Research topics and projects
My broad research topics are data and graph mining. I have a long standing effort to work on dynamic graph models (e.g., link streams), to understand their properties (cliques mining) and apply them to several tasks such as recommender systems. A few years ago, I have started working on Graph Neural Networks, mostly in the context of recommender systems.
Besides that, I have been working on symbolic computer music for several years. We contribute to improving data mining on musical scores, aiming at providing better indexing/research, storage and analysis possibilities.
You will have a better idea of my works by looking at my publications, or feel free to ask me.
My research work is funded through several “projects”, detailed below.
There is a list of internship offers at the end of this page, with regular updates.
Ongoing research projects
Mining musical scores : MuNIR
MuNIR is a project aiming at offering a platform to index, search and analyse music scores. It aims at continuing the Neuma project, which was done by my colleague before I joined their team. It is now a partnership between the CNAM and researchers from McGill University.
DAPHNE
DAPHNE is an ANR project centered on studying the careers of European scholars in Paris centuries ago. My team and I work on developping quality software to help extract and visualize prosopographic information.
Genèse du CEDRIC
With colleagues in the History departement, we start a project on the genesis of my computer science lab. The main objective is to understand how a young research field (computer science) was gradually accepted as a legit one. We have a digital humanities perspective, analysing first-hand archive documents and testimonies.
Past research projects
AMMICO (2013-2016)
I was involved in the AMMICO project during my 18-month post-doctoral contract with Univ. Paris 13 and Emmanuel Viennet’s team. The project aims at bringing museums, hardware builders and AI and Physics research labs to design new handheld devices for enhancing museum visits. The core idea is to propose geolocation-based recommendations of artworks.
More details on the AMMICO project.
Antipaedo project (2009-2014)
The Antipaedo project funded my Ph. D. research years. There were a French Research Agency project (MAPAP) and a EU commission project (MAPE) around this project, but our effort went beyond those two projects. The aim was to obtain knowledge on paedophile activity in P2P networks. Several methods were used (server side measurements, honeypots, etc.) but most of the project relied on keyword-based queries analysis.
More details can be found on the Antipaedo website.
Post-doctoral positions
Offers
There are currently no funded PhD to be filled in my team. However, I accept
proposals by email, at fournier at cnam.fr
.
Current advised post-docs
- Denis Coquenet (2022-10-01 — , avec Nicolas Thome et Clément Rambour). Explicabilité des modèles transformers pour la vision et la recommandation.
Former post-docs
- Victor David (2021-09-13 — 2021-09-12, avec Nicolas Travers). Représentation de données prosopographiques par des graphes de connaissance.
- Pedro Ramaciotti-Morales (2018-09-01 — 2019-08-31, avec Lionel Tabourier). Diversité pour les Heterogeneous Information Networks.
- Tiphaine Viard (2016-09-01 — 2017-08-31, avec Matthieu Latapy). Graphes dynamiques et systèmes de recommandation.
Ph. D.
Offers
Voici la liste des sujets de doctorat que je propose en ce moment :
-
No funding
.
PDF description.
Co-encadrement avec Lionel Tabourier (LIP6, Sorbonne Universités/CNRS).
Je concentre évidemment mes efforts pour pourvoir les offres présentées ci-dessus. Néanmoins, si vous êtes intéressé(e) par une thèse sur un sujet proche de mes thématiques de recherche, contactez-moi par mail. Si je n’ai pas de financement à proposer, on pourra élaborer ensemble un dossier permettant d’en obtenir.
Current Ph.D.
- Yannis Karmim. Graphes dynamiques pour les Graph Neural Networks et la recommandation. 2021-09 —
- Tiange Zhu. Extraction de motifs et classification automatique de partitions musicales. 2021-02 —
Former Ph. D.
- Francesco Foscarin. Transcription musicale.
Du 2017-10-04 au 2020-12-10.
Co-encadrement avec Philippe Rigaux et Florent Jacquemard (INRIA/CNAM).
Internships
Offers
There are currently no pending funded internship offer. However, I accept
proposal by email for students at M2 level (end of engineering studies,
typically between March and August). Short internships at M1 level are possible
too. Send me a CV and a detailed email at fournier at cnam.fr
.
Former interns (automatic list, in French)
- Floriane Béhanzin, Université de Rouen. Graph Neural Networks et dynamique pour les systèmes de recommandation. Niveau M2. 6 mois, 2021.
- Chen Dang, Université Paris Dauphine. Web Image Context Extraction. Niveau M2. 6 mois, 2020.
- Mirwaisse Djanbaz, Université Paris Orsay. Systèmes de recommandation et graphes dynamiques. Niveau M2. 6 mois, 2019.
- Fahde Karim, Université de Reims Champagne-Ardenne. Conception d’une plateforme d’analyse et de visualisation de données prosopographiques. Niveau M2. 6 mois, 2016.
- Guillaume Souveton, Université de St-Etienne. Sélection de descripteurs pour la fouille de partitions musicales. Niveau M2. 6 mois, 2016.
- Berrenur Saylam, ENS Lyon. Indicateurs de diversité dans les graphes de recommandation. Niveau M2. 6 mois, 2016.
- Laure Fouard, Université Paris VII. Évolution temporelle de l'utilisation de mots-clefs : analyse de traces P2P. Niveau M2. 4 mois, 2011.
Scientific community
Member of the PC for the following conferences:
- WSDM 2023 (ACM Intl Conf. on Web Search and Data Mining)
- TheWebConf (WWW2022)
- GEM - Graph Embedding and Mining @ ECML-PKDD (2020)
- MARAMI (2014)
- HaPoP-3 (2016)
Peer-review for:
- SNAM (2016)
- IJAIT (2015)
- Child Abuse & Neglect (2014)
- NetSciCom (2013)
- TRAC (2013)
- WWW (2012)
- RCIS (2014)