Publications

Autres listes de mes publications

Revues à comité de lecture

Journal ITEM: Improving Training and Evaluation of Message-Passing based GNNs for top-k recommendation. Karmim, Y., Ramzi, E., Fournier-S’niehotta, R., & THOME, N. (2024). Transactions on Machine Learning Research. Retrieved from https://openreview.net/forum?id=9B6LM2uoEs.
```
@article{karmim2024item,
  file = {/bibliography/files/2024-ITEM_KarmimRamziFournierSniehottaThome.pdf},
  title = {{ITEM}: Improving Training and Evaluation of Message-Passing based {GNN}s for top-k recommendation},
  author = {Karmim, Yannis and Ramzi, Elias and Fournier-S'niehotta, Raphael and THOME, Nicolas},
  journal = {Transactions on Machine Learning Research},
  issn = {2835-8856},
  year = {2024},
  url = {https://openreview.net/forum?id=9B6LM2uoEs},
  note = {}
}
```

Journal A Framework for Content-Based Search in Large Music Collections. Zhu, T., Fournier-S’niehotta, R., Rigaux, P., & Travers, N. (2022). Big Data and Cognitive Computing, 6(1). DOI: 10.3390/bdcc6010023.

@article{bdcc6010023,
  author = {Zhu, Tiange and Fournier-S’niehotta, Raphaël and Rigaux, Philippe and Travers, Nicolas},
  title = {A Framework for Content-Based Search in Large Music Collections},
  journal = {Big Data and Cognitive Computing},
  volume = {6},
  year = {2022},
  number = {1},
  url = {https://www.mdpi.com/2504-2289/6/1/23},
  file = {/bibliography/files/ZhuFournierRigauxTravers-MusicSearch-BDCC22.pdf},
  issn = {2504-2289},
  article-number = {23},
  doi = {10.3390/bdcc6010023}
}

We address the problem of scalable content-based search in large collections of music documents. Music content is highly complex and versatile and presents multiple facets that can be considered independently or in combination. Moreover, music documents can be digitally encoded in many ways. We propose a general framework for building a scalable search engine, based on (i) a music description language that represents music content independently from a specific encoding, (ii) an extendible list of feature-extraction functions, and (iii) indexing, searching, and ranking procedures designed to be integrated into the standard architecture of a text-oriented search engine. As a proof of concept, we also detail an actual implementation of the framework for searching in large collections of XML-encoded music scores, based on the popular ElasticSearch system. It is released as open-source in GitHub, and available as a ready-to-use Docker image for communities that manage large collections of digitized music documents.

Journal Measuring diversity in heterogeneous information networks. Ramaciotti Morales, P., Lamarche-Perrin, R., Fournier-S’niehotta, R., Poulain, R., Tabourier, L., & Tarissan, F. (2021). Theoretical Computer Science, 859, 80–115. DOI: https://doi.org/10.1016/j.tcs.2021.01.013.

@article{RAMACIOTTIMORALES202180,
  title = {Measuring diversity in heterogeneous information networks},
  journal = {Theoretical Computer Science},
  volume = {859},
  pages = {80-115},
  year = {2021},
  issn = {0304-3975},
  doi = {https://doi.org/10.1016/j.tcs.2021.01.013},
  url = {https://www.sciencedirect.com/science/article/pii/S0304397521000256},
  author = {{Ramaciotti Morales}, Pedro and Lamarche-Perrin, Robin and Fournier-S'niehotta, Raphaël and Poulain, Rémy and Tabourier, Lionel and Tarissan, Fabien},
  keywords = {Diversity measures, Heterogeneous information networks, Random walks on graphs, Recommender systems, Social network analysis},
  file = {/bibliography/files/Diversity-HIN-2020-Ramaciotti-TCS.pdf}
}

Diversity is a concept relevant to numerous domains of research varying from ecology, to information theory, and to economics, to cite a few. It is a notion that is steadily gaining attention in the information retrieval, network analysis, and artificial neural networks communities. While the use of diversity measures in network-structured data counts a growing number of applications, no clear and comprehensive description is available for the different ways in which diversities can be measured. In this article, we develop a formal framework for the application of a large family of diversity measures to heterogeneous information networks (HINs), a flexible, widely-used network data formalism. This extends the application of diversity measures, from systems of classifications and apportionments, to more complex relations that can be better modeled by networks. In doing so, we not only provide an effective organization of multiple practices from different domains, but also unearth new observables in systems modeled by heterogeneous information networks. We illustrate the pertinence of our approach by developing different applications related to various domains concerned by both diversity and networks. In particular, we illustrate the usefulness of these new proposed observables in the domains of recommender systems and social media studies, among other fields.

Journal Augmenting content-based rating prediction with link stream features. Viard, T., & Fournier-S’niehotta, R. (2019). Computer Networks, 150, 127–133. DOI: https://doi.org/10.1016/j.comnet.2018.12.002.

@article{VIARD2019127,
  file = {/bibliography/files/AugmentingCBRatingPredictionLinkStreamFeatures_ViardFournier_ComNet.pdf},
  title = {Augmenting content-based rating prediction with link stream features},
  journal = {Computer Networks},
  volume = {150},
  pages = {127 - 133},
  year = {2019},
  issn = {1389-1286},
  doi = {https://doi.org/10.1016/j.comnet.2018.12.002},
  url = {http://www.sciencedirect.com/science/article/pii/S1389128618313215},
  author = {Viard, Tiphaine and Fournier-S’niehotta, Raphaël},
  keywords = {Link streams, Recommender systems, Interpretability, Temporal networks, Graphs}
}

While graph-based collaborative filtering recommender systems have been introduced several years ago, there are still several shortcomings to deal with, the temporal information being one of the most important. The new link stream paradigm is aiming at extending graphs for correctly modelling the graph dynamics, without losing crucial information. We investigate the impact of such link stream features for recommender systems. We design link stream features, that capture the intrinsic structure and dynamics of the data. We show that such features encode a fine-grained and subtle description of the underlying system. We focused on a traditional recommender system context, the rating prediction on the MovieLens20M movie dataset and the Goodreads book dataset. We input link stream features along with some content-based ones into a gradient boosting machine (XGBoost) and show that itoutperforms significantly a sole content-based solution. These encouraging results call for further exploration of this original modelling and its integration to complete state-of-the-art recommender systems algorithms. Link streams and graphs, as natural visualizations of recommender systems, may offer more interpretability in a time when algorithm transparency is an increasingly important topic of discussion. We also hope that these results will sparkle interesting discussions in the community about the connections between link streams and traditional methods (matrix factorization, deep learning).

Journal Modeling Music as Synchronized Time Series: Application to Music Score Collections. Fournier-S’niehotta, R., Rigaux, P., & Travers, N. (2017). (IS’18) Information Systems, 1–36. DOI: 10.1016/j.is.2017.12.003.

@article{FRT17,
  author = {Fournier-S'niehotta, Raphaël and Rigaux, Philippe and Travers, Nicolas},
  journal = {(IS'18) Information Systems},
  title = {{Modeling Music as Synchronized Time Series: Application to Music Score Collections}},
  year = {2017},
  pages = {1--36},
  doi = {10.1016/j.is.2017.12.003},
  file = {:/bibliography/files/ModelingMusicSynchronizedTimeSeriesApplicationMusicScoreCollections_FournierSniehottaTraversRigaux_IS17.pdf:},
  url = {https://doi.org/10.1016/j.is.2017.12.003}
}

Music is a temporal organization of sounds, and we can therefore assume that any music representation has a structure that reflects some conceptual principles. This structure is hardly explicitly accessible in many encodings, such as, for instance, audio files. However, it appears much more clearly in the language of music notation.

              We propose to use the music notation language as a framework to model and manipulate the content of digital music objects, whatever their specific encoding may be. We describe an algebra that relies on this structured music representation to extract, restructure, and search such objects. The data model leverages the hidden structure of digital music encodings to enable powerful manipulations of their content.

              We apply the model to collections of music scores. We describe a system, based
              on an extension of XQuery with our algebra, that provides search,
              reorganization, and extraction functionalities on top of large collections of
              XML-encoded digital scores. Beyond its application to music objects, our work
              shows how one can rely on a structured content embedded in a complex XML
              encoding to develop robust collection management tools with minimal
              implementation effort.

Journal Temporal Patterns of Pedophile Activity in a P2P Network: First Insights about User Profiles from Big Data. Fournier-S’niehotta, R., & Latapy, M. (2015). International Journal of Internet Science, 10(1), 8–19. Retrieved from http://www.ijis.net/ijis_inpress/ijis_inpress_Fournier_and_Latapy_pre.html.

@article{FournierL15,
  author = {Fournier, Raphaël and Latapy, Matthieu},
  journal = {International Journal of Internet Science},
  title = {Temporal Patterns of Pedophile Activity in a {P2P} Network: First Insights about User Profiles from Big Data},
  year = {2015},
  issn = {1662-5544},
  number = {1},
  pages = {8--19},
  volume = {10},
  file = {/bibliography/files/TemporalPattersPedoActP2P_Fournier-Latapy_IJIS2015.pdf},
  url = {http://www.ijis.net/ijis_inpress/ijis_inpress_Fournier_and_Latapy_pre.html}
}

Recent studies have shown that child abuse material is shared
              through peer-to-peer (P2P) networks, which allow users to exchange files
              without a central server. Obtaining knowledge on the extent of this
              activity has major consequences for child protection, policy making and
              Internet regulation. Previous works have developed tools and analyses to
              provide overall figures in temporally-limited measurements. Offenders’
              behavior is mostly studied through small-scale interviews and there is few
              information on the times at which they engage in such activity. Here we
              show that the proportion of search-engine queries for pedophile content
              gradually has grown by a factor of almost 3 in three years. We also find
              that during the day, certain hours are, on average, privileged by seekers.
              Our results demonstrate that P2P networks are actively used to search for
              pedophile content and we find new and large-scale results on pedophile
              offenders’ profile, indicating that a substantial proportion is
              well-integrated into family life and professional work activities.

Journal Comparing paedophile activity in different P2P systems. Fournier-S’niehotta, R., Cholez, T., Latapy, M., Chrisment, I., Magnien, C., Daniloff, I., & Festor, O. (2014). Social Sciences, 3(3). DOI: 10.3390/socsci3030314.

@article{FournierCLCMDF14,
  file = {/bibliography/files/ComparingPaedoActP2P_FournierCholezLatapyMagnienChrismentDaniloffFestor.pdf},
  title = {Comparing paedophile activity in different {P2P} systems},
  author = {Fournier, Raphaël and Cholez, Thibault and Latapy, Matthieu and Chrisment, Isabelle and Magnien, Clémence and Daniloff, Ivan and Festor, Olivier},
  journal = {Social Sciences},
  volume = {3},
  year = {2014},
  number = {3},
  url = {\url{http://www.mdpi.com/2076-0760/3/3/314}},
  issn = {2076-0760},
  doi = {10.3390/socsci3030314}
}

Peer-to-peer (P2P) systems are widely used to exchange content
              over the Internet. Knowledge of pedophile activity in such networks remains
              limited, despite having important social consequences. Moreover, though
              there are different P2P systems in use, previous academic works on this
              topic focused on one system at a time and their results are not directly
              comparable. We design a methodology for comparing KAD and eDonkey, two P2P
              systems among the most prominent ones and with different anonymity levels.
              We monitor two eDonkey servers and the KAD network during several days and
              record hundreds of thousands of keyword-based queries. We detect
              pedophile-related queries with a previously validated tool and we propose,
              for the first time, a large-scale comparison of pedophile activity in two
              different P2P systems. We conclude that there are significantly fewer
              pedophile queries in KAD than in eDonkey (approximately 0.09% vs. 0.25%).

Journal A Social Formalism and Survey for Recommender Systems. Bernardes, D., Diaby, M., Fournier-S’niehotta, R., Fogelman-Soulié, F., & Viennet, E. (2014). SIGKDD Explorations, 16(2), 20–37. DOI: 10.1145/2783702.2783705.

@article{BernardesDFFV14,
  author = {Bernardes, Daniel and Diaby, Mamadou and Fournier, Rapha{\"{e}}l and Fogelman{-}Souli{\'{e}}, Fran{\c{c}}oise and Viennet, Emmanuel},
  journal = {SIGKDD Explorations},
  title = {A Social Formalism and Survey for Recommender Systems},
  year = {2014},
  number = {2},
  pages = {20--37},
  volume = {16},
  bibsource = {dblp computer science bibliography, http://dblp.org},
  biburl = {\url{http://dblp.uni-trier.de/rec/bib/journals/sigkdd/BernardesDFFV14}},
  doi = {10.1145/2783702.2783705},
  file = {/bibliography/files/SocialFormalismSurveyRS-BernardesDiabyFournierViennetFogelman.pdf},
  timestamp = {Tue, 23 Jun 2015 10:19:35 +0200},
  url = {\url{http://doi.acm.org/10.1145/2783702.2783705}}
}

We present a general formalism for Recommender Systems based on
               Social Network Analysis. After introducing the classical categories of
               recommender systems, we present our Social Filtering formalism and show
               that it extends association rules, classical Collaborative Filtering and
               Social Recommendation, while providing additional possibilities. This
               allows us to survey the literature and illustrate the versatility of our
               approach on various publicly available datasets, comparing our results
               with the literature.

Journal Quantifying paedophile activity in a large P2P system. Latapy, M., Magnien, C., & Fournier-S’niehotta, R. (2013). Information Processing and Management, 49(1), 248–263. DOI: 10.1016/j.ipm.2012.02.008.

@article{LatapyMF13,
  author = {Latapy, Matthieu and Magnien, Cl{\'e}mence and Fournier, Rapha{\"e}l},
  journal = {Information Processing and Management},
  title = {Quantifying paedophile activity in a large P2P system},
  year = {2013},
  number = {1},
  pages = {248--263},
  volume = {49},
  biburl = {\url{http://dblp.uni-trier.de/rec/bib/journals/ipm/LatapyMF13}},
  doi = {10.1016/j.ipm.2012.02.008},
  file = {/bibliography/files/ipm.pdf},
  timestamp = {Fri, 30 Nov 2012 10:28:44 +0100},
  url = {\url{http://dx.doi.org/10.1016/j.ipm.2012.02.008}}
}

Increasing knowledge of paedophile activity in P2P systems is a
               crucial societal concern, with important consequences on child protection,
               policy making, and internet regulation. Because of a lack of traces of P2P
               exchanges and rigorous analysis methodology, however, current knowledge of
               this activity remains very limited. We consider here a widely used P2P
               system, eDonkey, and focus on two key statistics: the fraction of
               paedophile queries entered in the system and the fraction of users who
               entered such queries. We collect hun- dreds of millions of keyword-based
               queries; we design a paedophile query detection tool for which we
               establish false positive and false negative rates using assessment by
               experts; with this tool and these rates, we then estimate the fraction of
               paedophile queries in our data; ﬁnally, we design and apply methods for
               quantifying users who entered such queries. We conclude that approximately
               0.25% of queries are paedophile, and that more than 0.2% of users enter
               such queries. These statistics are by far the most precise and reliable
               ever obtained in this domain.

Conférences

Doctorat

Thesis Modéliser avec des graphes, pour l’informatique musicale et les systèmes de recommandation (Habilitation à diriger des recherches). Fournier-S’niehotta, R. (2024). CNAM Paris.
```
@phdthesis{Fournier-Sniehotta2024,
  ids = {mon-hdr},
  title = {Modéliser avec des graphes, pour l'informatique musicale et les systèmes de recommandation},
  author = {Fournier-S'niehotta, Raphaël},
  year = {2024},
  school = {CNAM Paris},
  type = {Habilitation à diriger des recherches}
}
```
Thesis Détection et analyse d’une thématique rare dans de grands ensembles de requêtes : l’activité pédophile dans le P2P. PhD thesis. Fournier-S’niehotta, R. (2012). UPMC. Retrieved from https://tel.archives-ouvertes.fr/tel-02444174
```
@phdthesis{Fournier-Sniehotta2012,
  file = {/bibliography/files/These_FournierRaphael_vfinale.pdf},
  title = {Détection et analyse d'une thématique rare dans de grands ensembles de requêtes : l'activité pédophile dans le {P2P}},
  author = {Fournier-S'niehotta, Raphaël},
  year = {2012},
  school = {UPMC},
  url = {https://tel.archives-ouvertes.fr/tel-02444174}
}
```

Livres

Book Informatique. Delacroix, J., Barthélémy, F., Fournier-S’niehotta, R., Gil-Michalon, I., Lambert, A., Plateau, A., … Waymel, E. (2017) (1e ed.). Paris : Dunod. Retrieved from https://www.dunod.com/sciences-techniques/informatique.

@book{dunodL1,
  author = {Delacroix, Joëlle and Barthélémy, François and Fournier-S'niehotta, Raphaël and Gil-Michalon, Isabelle and Lambert, Amélie and Plateau, Agnès and Rovédakis, Stéphane and Simonot, Marianne and Thion, Virginie and Waymel, Emmanuel},
  publisher = {Dunod},
  title = {Informatique},
  year = {2017},
  address = {Paris},
  edition = {1e},
  isbn = {3257227892},
  month = jun,
  series = {Fluoresciences},
  url = {https://www.dunod.com/sciences-techniques/informatique}
}