Machine Learning group

The main objective comprises modeling and statistical learning studies designed to grasp data diversity (dimensionality, structures, non-stationarity), as well as the nature of the expected solutions (prior knowledge).

1. Multi-source, multi-view, multi-modal learning

2. Semi-supervised or weakly supervised learning

3. Transfer learning, domain adaptation, optimal transport

4. Robust machine learning and trusted AI

5. Deep neural networks

6. Recurrent neural networks

7. Generative adversarial networks

8. Graph analysis, graph matching, metrics on graphs

9. Graph-based models, multi-scale representations

10. Sparse models, dictionary learning and tensor decomposition

11. Information fusion, ensemble learning, random forests

12. Spatial and temporal information modeling

13. Federated learning and serious games

• Handwriting recognition
• Analysis of handwritten documents
• Analysis of old documents
• Information retrieval
• Medical imaging
• Analysis of sports gestures
• Analysis and synthesis of molecules
• Analysis of road scenes
• Road traffic analysis
• Risk management
• Historical spatial dynamics
• Analysis and description of audio scenes
• Time series analysis

PRODIGE-AI : PRObability, ranDom matrIx theory, Geometry and gEneralization for generative-AI (PEPR IA, 2026-2030)
Partners : INRIA (Lyon), I2M (Marseille) , IMT (Toulouse), LHC (Saint Etienne), LIS (Marseille), LPT (Toulouse), LJAD (Nice)

Share2Detect: Sharing is detecting in Federated Learning (ANR, 2026-2030)

HyFoundationS: Hyperspectral Foundation Models for Chemical Soil Analysis (ANR, 2026-2030)
Partner : Institut CARMeN (ex-COBRA Lab)

POPSTRAS: Extraction d’informations à partir de tableaux semi-structurés manuscrits du Fichier Domiciliaire pour une histoire de la population de Strasbourg (1871-1939) (ANR, 2026-2030)
Partners : SAGE, UMR IDEES

EulCat : l’IA conversationnelle Normande (FEDER/Région Normandie, 2025-2027)
Partners : REDlab, JUST IA, CRIANN

OCTOPUSSY : Nouvelles techniques computationnelles pour l'optimisation de polymères (ANR, 2024-2028)
Partners : PBS, PIMM, COBRA, University of York

ODD : Online Deep anomaly Detection (ANR, 2023-…)

Chaire IA RAIMO : A Road toward safe Artificial Intelligence in MObility (ANR, 2020-2024)

DynaTeam : Dynamique de coordination d'équipe (ANR, 2023-…)
Partners : CETAPS, Équipe RI2C du LITIS

FAMOUS : Fair Multimodal Learning (ANR, 2023-…)
Partners : LIS, LHC, INT, Euranova

FINLAM : Foundation INtegrated models for Libraries Archives and Museum (ANR, 2023-…)
Partners : TEKLIA, BnF

SHARP : Machine Learning for Safe Vehicle Charging Points (ANR, 2023-…)
Partners : Citeos Solutions Digitales (VINCI Energies), GREYC

CoDeGNN : Convolution and Decimation for Graph Neural Networks (ANR, 2021-2025)
Partners : GREYC, LIFAT

MediSeg : Deep medical image segmentation: what’s next? (ANR, 2021-2025)
Partners : LMI, ImVIA, Équipe QuantIF du LITIS

MultiTrans : Gradual Multi Transfer Learning for Safe Autonomous Driving (ANR, 2021-2025) Partners : Équipe STI du LITIS, I3S, Valeo.ai

EXO-POPP : Optical Extraction of Handwritten Named Entities for the Marriage Certificates of the Population of Paris (1921-1946) (ANR, 2021-2024)
Main partner : LARHRA

CATCH : Automatic Understanding of Human Sensors Testimonials (ANR, 2021-2023)
Partners : Saagie, Atmo Normandie

LabCom L-Lisa : Common Lab LITIS-Saagie (ANR, 2021-2025)
Partner : Saagie

SCHISM : Supporting CHemoinformatics via Interactive unsupervised and Semi-supervised data Mining (FEDER, 2020-2024)

Partners : GREYC, CERMN

DeepART : Medical Image collection, segmentation and anonymization for DEEP learning in Adaptive Radiation Therapy (2020-2022)
Partners : Centre François Baclesse, Centre Henri Becquerel, GREYC

AMCAS : Machine Learning for Understanding Secondary Atomization (2020-2022)
Partner : CORIA

THESIS : THermoplastic Erosion Shield for new generation Ice protection System (2019-2022)
Partners : Zodiac Aerosafety Systems (nowadays SAFRAN), Analyse and Surface, Dedienne Multi-plasturgy Group, IRSEEM, GPM, CORIA

ESCAPE SG : Serious game for evacuation management (2019-2023)

Partners : UMR IDEES, UMI UMMISCO

APi : Taming the Beast of the Preimage in Machine Learning for Structured Data : Signal, Image and Graph (ANR, 2018-2023)

Partners : LTCI, LIG

PAUSE APi : in complement to ANR APi, this funding is from the national program for the urgent aid and reception of scientists in exile (PAUSE, from the Collège de France)

NormanD'eep : Deep learning and application (2018-2022)
Partner : GREYC

ASTURIAS : Newspaper image analysis (2018-2022)
Partners : GREYC, MIND group of LITIS

DeepInFrance : Machine learning with deep neural networks (ANR, 2018-2022)
Partners : INRIA Grenoble, I3S, GREYC, LIF-AMU, LIP6

LEAUDS : Learning to understand audio scenes (ANR, 2018-2022)
Partners : Netatmo, INRIA Nancy - Grand Est

ESCAPE : Exploring by Simulation Cities Awareness on Population Evacuation (ANR, 2017-2022)
Partners : UMR IDEES, UMI UMMISCO

DynACEV : Dynamics of Learning: Behaviour and Lived Experiences. The role of exploratory strategies (ANR, 2017-2022)
Partners : CETAPS, ISSUL (Lausanne), MIP, SOPESES (Nouvelle-Zélande), SPARC (UK)

OATMIL : Optimal Transport and Machine Learning (ANR, 2017-2022)
Partners : IRISA, Lagrange, Technicolor

HBDEX : Analysis of historical financial data (ANR, 2017-2022)
Partners : PSE, CAMS, IRISA

POPP : P**roject for the oceration of the Paris population census** (1921-1946) (2020-2021) - voir démonstrateur ici
Partner : LARHRA

EurHisFirm: The European project aims at designing a world-class research infrastructure to collect, merge, extract, collate, align and share detailed historical high-quality firm level data for Europe (H2020 EU, 2018-2021)

Partners : Consortium of 16 European partners

Read E-Doc : Development of a system for extracting specific alphanumeric information from dematerialized container shipment slips (2019-2021)

Partner : Toshiba

INTERNATIONAL

DEEDS: Deep Ensembles for Evolving Data Streams (CAPES COFECUB, 2025-2029)

DD-AnDet: Data-driven Anomaly Detectors for Time Series Data and Big Data (CAPES STIC-AmSud, 2024-2026)

MMLARP : Multimodal/Multiview Learning Applied to Real Problems (CAPES STIC-AmSud, 2024-2025)

WeSmile : Weakly supervised medical image segmentation (PHC Van Gogh, 2020-2022)
Partner : Eindhoven TU

EurHisFirm : The European project aims at designing a world-class research infrastructure to collect, merge, extract, collate, align and share detailed historical high-quality firm level data for Europe (H2020 EU, 2018-2021)
Partners : Consortium of 16 European partners

Our PhD students in industry (CIFRE programs) with companies DataHertz, Lokad, TEKLIA, Weinig Luxscan, Actemium, SOLYSTIC, Itekube

International Collaborations

• Heidelberg University (Germany), Linear optimization in integers
• SCK CEN - Belgian Center for Nuclear Energy (Belgium), geostatistiques
• Université Catholique de Louvain (Belgium), dimensionality reduction, medical imagery
• KU Leuven (Belgium), tensor decomposition
• Eindhoven TU (Netherland), weakly supervised learning, medical imaging
• Cornell University (United States of America), variable selection
• Rutgers University (United States of America), statistical model selection
• ÉTS de Montréal (Canada), handwritten and printed document recognition
• Universidade de Sao Paulo (Brazil), dictionary learning
• Pontifícia Universidade Católica do Paraná (Brazil), analysis of handwritten documents
• Indian Statistical Institute de Calcutta (India), biometrics and handwriting analysis

National Academic Collaborations

• INRIA Grenoble, (group Thoth), deep learning
• I3E, (group Maasai) deep learning
• LIS (ex. LIF, group QARMA), deep learning
• LIP6 (group MLIA), deep learning
• INRIA Nancy - Grand Est (group MultiSpeech), audio scenes recognition
• LTCI (group S2A), deep learning
• LIG (group APTIKAL), time series analysis
• MIP, sports gesture analysis
• Lagrange, (group SI), optimal transport
• IRISA (group OBELIX), optimal transport
• IRISA (group IntuiDoc), historical financial data analysis
• CAMS (group MSC), historical financial data analysis
• PSE, historical financial data analysis
• LARHRA, optical recognition of ancient documents
• CEA LIST, domain adaptation, video sequence analysis
• UMI UMMISCO, serious games

Academic Collaborations in Normandy

• GREYC (several research groups including Image et CoDaG), deep learning, imaging
• Centre François Baclesse, medical imaging
• Centre Henri Becquerel, medical imaging
• CERMN, chemoinformatics
• GMP, physical models
• CORIA, computational fluid mechanics, physical models
• UMR IDEES, serious games
• CETAPS, sports gesture analysis
• IRSEEM, road scene analysis
• CEREMA Normandie-Centre, road scene analysis
• and mathematics laboratories LMI et LMRS

Industrial Partners

• Groupe Hamelin, handwritten text recognition
• Toshiba, handwritten document analysis
• Saagie, MLOps
• Netatmo, sound scene analysis
• Atmo Normandie, time series analysis
• DataHertz, embedded machine learning
• Lokad, differentiable programming
• TEKLIA, document recognition
• Weinig Luxscan, transfer learning
• Actemium, road scene analysis
• SOLYSTIC, handwritten text recognition
• Itekube, human-machine interfaces
• Tellux, transfer learning, hyperspectral imaging

• Network of Norman Interest (RIN) “Normandie Digitale”
• Research federation NormaSTIC (FR CNRS 3638) with the GREYC lab
• AFRIF (Association Française pour la Reconnaissance et l'Interprétation des Formes)
• GRETSI (Groupe de Recherche et d'Etudes du Traitement du Signal et des Images)
• GRCE (Groupe de Recherche en Communication Ecrite)
• GdR ISIS (Information, Signal, Image et ViSion)
• SSFAM (Société Savante Francophone d'Apprentissage Machine)
• as well as GdR MaDICS et GdR IA

• PERMANENTS : 22
• PhD STUDENTS : ≈15
• POSTDOC : ≈4
• ASSOCIATED : 5

Our published papers are available on HAL.