 
    Marc Tommasi
                                    76 
                                    
                            Documents
                                Présentation
Publications
                    
                    
                    7
                
                            
                    
                    
                    5
                
                            
                    
                    
                    5
                
                            
                    
                    
                    5
                
                            
                    
                    
                    4
                
                            
                    
                    
                    4
                
                            
                    
                    
                    4
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                    
                    
                    
                    25
                
                            
                    
                    
                    22
                
                            
                    
                    
                    10
                
                            
                    
                    
                    8
                
                            
                    
                    
                    8
                
                            
                    
                    
                    5
                
                            
                    
                    
                    5
                
                            
                    
                    
                    5
                
                            
                    
                    
                    5
                
                            
                    
                    
                    4
                
                            
                    
                    
                    4
                
                            
                    
                    
                    4
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    3
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    2
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                            
                    
                    
                    1
                
                    
                    
                    
                    40
                
                            
                    
                    
                    24
                
                            
                    
                    
                    16
                
                            
                    
                    
                    16
                
                            
                    
                    
                    9
                
                            
                    
                    
                    7
                
                            
                    
                    
                    2
                
                    Publications
|  | Exploiting Context-dependent Duration Features for Voice Anonymization Attack SystemsInterspeech 2025, Aug 2025, Rotterdam, NetherlandsCommunication dans un congrès
                    
        
        
            
                
                    hal-05099074
                                            v1 | 
|  | Privacy Amplification Through Synthetic Data: Insights from Linear RegressionICML 2025 - 42nd International Conference on Machine Learning, 2025, Vancouver, Canada. ⟨10.48550/arXiv.2506.05101⟩Communication dans un congrès
                    
        
        
            
                
                    hal-05237878
                                            v1 | 
|  | Analysis of Speech Temporal Dynamics in the Context of Speaker Verification and Voice Anonymization2025 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2025), Apr 2025, Hyderabad, India. ⟨10.1109/ICASSP49660.2025.10887896⟩Communication dans un congrès
                    
        
        
            
                
                    hal-04853872
                                            v1 | 
|  | TAROT: Task-Oriented Authorship Obfuscation Using Policy Optimization MethodsProceedings of the Sixth Workshop on Privacy in Natural Language Processing, Apr 2025, Albuquerque, United States. pp.14-31, ⟨10.18653/v1/2025.privatenlp-main.2⟩Communication dans un congrès
                    
        
        
            
                
                    hal-05299966
                                            v1 | 
|  | Rényi Pufferfish Privacy: General Additive Noise Mechanisms and Privacy Amplification by Iteration via Shift Reduction LemmasInternational Conference on Machine Learning (ICML 2024), 2024, Vienna (Austria), AustriaCommunication dans un congrès
                    
        
        
            
                
                    hal-04363020
                                            v2 | 
|  | Improved Stability and Generalization Guarantees of the Decentralized SGD AlgorithmICML 2024 - The Forty-first International Conference on Machine Learning, Jul 2024, Vienne, AustriaCommunication dans un congrès
                    
        
        
            
                
                    hal-04611418
                                            v1 | 
|  | Refined Convergence and Topology Learning for Decentralized SGD with Heterogeneous DataProceedings of The 26th International Conference on Artificial Intelligence and Statistics (AISTATS 2023), 2023, Valencia, Spain, SpainCommunication dans un congrès
                    
        
        
            
                
                    hal-03905091
                                            v2 | 
|  | High-Dimensional Private Empirical Risk Minimization by Greedy Coordinate DescentAISTATS 2023 - International Conference on Artificial Intelligence and Statistics, Apr 2023, Valencia, Spain. pp.4894-4916Communication dans un congrès
                    
        
        
            
                
                    hal-03714465
                                            v3 | 
|  | Differential Privacy has Bounded Impact on Fairness in ClassificationInternational Conference on Machine Learning, Jul 2023, Honolulu, United StatesCommunication dans un congrès
                    
        
        
            
                
                    hal-03902203
                                            v2 | 
|  | FLamby: Datasets and Benchmarks for Cross-Silo Federated Learning in Realistic Healthcare SettingsNeurIPS 2022 - Thirty-sixth Conference on Neural Information Processing Systems, Nov 2022, New Orleans, United StatesCommunication dans un congrès
                    
        
        
            
                
                    hal-03900026
                                            v1 | 
|  | Extraction d'informations liées au locuteur depuis un modèle acoustique personnaliséJEP 2022, Jun 2022, île de Noirmoutier, FranceCommunication dans un congrès
                    
        
        
            
                
                    hal-03706944
                                            v1 | 
|  | Differentially Private Coordinate Descent for Composite Empirical Risk MinimizationICML 2022 - 39th International Conference on Machine Learning, Jul 2022, Baltimore, United States. pp.14948-14978Communication dans un congrès
                    
        
        
            
                
                    hal-03424974
                                            v3 | 
|  | Enhancing speech privacy with slicingInterspeech 2022 - Human and Humanizing Speech Technology, Sep 2022, Incheon, South KoreaCommunication dans un congrès
                    
        
        
            
                
                    hal-03369137
                                            v2 | 
|  | Retrieving Speaker Information from Personalized Acoustic Models for Speech RecognitionIEEE ICASSP 2022, 2022, Singapour, SingaporeCommunication dans un congrès
                    
        
        
            
                
                    hal-03539741
                                            v1 | 
|  | On speaker verification from the neural network footprint of personalized acoustic modelsJournées d'Études sur la Parole - JEP2022, Jun 2022, Île de Noirmoutier, FranceCommunication dans un congrès
                    
        
        
            
                
                    hal-03626964
                                            v1 | 
|  | Privacy attacks for automatic speech recognition acoustic models in a federated learning frameworkICASSP 2022, 2022, Singapour, SingaporeCommunication dans un congrès
                    
        
        
            
                
                    hal-03539742
                                            v2 | 
|  | Study on Acoustic Model Personalization in a Context of Collaborative Learning Constrained by Privacy PreservationSPECOM 2021 - 23rd International Conference on Speech and Computer, Sep 2021, St Petersburg, Russia. pp.426 - 436, ⟨10.1007/978-3-030-87802-3_39⟩Communication dans un congrès
                    
        
        
            
                
                    hal-03369206
                                            v1 | 
|  | Design Choices for X-vector Based Speaker AnonymizationINTERSPEECH 2020, International Speech Communication Association (ISCA), Oct 2020, Shanghai, ChinaCommunication dans un congrès
                    
        
        
            
                
                    hal-02610447
                                            v2 | 
|  | Fully Decentralized Joint Learning of Personalized Models and Collaboration GraphsAISTATS 2020 - The 23rd International Conference on Artificial Intelligence and Statistics, Aug 2020, Palerme / Virtual, ItalyCommunication dans un congrès
                    
        
        
            
                
                    hal-03100057
                                            v1 | 
|  | Evaluating Voice Conversion-based Privacy Protection against Informed AttackersICASSP 2020 - 45th International Conference on Acoustics, Speech, and Signal Processing, IEEE Signal Processing Society, May 2020, Barcelona, Spain. pp.2802-2806Communication dans un congrès
                    
        
        
            
                
                    hal-02355115
                                            v2 | 
|  | A comparative study of speech anonymization metricsINTERSPEECH 2020, Oct 2020, Shanghai, ChinaCommunication dans un congrès
                    
        
        
            
                
                    hal-02907918
                                            v1 | 
|  | Privacy-Preserving Adversarial Representation Learning in ASR: Reality or Illusion?INTERSPEECH 2019 - 20th Annual Conference of the International Speech Communication Association, Sep 2019, Graz, AustriaCommunication dans un congrès
                    
        
        
            
                
                    hal-02166434
                                            v1 | 
| Communication-Efficient Decentralized Boosting while Discovering the Collaboration GraphNeurIPS 2018 - Workshop on Machine Learning on the Phone and other Consumer Devices, 2nd Workshop on Machine Learning on the Phone and other Consumer Devices (MLPCD 2), Dec 2018, Montréal, CanadaCommunication dans un congrès
                    
        
        
            
                
                    hal-02166758
                                            v1 | |
|  | Personalized and Private Peer-to-Peer Machine LearningAISTATS 2018 - 21st International Conference on Artificial Intelligence and Statistics, Apr 2018, Lanzarote, Spain. pp.1-20Communication dans un congrès
                    
        
        
            
                
                    hal-01745796
                                            v1 | 
| Personalized and Private Peer-to-Peer Machine LearningNIPS 2017 - Workshop on Machine Learning on the Phone and other Consumer Devices, Dec 2017, Long Beach, United StatesCommunication dans un congrès
                    
        
        
            
                
                    hal-01665422
                                            v1 | |
|  | Decentralized Collaborative Learning of Personalized Models over NetworksInternational Conference on Artificial Intelligence and Statistics (AISTATS), Apr 2017, Fort Lauderdale, Florida., United StatesCommunication dans un congrès
                    
        
        
            
                
                    hal-01533182
                                            v1 | 
|  | Learning Sequential Tree-to-Word Transducers8th International Conference on Language and Automata Theory and Applications, Mar 2014, Madrid, SpainCommunication dans un congrès
                    
        
        
            
                
                    hal-00912969
                                            v1 | 
|  | Hypernode Graphs for Spectral Learning on Binary Relations over SetsConférence Francophone sur l'Apprentissage Automatique (Cap 2014), Jul 2014, Saint-Etienne, France. ⟨10.1007/978-3-662-44851-9_42⟩Communication dans un congrès
                    
        
        
            
                
                    hal-01104618
                                            v1 | 
|  | Clustering Spectral avec Contraintes de Paires réglées par Noyaux GaussiensCAP 2014, Jul 2014, Saint-Etienne, FranceCommunication dans un congrès
                    
        
        
            
                
                    hal-01105339
                                            v1 | 
|  | Fast Gaussian Pairwise Constrained Spectral ClusteringECML/PKDD - 7th European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, Sep 2014, Nancy, France. pp.242 - 257, ⟨10.1007/978-3-662-44848-9_16⟩Communication dans un congrès
                    
        
        
            
                
                    hal-01017269
                                            v1 | 
|  | Hypernode Graphs for Spectral Learning on Binary Relations over SetsECML/PKDD - 7th European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, Sep 2014, Nancy, FranceCommunication dans un congrès
                    
        
        
            
                
                    hal-01017025
                                            v1 | 
|  | Learning from Multiple Graphs using a Sigmoid KernelThe 12th International Conference on Machine Learning and Applications (ICMLA'13), Dec 2013, Miami, United StatesCommunication dans un congrès
                    
        
        
            
                
                    hal-00913237
                                            v1 | 
|  | Spectral Estimation of Conditional Random Graph Models for Large-Scale Network dataUAI 2012 - 28th Conference on Uncertainty in Artificial Intelligence, 2012, Avalon, United StatesCommunication dans un congrès
                    
        
        
            
                
                    hal-00714446
                                            v1 | 
|  | Fiedler Random Fields: A Large-Scale Spectral Approach to Statistical Network ModelingNeural Information Processing Systems (NIPS), Dec 2012, Lake Tahoe, United StatesCommunication dans un congrès
                    
        
        
            
                
                    hal-00750345
                                            v1 | 
|  | Normalization of Sequential Top-Down Tree-to-Word Transducers5th International Conference on Language Automata Theory and Appliciations, May 2011, Tarragona, SpainCommunication dans un congrès
                    
        
        
            
                
                    inria-00566291
                                            v1 | 
|  | Relevant Representations for the Inference of Rational Stochastic Tree LanguagesInternational Colloquium on Grammatical Inference, 2008, St Malo, France. pp.57-70Communication dans un congrès
                    
        
        
            
                
                    hal-00293511
                                            v3 | 
| Automatic Wrapper Induction from Hidden-Web Sources with Domain KnowledgeInternational Workshop on Web information and data managment, Oct 2008, Napa, United States. pp.9-16Communication dans un congrès
                    
        
        
            
                
                    inria-00337098
                                            v1 | |
|  | On Probability Distributions for Trees: Representations, Inference and LearningNIPS Workshop on Representations and Inference on Probability Distributions, Dec 2007, Whistler, CanadaCommunication dans un congrès
                    
        
        
            
                
                    inria-00294636
                                            v1 | 
|  | Champs Conditionnels Aléatoires pour l'Annotation d'ArbresCAp 2006, May 2006, Trégastel, FranceCommunication dans un congrès
                    
        
        
            
                
                    inria-00117014
                                            v1 | 
| XML Document Transformation with Conditional Random FieldsINEX 2006, Dec 2006, Dagstuhl, GermanyCommunication dans un congrès
                    
        
        
            
                
                    inria-00147052
                                            v1 | |
|  | Interactive Tuples Extraction from Semi-Structured DataWeb Intelligence, Dec 2006, Hong Kong, ChinaCommunication dans un congrès
                    
        
        
            
                
                    inria-00581253
                                            v1 | 
|  | Conditional Random Fields for XML TreesWorkshop on Mining and Learning in Graphs, Sep 2006, Berlin, GermanyCommunication dans un congrès
                    
        
        
            
                
                    inria-00118761
                                            v1 | 
|  | Querying Unranked Trees with Stepwise Tree Automata19th International Conference on Rewriting Techniques and Applications, 2004, Aachen, Georgia. pp.105--118Communication dans un congrès
                    
        
        
            
                
                    inria-00536529
                                            v1 | 
|  | Residual Finite Tree Automata7th International Conference on Developments in Language Theory, Jul 2003, Szeged, Hungary, Hungary. pp.171 -- 182Communication dans un congrès
                    
        
        
            
                
                    inria-00091272
                                            v2 | 
|  | Learning Multi-label Alternating Decision Trees from Texts and DataInternational Conference on Machine Learning and Data Mining, 2003, Leipzig, Georgia. pp.35-49Communication dans un congrès
                    
        
        
            
                
                    inria-00536733
                                            v1 | 
| Text Classification from Positive and Unlabeled ExamplesProceedings of the 9th International Conference on Information Processing and Management of Uncertainty in Knowledge-Based Systems, IPMU'02, 2002, Grenoble, France. pp.1927--1934Communication dans un congrès
                    
        
        
            
                
                    inria-00538889
                                            v1 | |
| Homomorphisms and concurrent term rewritingFCT 1999 - International Symposium on Fundamentals of Computation Theory, Aug 1999, Iasi, Romania. pp.475-487, ⟨10.1007/3-540-48321-7_40⟩Communication dans un congrès
                    
                    istex
        
        
            
                
                    hal-02193779
                                            v1 | |
| Deciding the Satisfiability of Quantifier free Formulae on One-Step RewritingRewriting Techniques and Applications, 10th International Conference, RTA-99, Trento, Italy, July 2-4, 1999, Proceedings, 1999, Unknown, Unknown Region. pp.103--117, ⟨10.1007/3-540-48685-2_9⟩Communication dans un congrès
                    
                    istex
        
        
            
                
                    hal-01820502
                                            v1 | |
| Some new Decidability Results on Positive and Negative Set ConstraintsProceedings of First International Conference on Constraints in Computational Logics, CCL'94, 1994, Munich, Germany. pp.336--351Communication dans un congrès
                    
        
        
            
                
                    inria-00538880
                                            v1 | |
| Solving Systems of Set Constraints using Tree AutomataProceedings of the 10th Annual Symposium on Theoretical Aspects of Computer Science, STACS'93, 1993, Würzburg, Germany. pp.505--514Communication dans un congrès
                    
        
        
            
                
                    inria-00538878
                                            v1 | |
| Solving Systems of Set Constraints with Negated Subset RelationshipsProceedings of the $34^th$ Symposium on Foundations of Computer Science, FOCS'93, 1993, Palo Alto California, United States. pp.372--380Communication dans un congrès
                    
        
        
            
                
                    inria-00538879
                                            v1 | 
|  | Hypernode Graphs for Learning from Binary Relations between Groups in NetworksNetworks: From Graphs to Rich Data, NIPS Workshop., Dec 2014, Montreal, Canada. , 2014Poster de conférence
                    
        
        
            
                
                    hal-01088036
                                            v1 | 
|  | Tree Automata Techniques and Applicationspp.262, 2008Ouvrages
                    
        
        
            
                
                    hal-03367725
                                            v1 | 
| Champs Markoviens Conditionnels pour l'extraction d'informationEric Gaussier and François Yvon. Modèles probabilistes pour l'accès à l'information textuelle, Hermès, pp.223-267, 2011Chapitre d'ouvrage
                    
        
        
            
                
                    inria-00514525
                                            v1 | 
| L’apprentissage automatique : le diable n’est pas dans l’algorithme2015Autre publication scientifique
                    
        
        
            
                
                    hal-01246178
                                            v1 | |
| Sequential Tree-to-Word Transducers: Normalization, Minimization, and Learning2015Autre publication scientifique
                    
        
        
            
                
                    hal-01186993
                                            v1 | 
| Tau-Eval: A Unified Evaluation Framework for Useful and Private Text Anonymization2025Pré-publication, Document de travail
                    
        
        
            
                
                    hal-05300069
                                            v1 | |
|  | Federated Learning for MRI-based BrainAGE: a multicenter study on post-stroke functional outcome prediction2025Pré-publication, Document de travail
                    
        
        
            
                
                    hal-05244644
                                            v1 | 
|  | Probability Estimation over Large-Scale Random Networks via the Fiedler Delta Statistic2013Pré-publication, Document de travail
                    
        
        
            
                
                    hal-00922432
                                            v1 | 
|  | Structures arborescentes et apprentissage automatiqueAutre [cs.OH]. Université Charles de Gaulle - Lille III, 2006HDR
                    
        
        
            
                
                    tel-00117063
                                            v1 | 
                Chargement...
            
            Chargement...