Section: Partnerships and Cooperations

National Initiatives

ANR Project Qcompere

Participants : Guillaume Fortier, Cordelia Schmid, Jakob Verbeek.

This three-and-a-half year project started in November 2010. It is aimed at identifying people in video using both audio (using speech and speaker recognition) and visual data in challenging footage such as news broadcasts, or movies. The partners of this project are the CNRS laboratories LIMSI and LIG, the university of Caen, Inria's LEAR team, as well as two industrial partners Yacast and Vecsys Research.

ANR Project Physionomie

Participants : Frédéric Jurie [University of Caen] , Jakob Verbeek, Shreyas Saxena.

Face recognition is nowadays an important technology in many applications ranging from tagging people in photo albums, to surveillance, and law enforcement. In this 3-year project (2013–2016) the goal is to broaden the scope of usefulness of face recognition to situations where high quality images are available in a dataset of known individuals, which have to be identified in relatively poor quality surveillance footage. To this end we will develop methods that can compare faces despite an asymmetry in the imaging conditions, as well as methods that can help searching for people based on facial attributes (old/young, male/female, etc.). The tools will be evaluated by law-enforcement professionals. The participants of this project are: Morpho, SensorIT, Université de Caen, Université de Strasbourg, Fondation pour la Recherche Stratégique, Préfecture de Police, Service des Technologies et des Systèmes d’Information de la Sécurité Intérieure, and LEAR.

ANR Project Macaron

Participants : Julien Mairal, Zaid Harchaoui, Laurent Jacob [CNRS, LBBE Laboratory] , Michael Blum [CNRS, TIMC Laboratory] , Joseph Salmon [Telecom ParisTech] .

The project MACARON is an endeavor to develop new mathematical and algorithmic tools for making machine learning more scalable. Our ultimate goal is to use data for solving scientific problems and automatically converting data into scientific knowledge by using machine learning techniques. Therefore, our project has two different axes, a methodological one, and an applied one driven by explicit problems. The methodological axis addresses the limitations of current machine learning for simultaneously dealing with large-scale data and huge models. The second axis addresses open scientific problems in bioinformatics, computer vision, image processing, and neuroscience, where a massive amount of data is currently produced, and where huge-dimensional models yield similar computational problems.

This is a 3 years and half project, funded by ANR under the program “Jeunes chercheurs, jeunes chercheuses”, which started in October 2014. The principal investigator is Julien Mairal.

PEPS CNRS BMI (Biology - Mathematics - Computer Science), Project FlipFlop

Participants : Elsa Bernard [Institut Curie, Ecoles des Mines-ParisTech] , Laurent Jacob [CNRS, LBBE Laboratory] , Julien Mairal, Jean-Philippe Vert [Institut Curie, Ecoles des Mines-ParisTech] , Anne-Hélène Monsoro-Burq [Institut Curie] .

The project is concerned with large-scale sparse estimation techniques for processing RNA-Seq data. It led to a joint publication [4] with partners from Inria Grenoble, Institut Curie in Paris, and the LBBE laboratory in Lyon. The principal investigator was Laurent Jacob (CNRS, LBBE laboratory). The project started in Jun 2012 and ended in Dec 2014.

MASTODONS Program CNRS - Project Gargantua

Participants : Zaid Harchaoui, Julien Mairal.

The project is concerned with machine learning and mathematical optimization for big data. The partners are from LJK (Grenoble), LIG (Grenoble), LIENS (ENS, Paris), Lab. P. Painleve (Lille). Principal investigator/leader: Zaid Harchaoui. Dates: May 2013-Dec. 2014

Equipe-action ADM du Labex Persyval (Grenoble) “Khronos”

Participants : Zaid Harchaoui, Massih-Reza Amini [LIG] .

The partners of this project are from the laboratories LJK, LIG, GIPSA, TIMC, CEA. The principal investigators/leaders are Zaid Harchaoui (Inria and LJK), Massih-Reza Amini (LIG). The project started in Jan. 2014 and ends in Dec. 2016.