logo EDITE Patrick GALLINARI
Identité
Patrick GALLINARI
État académique
Thèse soutenue
Titulaire d'une HDR (ou équivalent) 1990-01-01
Laboratoire: personnel permanent
Direction de thèses (depuis 2007)
3.5
Propositions de sujets de thèse
Voisinage
Ellipse bleue: doctorant, ellipse jaune: docteur, rectangle vert: permanent, rectangle jaune: HDR. Trait vert: encadrant de thèse, trait bleu: directeur de thèse, pointillé: jury d'évaluation à mi-parcours ou jury de thèse.
Productions scientifiques
edite:133279232757
A Learning to Rank framework applied to text-image retrieval
2011
http://asso-aria.org/coria/2011/159.pdf
Apprentissage des schemas de propagation dans les multi-graphes
CORIA 2011
http://www.somabec.com/9782746231399/SEMANTIQUE_ET_MULTIMODALITE_EN_ANALYSE_DE_L_INFORMATION.htm?HERMES
Apprentissage automatique pour l'annotation d'images
2011
edite:1332792341127
Classification and Annotation in Social Corpora using Multiple Relations
CIKM 2011
edite:1332792345155
Datum-Wise Classification: A Sequential Approach to Sparsity
ECML/PKDD 2011
edite:1332792380248
Iterative Multi-Label Multi-Relational Classification Algorithm for Complex Social Networks
2011
edite:1332792381258
Learning efficient error correcting output codes for large hierarchical multi-class problems
Workshop on Large Scale Hierarchical Classification (at ECML) 2011
edite:1332792383265
Link Pattern Prediction with tensor decomposition in multi-relational networks
CIDM 2011
edite:1332792401344
Predicting Most Rated Items in Weekly Recommendation with Temporal Regression
This paper presents our approach to contextual recommendation for the Filmtipset Weekly Recommendation Track of the CAMRA 2010 Challenge[5]. The goal of this task is to predict which items will be rated by each user on specific weeks of the year, namely the week containing Christmas day, and the week leading up to the Oscars, based on ratings collected prior to the test period. Our approach aims at modeling the short-term evolution of the probability that an item is rated before each test period (in a user-independent way), and then forecasting these probabilities on the test week. To that end, we use a temporal regression technique providing non-personalized recommendation with better test performances than other non-personalized recommendation baselines. We then tried, with success, to generate time-dependent collaborative personalized recommendations providing us our best results.
ACM SIGKDD Workshop on Context-Aware Recommender Systems 2011
edite:1332792403362
Réseau de neurones profond et SVM pour la classification de sentiments
CORIA: COnférence en Recherche d'Information et Applications 2011
edite:1332792418409
Temporal Link Prediction by Integrating Content and Structure Information
CIKM 2011
edite:1332792418410
Text Classification: A Sequential Reading Approach
ECIR 2011
edite:1332792429480
A Ranking based Model for Automatic Annotation in a Social Network
ICWSM 2010 2010
edite:1332792435514
Apprentissage d'un Espace de Concepts de Mots pour une Nouvelle Représentation des Données Textuelles
Vol. 13, No. 1, pp. 63-82 2010
edite:1332792456625
Exploitation du contenu visuel pour améliorer la recherche textuelle d'images en ligne
Vol. 13, No. 1, pp. 187-209 2010
edite:1332792463657
Improving Document Clustering in a Learned Concept Space
Most document clustering algorithms operate in a high dimensional bag-of-words space. The inherent presence of noise in such representation obviously degrades the performance of most of these approaches. In this paper we investigate an unsupervised dimensionality reduction technique for document clustering. This technique is based upon the assumption that terms co-occurring in the same context with the same frequencies are semantically related. On the basis of this assumption we first find term clusters using a classification version of the EM algorithm. Documents are then represented in the space of these term clusters and a multinomial mixture model (MM) is used to build document clusters. We empirically show on four document collections, Reuters-21578, Reuters RCV2, 20-Newsgroups and WebKB, that this new text representation noticeably increases the performance of the MM model. By relating the proposed approach to the Probabilistic Latent Semantic Analysis (PLSA) model we further propose an extension of the latter in which an extra latent variable allows the model to co-cluster documents and terms simultaneously. We show on these four datasets that the proposed extended version of the PLSA model produces statistically significant improvements with respect to two clustering measures over all variants of the original PLSA and the MM models.
Vol. 46, No. 2, pp. 180-192 2010
edite:1332792463658
Improving Image Annotation in Imbalanced Classification Problems with Ranking SVM
Multilingual Information Access Evaluation Vol. II Multimedia Experiments 2010
edite:1332792477706
Modèles d'Ordonnancement pour l'Annotation Automatique d'Images dans les Réseaux Sociaux
CORIA 2010 2010
edite:1332792479716
Multiview Clustering of Multilingual Documents
Proceedings of the 33rd Annual ACM SIGIR Conference (SIGIR 2010) 2010
edite:1332792487753
Overview of the INEX 2009 XML Mining Track: Clustering and Classification of XML Documents
INEX 2009 2010
edite:1332792488771
Prediction de Motifs Relationnels par Decomposition Tensorielle dans les Reseaux Sociaux
Workshop Reiso 2010 2010
edite:1332792493799
Report on INEX 2009
Vol. 44, No. 1, pp. 38-57 2010
edite:1332792515917
A comparative study of diversity methods for hybrid text and image retrieval approaches
Evaluating Systems for Multilingual and Multimodal Information Access -- 9th Workshop of the Cross-Language Evaluation Forum 2009
edite:1332792523948
A self-training method for learning to rank with unlabeled data
Burges, Belgique 2009
edite:1332792527978
Apprentissage de fonctions d'ordonnancement avec un flux de données non-étiquetées
Hammamet, Tunisie 2009
http://www.clef-campaign.org/2009/working_notes/Glotin2_paperCLEF2009_VCDT_AVEIR.pdf.pdf
Comparison of Various AVEIR Visual Concept Detectors with an Index of Carefulness
CLEF working notes 2009 2009
edite:13327925421075
Exploiting Visual Concepts to Improve Text-Based Image Retrieval
European Conference on Information Retrieval (ECIR) 2009
edite:13327925601231
Probabilistic Multi-classifier by SVM from voting rule to voting features
pp. 2-2 2009
edite:13327925601239
Ranking with ordered weighted pairwise classification
Montreal, Quebec, Canada 2009
edite:13327925601242
Report on INEX 2008
Vol. 43, No. 1, pp. 17-36 2009
edite:13327925611256
SGD-QN: Careful Quasi-Newton Stochastic Gradient Descent
Vol. 10, pp. 1737----1754 2009
edite:13327925611257
SICA : Simulated Iterative Classification - A new larning method for graph labelling
ECML PKDD 2009 2009
edite:13327925621277
Structured Prediction with Reinforcement Learning
2009
edite:13327925661313
Une extension du modèle sémantique latent probabiliste pour le partitionnement non-supervisé de documents textuels
Hammamet, Tunisie 2009
edite:13327925661323
Using Visual Concepts and Fast Visual Diversity to Improve Image Retrieval
Evaluating Systems for Multilingual and Multimodal Information Access - 9th Workshop of the Cross-Language Evaluation Forum 2009
http://asso-aria.org/coria/2009/83.pdf
Utilisation de concepts visuels et de la diversité visuelle pour améliorer la recherche d'images
Actes de Conférence en Recherche d'Informations et Applications (CORIA'09) 2009