International Projects

BioASQ (view details)

  • Title: Intelligent Information Management, Targeted Competition Framework
  • Type: FP7-ICT-2011.4.4(d)
  • Date: 2012-2014
  • Involved team member(s): Eric Gaussier, Ioannis Partalas
  • Description: BioASQ will push for a solution to the information access problem of biomedical experts by setting up a challenge on biomedical semantic indexing and question answering (QA). The challenge will run in two stages, designed to (a) adapt traditional semantic indexing and QA methods to the needs of biomedical experts, and (b) collect feedback and improve the experimental setting itself.

National Projects

ARESOS (view details)

  • Title: Reconstruction, Analyse et Accès aux Données dans les Grands Réseaux Socio-Sémantiques
  • Type: programme CNRS MASTODONS
  • Date: 2012-2014
  • Involved team member(s): Eric Gaussier, Gilles Bisson
  • Description: The aim of this project is to construct, from the analysis of large textaul corpora, multipartite dynamic content and user networks and to analyze and model information diffusion in such networks. An important goal of the project is to understand, from a multi-scale analysis of the dynamics of content/user networks, the dynamics of our society.


  • Title: Comportements émergents dans les réseaux économiques et sociaux
  • Type: CAPES-COFECUB N° Sh 749-12
  • Funding:
  • [/members]

  • Date: janvier 2012 – janvier 2014
  • Involved team member(s): Mirta Gordon
  • Description: Le projet scientifique se structure autour de 3 lignes de recherche: l’étude de modèles de choix discrets dans des systèmes avec des interactions sociales, l’étude de la distribution de la richesse et de l’inégalité, et l’étude et modélisation de la criminalité urbaine.

Class-Y (view details)

  • Title: Classification With a Very Large Number of Categories
  • Type: ANR
  • Date: 2012-2014
  • Involved team member(s): Eric Gaussier, Ioannis Partalas, Rohit Babbar, Massih-Reza Amini, Cecile Amblard
  • Description: This project concerns with the classification of data in very large number of classes (large-scale classification). More specifically, the existing methods will be revisited and new methods will be proposed.


  • Title: DYNamics Of The Endocytic Proteome: A systemic perspective of trafficking and maturation (SVSE6)
  • Type: ANR Blanc
  • Date: 2012-2016
  • Involved team member(s): Ahlame Douzal, Gilles Bisson, Christophe Brouard
  • Description: The objective of the DYNOTEP project, in partnership with the CEA, is to provide a systemic perspective on the endocytic process from the angle of protein dynamics. The project is at the crossroad of fundamental research on endocytosis and experimental development on proteome dynamics. Modern large‐scale mass spectrometry‐based quantitative proteomics methods now offer the possibility to establish the repertoire of proteins and to gain insight into their temporal distribution profile in dynamic biological systems. While the technology to achieve the throughput and data quality required for studying proteome dynamics is available, the concepts for multidimensional temporal profile analysis are emerging, and the underlying mathematical tools still require refinement.

Gargantua (view details)

  • Title: biGdata ; AppRentissaGe et optimisAtioN maThématiqUe pour les données gigAntésques
  • Type: Projet Mastodons
  • Date: 2013-2014
  • Involved team member(s): Massih-Reza Amini
  • Description: Le but du projet est de former un consortium d’équipes comprenant des chercheurs en optimisation mathématique, analyse convexe, apprentissage automatique, et statistique, pour attaquer les verrous scientifiques relatifs à la conception d’algorithmes numériquement efficaces pour l’analyse de données structurées gigantesques. Le consortium est inter-disciplinaire et réunit des chercheurs du laboratoire Jean Kuntzmann, le laboratoire d’informatique de Grenoble, le laboratoire Paul Painlevé, et le département d’informatique de l’ENS. Le projet s’articule autour de deux grands axes de recherche. Le premier axe concerne la conception d’algorithmes d’optimisation stochastique numériquement efficaces pour l’apprentissage grande échelle lorsqu’une structure sous-jacente est en présense, et l’étude théorique de leur performances. Le deuxième axe concerne la conception d’algorithmes d’optimisation stochastique pour la calibration des hyperparamétres d’algorithmes d’apprentissage, et l’étude théorique de leur performances.

MacCoy Critical

  • Title: Models for Adaptative feedback enriChment and Orchestration based virtual realitY in Critical situations (MacCoy Critical)
  • Type: Appel Générique ANR 2014
  • Date: 2014-2017
  • Involved team member(s): Catherine Garbay
  • Description: MacCoy Critical is a 42 month multidisciplinary project involving researchers in human sciences, computer sciences, and end-users. Its goal is to study and improve the design of human learning devices based on simulation and virtual environments. Two applications domains knowing a great impact in terms if public health and security (medicine and driving) are considered. It is aimed more precisely at investigating the non technical competences that are involved in the management of critical situations.


  • Title: Propagation d’Actes Incivils – une approche systèmes compleXes
  • Type: PEPS CNRS — HuMaIn : Humanités – Mathématiques – Sciences de l’Information
  • Date: 2013
  • Involved team member(s): Mirta Gordon
  • Description: Le projet réunit une équipe pluridsciplinaire de criminologues, physiciens et mathématiciens autour de la question de la propagation d’actions illégales dans des contextes de contagions sociales. L’objectif est de faire se rencontrer des approches de modélisations très récentes sur ces questions et l’analyse de données recueillies sur les émeutes urbaines de 2005 en France.

PROSPECTOM (view details)

  • Title: Visual Analytics of Proteomes using Statistical Learning and integrating Multiple-view coming from Spectrometrics and Omics Databases
  • Type: défi Mastodons “grandes masse de données” du CNRS
  • Date: 2012-2014
  • Involved team member(s): Gilles Bisson
  • Description: The objective of this project is to create a synergy between Proteomics, Systems Biology, Learning and Information Visualization communities. More precisely, the first goal is to allow machine learning specialists, knowledge management and “Visual Analytics” to discover new application areas and the future challenges and issues that complex data mining will be faced with. The second goal, is to provide to the proteomics community an access to the latest knowledge extraction/fusion tools to add a semantic layer to the protein identification tools.

Regional Projects

Décision collectives en environnements interactifs et collaboratifs complexes : application à la gestion de crise.

  • Title: Collective decisions in interactive and complexe work context: crisis management application ;
  • Type: Regional
  • Funding: Call for proposals ARC6 2013; ARC “ITC and innovative use of informatic tools”;
  • Date: oct. 2013- oct.2016
  • Involved team member(s): Catherine Garbay Lauren Thévin
  • Description: The aim of this project is to design technological environment to support distant collaborative work applied to crisis management. This project is the follow-up of a previous ANR project called Imagit ( ) and will use the tangible interface TangiSense. This application has to support the sharing and coordination of different ressources among humans organisations with diffrent goals, decision protocols and ways to interact. Collaborations: IRMA (; RFIdees (; Olivier Boissier ( ); Julie Dugdale (

New Theoretical Frameworks in Metric Learning: Application to Energy Management

  • Title: New Theoretical Frameworks in Metric Learning: Application to Energy Management
  • Type: Regional
  • Funding: Call for proposals ARC6 2013; ARC “ITC and innovative use of informatic tools”
  • Date: oct. 2013- oct.2016
  • Involved team member(s): Massih-Reza Amini, Eric Gaussier, Irina Nicolae
  • Description: This proposal is a fundamental research project whose main goal is to provide new theoretical frameworks and algorithms for automatically learning metrics from data. Based on the saying “Birds of a feather flock together”, metrics play a crucial role in a large set of learning methods, such as the widely used k-nearest neighbors, kernel-based methods in classification or the k-Means algorithm in clustering. Since manually tuning metrics for a given real-world problem is often difficult and tedious, our objective is to automatically acquire knowledge from training data to optimize good metrics. This requires to formally define the notion of goodness that would allow us to ensure theoretical guarantees (i) on the generalization ability of the metric (i.e. do the properties optimized over the training set still hold on new data?) and (ii) on the generalization capability of a classifier using that metric (i.e. can we derive upper bounds on the generalization error of the classifier?). The metric learning algorithms developed in this project will be used to deal with image classification tasks in order to not only increase the classification accuracy but also improve the creation of relevant visual dictionaries.
  • Collaborations: Amaury Habrard, Marc Sebban (LHC – St-Etienne)

Local Projects


  • Title: Développement d’une plateforme mobile destinée à participer à la surveillance d’un ensemble de personnes en situation de fragilité (Attentive)
  • Type: Projet exploratoire PERSYVAL
  • Date: 2013-2014
  • Description: The purpose of the project is to design a mobile communicating companion robot, to provide assistance, and to ensure follow-up of impaired people (elderly, disabled…). The role of the robot is to perceive the situations of the “accompanied” person, his behavior and needs, based on various sensors placed on both of them. The overall objective is that the robot should be able to accompany and assist impaired persons in a “natural” way, at home or during a walk, and raise alerts in case of detected problem (fall, weakening…).
    The work falls under the context of cooperation between two research laboratories: LIG (teams AMA, MAGMA, GETALP) and GIPSA-lab (team AGPIG).

KHRONOS (view details)

  • Title: KHRONOS
  • Type: Projet action Labex Persyval
  • Date: 2014-2016
  • Involved team member(s): Massih-Reza Amini
  • Description: The aim of the project is to explore the properties of the adaptive nonlinear filter in the case of indirect (incomplete and blurry) observations. The objective of this work is twofold. First, the statistical properties of the proposed algorithms should be studied under various observation signal scenarios. Second, we aim to devise the implementation of the adaptive filters through iterative saddle-point optimization, developed in [3,4], which should allow to treat efficiently large-scale data

THANATOS (view details)

  • Title: THeoretical ANAlysis and sTudy of One Shot learning
  • Type: Projet UJF – AGIR
  • Date: 2014-2015
  • Involved team member(s): Massih-Reza Amini, Yuri Maximov
  • Description: The project concerns a theoretical study of one shot learning. This learning paradigm has been developed for emerging applications, where the aim is to find a predictor with one or few labeled examples associated to the application of interest, with many other unlabeled examples and labeled examples from related learning tasks to the main application. This is for example the case in computer vision where we may have few labeled images for a new concept, together with many unlabeled images containing the concept and also many labeled images of other concepts. Another example, is learning to rank applications in information retrieval, where the aim is to find a ranking function for a new set of queries/documents for which we may have just the ranking of a subset of documents to a given query together with many rankings of documents to other passed queries.

Industrial Projects

Best of Media

  • Title: Activity prediciton in social networks
  • Type: Cifre
  • Date: 2011-2014
  • Involved team member(s): A. Douzal, E. Gaussier
  • Description: The aim of the project is to study machine learning techniques for activity prediciton in social networks.


  • Title: Sequence learning for information extraction
  • Type:Cifre
  • Date: 2014-2017
  • Involved team member(s): M.-R. Amini, E. Gaussier
  • Description: The aim of the project is to study information extraction via sequence learning methods from heterogeneous documents.


  • Title: Temporal data analysis and learning for efficiency energetics on sensor data
  • Type: Cifre
  • Date:2013-2017
  • Involved team member(s): Ahlame Douzal, Cao Tri Do
  • Description: The aim of the project is to study new temporal data analysis and learning algorithms for the regression and the classification (supervised and semi-supervised) of sensor data. .

Xerox Research Centre Europe

  • Title: Predictive models for relational data
  • Type: Cifre
  • Date: 2014-2017
  • Involved team member(s): E. Gaussier
  • Description: The aim of the project is to study matrix factorization methods in a realtional context with the will to capture and infer new relations between objects.

Past Projects

Interactive (view details);

  • Title: accident avoidance by active intervention for Intelligent Vehicles;
  • Type: FP7-ICT
  • Date: 01/2010-12/2013
  • Involved team member(s): Olivier Aycard, Ricardo Omar Chavez Garcia
  • Description: InteractIVe develops advanced driver assistance systems (ADAS) for safer and more efficient driving. interactIVe introduces safety systems that autonomously brake and steer. The driver is continuously supported by interactIVe assistance systems. They warn the driver in potentially dangerous situations. The systems do not only react to driving situations, but are also able to actively intervene in order to protect occupants and vulnerable road users. Seven demonstrator vehicles – six passenger cars of different vehicle classes and one truck – will be built up to develop, test, and evaluate the next generation of safety systems.

Fragrances (view details)

  • Title:Filtrage, Recherche et Annotations dans des Graphes d’Interaction Sociaux;
  • Type: ANR-08-CORD-008
  • Date: 2009-2013
  • Involved team member(s): Gilles Bisson, Eric Gaussier
  • Description: The goal of the FRAGRANCES project is to develop a new algorithms and tools for the information retrieval tasks in the context of the interaction networks, and more precisely of the Social Networks. The contributions of our team to this project concerns the information diffusion modeling (Cédric Lagnier PhD thesis) and the multi-view clustering (Clément Grimal PhD thesis).


  • Title: IMAGIT
  • Type: ANR-10-CORD-0017
  • Date: oct. 2010- dec. 2013
  • Involved team member(s): Catherine Garbay, Lauren Thevin
  • Description: The project consist to design, realize and test a new environment making it possible a whole of users to interact in a way distributed using a set of interaction tables, but also tangible and virtual objects. The interactive table is based on RFID technology for the identification and localization of tangible objects.

Metricc (view details)

  • Title: MEmoire de Traduction, Recherche d’Information et Corpus Comparables
  • Type: ANR-08-CORD-008
  • Date: 2009-2013
  • Involved team member(s): Gilles Bisson, Eric Gaussier
  • Description: TThe goal of the METRICC project was to develop new algorithms and tools for the exploitation of compararble corpora in cross-language settings, in particular for cross-language inforamtion retrieval.


  • Title: Predicting activity in social media
  • Type: Cifre
  • Date:2011-2014
  • Involved team member(s): Eric Gaussier, Ahlame Douzal, Francois Kawala
  • Description: Themes covered by this project are detection, analysis and modeling of users behaviours that induce and explain macro changes (eg. users interest wrt. a given topic) which are related to the monetization of contents. We focus on models that can cope with distinct informations sources (eg. content and behavioural). The project could span on :
    • Analysis and selection of informations sources
    • Content based feature selection
    • Behaviour based feature selection
    • Modeling of information diffusion for online social networks