UMR 518 - Applied Mathematics and Computer Science (MIA)

The Mia-Paris unit brings together statisticians and computer scientists specializing in the modeling of machine learning for biology, ecology, the environment, agronomy and food science. The team is skilled in methods of statistical inference (complex models, models with latent variables, Bayesian inference, learning, model selection etc.) and algorithms (generalization, domain transfer, knowledge representation).

The unit develops original statistical and computing methods, which may be generic or driven by precise problems in life sciences. Its activities are based on a strong background in the disciplines concerned: ecology, environmental science, food science, molecular biology, and systems biology.

Unit website : https://www6.inra.fr/mia-paris

Key figures :

Number of permanent staff in the unit: 22

Doctoral school(s)
DS 581 - Agriculture, food, biology, environment, health
Website : http://www.agroparistech.fr/abies/
Co-accredited institutions : Agreenium, Université Paris-Saclay, Université Paris-Est, Université de Reims Champagne-Ardenne
Head : Alexandre PÉRY
Research teams
  • Studies of environmental and climatic risks, particularly in the domains of pollution and hydrology.
  • Development of statistical techniques for tackling domains in which data are becoming increasingly complex, such as ecology.

Research topics: spatial and spatiotemporal statistics (Bayesian hierarchical models, point processes, studies of dependence, conditional simulations of processes), multivariate and spatialized extremes, numerical experiments, uncertainty propagation and Bayesian decision theory, analysis and inference of random graphs, trajectory modeling.

  • Creation and development of original statistical methods, principally for high-throughput technologies from molecular biology.

Research topics: segmentation and detection of break-points, modeling of time series, modeling of mixtures and models with hidden structures, analysis and inference of random graphs, detection of motifs, machine learning (model selection, variable selection, classification).

  • Development of methods for the exploitation of data from multiple, heterogeneous sources, based on an informed choice of shared, multi-scale semantic representation, in the fields of life and food sciences.
  • Study and use of machine learning methods capable of processing dynamic data, possibly from changing environments or from different tasks. One of the objectives of this work is to contribute to the enrichment of expert knowledge.

Research topics: modeling and analysis of heterogeneous data from multiple sources, human and machine multi-expertise (taking into account in semantics), collaborative and incremental learning methods.

Download PDF
Updated date: 28/05/2018

Supervisory bodies

Unit Manager

Liliane BEL

Location

Region : Île-de-France (Paris)
University site : Université Paris-Saclay
Address :
AgroParisTech - 16 rue Claude Bernard - F-75231 Paris Cedex 05