et Jean-Michel Trogneux, itinéraire d’une infox délirante". In: LeMonde (2022), p. 16-17 [web].

M. Goar, N. Chapuis, "Présidentielle 2022 : faut-il se couper de Twitter, huis clos politique devenu hostile ?". In: LeMonde (2022), p. 1 and p. 16-19 [web].

P. Latouche, C. Bouveyron, D. Marié, G. Fouetillou, "Présidentielle 2017 : une réorganisation politique du web social ?". In: TheConversation (2017) [web].

P. Latouche, C. Bouveyron, D. Marié, G. Fouetillou, "Présidentielle 2017 : une réorganisation politique du web social ?". In: Data analytics post (2017) [web].

P. Latouche, C. Bouveyron, D. Marié, G. Fouetillou, "Présidentielle 2017 : une réorganisation politique du web social ?". In: Panthéon-Sorbonne magazine (2017).

P. Latouche, C. Bouveyron, "Les échanges de données au peigne fin". In: CNRS, le journal (2017), p. 9 [web].

P. Latouche, C. Bouveyron, "Des réseaux, des textes, et de la statistique !". In: Lettre de l’INSMI (2016).


Preprints

D. Liang, M. Corneli, C. Bouveyron, P. Latouche. "Clustering by deep latent position model with graph convolution network" [web].

A. Leroy, P. Latouche, B. Guedj, and S. Gey. "Cluster-Specific Predictions with Multi-Task Gaussian Processes" [web].


Papers

A. Leroy, P. Latouche, B. Guedj, and S. Gey. "MAGMA: Inference and prediction with multi-task Gaussian processes (2022). In: Machine Learning (p. 1821-1849) [web].

S. Ouadah, P. Latouche, S. Robin. "Motif-based tests for bipartite networks" In: Electronic Journal of Statistics (2021), p. 293-330 [web].


D. Liang, M. Corneli, C. Bouveyron, P. Latouche. "DeepLTRS: A deep latent recommender system based on user ratings and reviews". In: Pattern Recognition Letters (2021), p. 267-274 [web].


N. Jouvin, C. Bouveyron, P. Latouche. "A Bayesian Fisher-EM Algorithm for Discriminative Gaussian Subspace Clustering" In: Statistics and Computing (2020), p. 1-20 [web].


E. Côme, P. Latouche, N. Jouvin, and C. Bouveyron. "Hierarchical clustering with discrete latent variable models and the integrated classification likelihood". In: Journal of Advances in Data Analysis and Classification (2020), p. 957-986 [web].


N. Jouvin, P. Latouche, C. Bouveyron, G. Bataillon, and A. Livartowski. "Greedy clustering of count data through a mixture of multinomial PCA". In: Computational Statistics (2020), p. 1-33 [web]. 



M. Corneli, C. Bouveyron, and P. Latouche. "Co-clustering of ordinal data via latent continuous random variables and a classification EM algorithm". In: Journal of Computational and Graphical Statistics (2020), p. 771-785 [
web]. 



S. Ouadah, S. Robin, and P. Latouche. "A degree-based goodness-of-fit test for heterogeneous random graph models". In: Scandinavian Journal of Statistics (2019), p. 156-181 [
web].



L. Bergé, C. Bouveyron, C. Corneli, and P. Latouche. "The latent topic block model for the co-clustering of textual interaction data". In: Journal of Computational Statistics and Data Analysis (2019), p. 247-270 [
web].



G. Fouetillou, P. Latouche, C. Bouveyron, and P-A. Mattei. "Exact dimensionalality selection for Bayesian PCA". In: Journal of Statistical Planning and Inference (2019), p. 196-211 [
web].



M. Corneli, C. Bouveyron, P. Latouche, and F. Rossi. "The dynamic stochastic topic block model for time evolving networks with textual edges". In: Statistics and Computing (2019), in press [web].



P. Latouche, C. Bouveyron, and P-A. Mattei. "Bayesian variable selection for globally sparse probabilistic PCA". In: Electronic Journal of Statistics 12.2 (2018), p. 3036-3070 [
web].



R. Rastelli, P. Latouche, and N. Friel. "Choosing the number of groups in a latent stochastic block model for dynamic networks". In: Network Science 6.4 (2018), p. 469-493 [
web].



P. Latouche, S. Robin, and S. Ouadah. "Goodness of fit of logistic regression models for random graphs". In: Journal of Computational and Graphical Statistics (2018), p. 98-109 [
web].



M. Corneli, P. Latouche, and F. Rossi. "Multiple change points detection and clustering in dynamic networks". In: Statistics and Computing (2018), p. 989-1007 [
web].



P. Latouche, C. Bouveyron, D. Marié, and G. Fouetillou. "Présidentielle 2017 : l’analyse des tweets renseigne sur les recompositions politiques". In: Statistique et Société 5.3 (2017) [
web].



J. Wyse, N. Friel, and P. Latouche. "Inferring structure in bipartite networks using the latent block model and exact ICL". In: Network Science 5.1 (2017), p. 45-69 [
web].



R. Zreik, P. Latouche, and C. Bouveyron. "The dynamic random subgraph model for the clustering of evolving networks". In: Computational Statistics (2016), p. 1-33 [
web].



P ; Latouche and S. Robin. "Variational Bayes model averaging for graphon functions and motifs frequencies inference in W-graph models". In: Statistics and Computing 26.6 (2016), p. 1173-1185 [
web].



P. Latouche, P-A Mattei et al. "Combining a relaxed EM algorithm with Occam’s razor for Bayesian variable selection in high dimension regression". In: Journal of Multivariate Analysis 146 (2016), p. 177-190 [
web].



M. Corneli, P. Latouche, and F. Rossi. "Exact ICL maximisation in a non stationary temporal extension of the stochastic block model for dynamic networks". In: Neurocomputing 192 (2016), p. 81-91 [
web].



M. Corneli, P. Latouche, and F. Rossi. "Block modelling in dynamic networks with non homogenous Poisson processes and exact ICL". In: Social Network Analysis and Mining 6.1 (2016), p. 55-85 [
web].



C. Bouveyron, P. Latouche, and R. Zreik. "The stochastic topic block model for the clustering of vertices in networks with textual edges". In: Statistics and Computing (2016), p. 1-21 [
web].



R. Zreik, P. Latouche, and C. Bouveyron. "Classification automatique de réseaux dynamiques avec sous-graphes : étude du scandale Enron". In: Journal de la Société Française de Statistique 156.3 (2015), p. 166-191 [
web].



E. Côme and P. Latouche. "Model selection and clustering in stochastic block models based on the exact integrated complete data likelihood". In: Statistical Modelling 15.6 (2015), p. 564-589 [
web].



P. Latouche, E. Birmelé, and C. Ambroise. "Model selection in overlapping stochastic block models". In: Electronic Journal of Statistics 8.1 (2014), p. 762-794 [
web].



Y. Jernite, P. Latouche et al. "The random subgraph model for the analysis of an ecclesiastical network in Merovingian Gaul". In: Annals of Applied Statistics 8.1 (2014), p. 377-405 [
web].



P. Latouche, E. Birmelé, and C. Ambroise. "Variational Bayesian inference and complexity control for stochastic block models". In: Statistical Modelling 12.1 (2012), p. 93-115 [
web].



P. Latouche, E. Birmelé, and C. Ambroise. "Overlapping stochastic block models with application to the French political blogosphere". In: Annals of Applied Statistics 5.1 (2011), p. 309-336 [
web].




Chapters

R. Zreik, C. Ducruet, C. Bouveyron, and P. Latouche. "Cluster dynamics in the collapsing Soviet shipping network". In : Advances in Shipping Data Analysis and Modeling Tracking and Mapping Maritime Flows in the Age of Big Data. Routledge (2017) [web].



R. Zreik, P. Latouche, C. Bouveyron, and C. Ducruet. "Cluster identification in maritime flows with stochastic methods". In : Maritime Networks : Spatial Structures and Time Dynamics. Routledge (2015) [
web].



P. Latouche, E. Birmelé, and C. Ambroise. "Overlapping clustering methods for networks". In : Handbook of Mixed Membership Models and Their Applications. Chapman et Hall/CRC (2014) [
web].



P. Latouche, E. Birmelé, and C. Ambroise. "Bayesian methods for graph clustering". In : Advances in Data Handling and Business Intelligence". Springer (2009) [
web].




Softwares

- Greedy (R package): an ensemble of algorithms that enable the clustering of networks and data matrix such as document/term matrix with different type of generative model

- MAGMA (R package): inference and prediction with multi-task Gaussian processes

- MAGMAclust (R package): inference and prediction with cluster-specific multi-task
Gaussian processes

- Topix (web):  allows to summarize massive and possibly extremely sparse data bases involving text

- FisherEM (R package): efficient method for the clustering of high-dimensional data

- ordinalLBM (R package): implements functions for simulation and estimation of the ordinal latent block model (OLBM)

MoMPCA (R package): inference and clustering for mixture of multinomial principal component analysis

Linkage (web): analysis of networks with textual edges

- GSPPCA (R package): implements the GSPPCA algorithm for high-dimensional unsupervised feature selection 

- Spinyreg (R package): spare regression using spike and slab prior distributions

- GofNetwork (R package): assess the goodness of fit of network models in the presence of covariates

- Mixer (R package written in C++): variational inference techniques for the
stochastic bloc model. Can be used to classify the vertices of a network depending on their connection profiles

- Rambo (R package): estimate the parameters, the number of classes and cluster vertices of a random network into groups with homogeneous connection profiles. The clustering is performed for directed graphs with typed edges (edges are assumed to be drawn from multinomial distributions) for which a partition of the vertices is available

- Netlab (Matlab): some of the most important pattern recognition algorithms described by C.M. Bishop in “Neural Networks for Pattern Recognition” (Oxford University Press, 1995)

- Genoscript (WebObject): a Web environment for transcriptom analysis