Researcher in Machine Learning / Data Mining, LIST, CEA Saclay, France

My fields of research are currently oriented towards deep learning on streams of data and incremental deep models, i.e. investigating reactive Online Deep Learning procedures for streams of data. I am also currently investigating deep models for fault detection, identification and diagnosis in production chains.
An other thematic that I am currently exploring is the problem of dynamic multiple-steps-ahead model selection, with an application to time series forecasting with exogenous data.



General domains of interest:
  • Deep learning / online reactive deep learning algorithms
  • Fault detection, identification and diagnosis
  • Time series forecasting / model selection
  • Automatic scheduling
  • Ressource allocation
    -----------------------------------------------------------
  • Machine learning
  • Data mining / visual data mining
  • Applications: mining and visualizing temporal series

Past research topics (Ph.D work):
  • image retrieval / satellite imagery
  • content-based image retrieval (CBIR) systems
  • active learning / relevance feedback / human-machine interaction
  • object retrieval using coarse-to-fine strategies
  • semi-supervised learning (theory)
An quick overview of my Ph.D work can be found here. My Ph.D manuscript can be downloaded here (pdf).


Current position: Research engineer at CEA Saclay.


Publications:
  • Ph.D: Fast learning methods adapted to the user specificities: application to earth observation image information mining (pdf).

  • Journal articles:
    • Pierre Blanchart, and Marine Depecker. A Non-linear Semantic-preserving Projection Approach to Visualize Multivariate Periodical Time-series, IEEE transactions on Neural Networks and Learning Systems (TNNLS), 2013.
    • Pierre Blanchart, Marin Ferecatu, and Mihai Datcu. Pattern retrieval in large image databases using multiscale coarse-to-fine cascaded active learning, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (JSTARS), 2013.
    • Pierre Blanchart, and Mihai Datcu. A Semi-Supervised Algorithm for Auto-Annotation and Unknown Structures Discovery in Satellite Image Databases, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (JSTARS), 2010 (pdf).
    • Pierre Blanchart, Adila Azzou, and Philippe Blanchart. Predicting the sintering curve of Porcelain by Support Vector Regression, Journal of the American Ceramic Society, JACERS 2011 (pdf).
  • Conference articles:

    • Pierre Blanchart and Cédric Gouy-Pailler. WHODID: Web-based interface for Human-assisted factory Operations in fault Detection, Identification and Diagnosis. ECML/PKDD, Skopje, 2017.
    • Andrey Besedin, Pierre Blanchart, Michel Crucianu and Marin Ferecatu. Evolutive deep models for online learning on data streams with no storage. 2nd ECML/PKDD Workshop on Large-scale Learning from Data Streams in Evolving Environments, Skopje, 2017.
    • Pierre Blanchart, and Marin Ferecatu. Local integrity constraints for structure detection and segmentation in high-resolution earth observation images. To appear in the proceedings of IEEE Conference on Image Processing (ICIP), Québec City, 2015 (pdf).
    • Pierre Blanchart, Marine Depecker, and Cédric Auliac. Large scale FEV charge scheduling under contractual power constraints: a priority rule based semi-online algorithm, to appear in the proceedings of Transport Research Arena (TRA), Paris, 2014.
    • Shiyong Cui, Mihai Datcu, and Pierre Blanchart. Cascaded active learning for SAR image annotation, in the proceedings of International Geoscience and Remote Sensing Symposium (IGARSS), Munich, 2012.
    • Pierre Blanchart, Marin Ferecatu, and Mihai Datcu. Mining large satellite image repositories using semi-supervised methods, in the proceedings of International Geoscience and Remote Sensing Symposium (IGARSS), Vancouver, Canada, 2011 (oral presentation).
    • Pierre Blanchart, Marin Ferecatu, and Mihai Datcu. Cascaded Active Learning for Object Retrieval using Multiscale Coarse to Fine Analysis, in the proceedings of IEEE Conference on Image Processing (ICIP), Brussels, Belgium, 2011 (poster presentation) (extended version pdf).
    • Pierre Blanchart, Marin Ferecatu, and Mihai Datcu. Indexation of large satellite image repositories using small training sets, in the proceedings of ESA-EUSC-JRC Conference on Image Information Mining, Ispra, Italy, 2011 (oral presentation).
    • Pierre Blanchart, Marin Ferecatu, and Mihai Datcu. Active Learning Using the Data Distribution for Interactive Image Classification and Retrieval, in the proceedings of IEEE Symposium on Computational Intelligence and Data Mining (CIDM), Paris, 2011 (oral presentation).
    • Pierre Blanchart, Marin Ferecatu, Mihai Datcu. Apprentissage actif et utilisation de la structure a priori des données : application à une base d'images satellites à haute résolution, in the proceedings of RFIA, Caen, France, 2010 (oral presentation).
    • Pierre Blanchart, and Mihai Datcu. Semi-supervised learning and discovery of unknown structures among data: Application to satellite image annotation, in the proceedings of IGARSS, Cape Town, South Africa, 2009 (oral presentation).
    • Pierre Blanchart, Liyun He, and François Le Gland. Information fusion for indoor localization, in the proceedings of FUSION'09, Seattle, United States, 2009 (oral presentation) (pdf).


Education:
  • 2008 - 2011: Ph.D student (Télécom ParisTech, France)
Subject: Fast learning methods adapted to the user specificities: application to earth observation image information mining
Laboratories: Télécom ParisTech, TSI Department (UMR CNRS 5141 LTCI) – CNES Toulouse
Advisors: Mihai Datcu (German Aerospace Center DLR), Marin Ferecatu (CNAM)
  • 2007 - 2008: Master Research in Signal, Image, Embedded systems, Control engineering, University of Rennes 1, France
  • 2005 - 2008: Engineer degree in Telecommunications, specialization "Signal and Image", Télécom Bretagne, Brest, France


Curriculum vitae: 

Contact:

Address:   Pierre Blanchart
                 CEA LIST
                 Laboratoire Analyse de Données et Intelligence des Systèmes
                 Centre de Saclay, bat. 565 – Point courrier 192
                 91191 Gif-sur-Yvette Cedex – France