General Program
Keynote Speaker
Detailed sessions
—————————–
General program of the day Booklet_Abstracts
| Time | Program |
|---|---|
| 8h15-8h45 | Welcome coffee |
| 8h45-8h50 | Opening |
| 8h50-10h10 | Oral Session 1 |
| 10h10-10h30 | Coffee break |
| 10h30-11h30 | Oral Session 2 |
| 11h30-11h45 | Sponsors presentation |
| 11h45-12h15 | GRCE general meeting |
| 12h15-1h15 | Lunch |
| 1h15-2h | Posters and demonstrations |
| 2h-3h | Keynote speaker: Elisa Fromont |
| 3h-3h20 | Coffee break |
| 3h20-4h20 | Oral Session 3 |
| 4h20-4h40 | Closing |
—————————–
Keynote speaker : Elisa FROMONT
Title : Explainable Time Series Classification
Abstract : I will introduce motivations about explainability in machine learning models and deep learning models as well as state-of-the-art methods. Then, I will present original works, especially on univariate or multivariate time series classification:
- Adversarial Regularization for Explainable-by-Design Time Series Classification – ICTAI 2020
- A Performance-Explainability Framework to Benchmark Machine Learning Methods: Application to Multivariate Time Series Classifiers – IJCAI Workshop on Explainable Artificial Intelligence (XAI), 2020
- XEM: An Explainable-by-Design Ensemble Method for Multivariate Time Series Classification, DAMI 2022
- XCM: An Explainable Convolutional Neural Network for Multivariate Time Series Classification, Mathematics
Bio : Elisa is a full professor at Université de Rennes 1 France, since 2017 and a Junior member of the Institut Universitaire de France (2019-2024). Elisa works at IRISA research institute in the INRIA LACODAM (“Large Scale Collaborative Data Mining”) team. From 2008 until 2017, she was associate professor at Université Jean Monnet in Saint-Etienne, France. She worked at the Hubert Curien research institute in the Data Intelligence team. Elisa received her Research Habilitation (HDR) in December 2015 from the University of Saint-Etienne. From 2006 until 2008, she was a postdoctoral researcher in the Machine Learning group of the Katholieke Universteit Leuven, Belgium. She received her PhD in December 2005 from Université de Rennes 1.
Research domains :
- Explainable AI (interpretability of data mining/machine learning models and explainable decisions)
- Data mining for time series analysis (including fraud and anomaly detection)
—————————–
Detailed program of sessions
Oral Session 1 (slides)
<td9h10-9h30
| Time | Title | Authors |
|---|---|---|
| 8h50-9h10 | Handwritting Recognition using Multi-Hypothesis based Semi-Supervised Learning for adaptation to specific documents | Alexandre Chapin, Yann Soullard, Aurélie Lemaître, Bertrand Couasnon (IRISA) |
| 9h10-9h30 | EAML: Ensemble Self-Attention-based Mutual Learning Network for Document Image Classification | Souhail Bakkali, Zuheng Ming, Mickael Coustaty, Marcal Rusinol, Oriol Ramos Terrades (L3I) |
| 9h30-9h50 | EXO-POPP : Extraction d’entités nommées dans des actes de mariage manuscrits | Thomas Constum, Pierrick Tranouez, Thierry Paquet (LITIS) |
| 9h50-10h10 | Localisation d’une fraude dans des documents | Aurélie Ling, Cuong van Kieu, Camille Kurtz, Nicole Vincent (LIPADE) |
Oral Session 2 (slides)
| Time | Title | Authors |
|---|---|---|
| 10h30-10h50 | Détection et association des queues de bulles : cas des multi-bulles | Christophe Rigaud (L3I), Jean-Christophe Burie (L3i), Samuel Petit (Geo Comix) |
| 10h50-11h10 | Open Source Maif pour une IA transparente & Maitrisée | Laure Bonafous et Sébastien Bidault (MAIF) |
| 11h10-11h30 | Arkindex pour l’expérimentation en traitement de documents numérisés | Mélodie Boillet, Solène Tarride et Christopher Kermorvant (TEKLIA) |
Oral Session 3 (slides
| Time | Title | Authors |
|---|---|---|
| 3h20-3h40 | Toward Deep neural network for pen trajectory reconstruction from kinematic sensors | Imbert Florent, Wassim Swaileh, Yann Soullard, Romain Tavenard, Eric Anquetil (IRISA) |
| 3h40-4h00 | EyCon: Computational Approaches to the Circulation of photographs in late 19th-/Early 20th-Century periodicals and beyond | Julien Schuh, Marina Giardinetti, Soumik Mallick et Daniel Foliard (LARCA) |
| 4h00-4h20 | Reconnaissance de texte dans les bandes dessinées par Smartphone | Clément Charrier, Frédéric Daubignard (Entreprise ALGONA), Rémy Leconge (Laboratoire PRISME), Frédéric Rayar (Laboratoire LIFAT), Sylvie Treuillet (PRISME) |
Session Posters et démonstrations – 1h15-2h
Posters
| Title | Authors | Affiliation |
|---|---|---|
| Leveraging Domain Knowledge for Document Information Extraction | Arthur Hemmer | L3I |
| Reconnaissance et segmentation de l’écriture manuscrite d’enfants | Simon Corbillé, Élisa Fromont, Éric Anquetil | IRISA |
| A comparative study for classification of French language corporate documents on imbalanced datasets | Tri-Cong Pham, Aurélie Joseph, Muriel Visani, Nicolas Sidere, Mickael Coustaty | L3I |
| Transformers pour la reconnaissance de documents anciens | Killian Barrere, Yann Soullard, Aurélie Lemaitre, Bertrand Coüasnon | IRISA |
| Information Extraction from Administrative Documents | Dipendra Sharma Kafle | L3I |
| Classification of Untranscribed Handwritten Notarial Documents by Textual Contents | José Ramon Prieto Fontcuberta | PRHLT |
| Détection de fraude de tickets de caisse en prenant en compte la sémantique | Beatriz Martínez Tornés | L3I |
| Transformation digitale d’images de documents techniques d’ingénierie et de plans industriels. | Mathieu François | LIRIS |
| Document Forgery Detection in the Context of Double JPEG Compression | Théo Taburet, Kais Rouis, Mickaël Coustaty, Petra Gomez-Krämer, Nicolas Sidère, Saddok Kébairi, Vincent Poulain d’Andecy | L3I |
Demonstrations (with or without poster)
| Title | Authors | Affiliation |
|---|---|---|
| Geo Comix Translator: a unique tool to speed-up translation process for comics, mangas and webtoons | Christophe Rigaud, Jean-Christophe Burie, Samuel Petit | L3I – Geo Comix |
| Arkindex pour l’expérimentation en traitement de documents numérisés | Mélodie Boillet, Solène Tarride et Christopher Kermorvant | TEKLIA |
| Open Source Maif pour une IA transparente & Maitrisée | Laure Bonafous and Sébastien Bidault | MAIF |