Mostafa Sadeghi
Présentation
I am a researcher with the Multispeech team at Inria, Nancy - Grand Est, France. My current research focuses on robust audio-visual speech processing (enhancement & separation), with an emphasis on combining the expressive power of deep neural networks with statistical and probabilistic machine learning techniques, e.g., using deep generative models. I received my PhD from Sharif University of Technology, Tehran, Iran, in April 2018. From August 2018 to October 2020, I was a postdoctoral researcher with the Perception team at Inria, Grenoble, France, working with Radu Horaud and Xavier Alameda-Pineda. Prior to that, I was a visiting PhD scholar at the Information Science and Engineering Department, Royal Institute of Technology (KTH), Stockholm, Sweden, and later, a research engineer at the Automatic Control Department.
Publications
Publications
|
|
Diffusion-based Unsupervised Audio-visual Speech EnhancementICASSP 2025 - International Conference on Acoustics Speech and Signal Processing, IEEE, Apr 2025, Hyderabad, India. pp.1-5 |
|
|
End-to-end Joint Punctuated and Normalized ASR with a Limited Amount of Punctuated Training DataEuropean Signal Processing Conference (EUSIPCO 2025), Sep 2025, Palermo, Italy |
|
|
Data-independent Beamforming for End-to-end Multichannel Multi-speaker ASRIEEE 27th International Workshop on Multimedia Signal Processing (MMSP 2025), IEEE, Sep 2025, Pékin, China |
|
|
COLaF : Corpus et Outils pour les Langues de France et variétés de français20e Conférence en Recherche d’Information et Applications (CORIA) 32ème Conférence sur le Traitement Automatique des Langues Naturelles (TALN) 27ème Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues (RECITAL) Les 18e Rencontres Jeunes Chercheurs en RI (RJCRI), Jun 2025, Marseille, France. pp.33-47 |
|
|
Towards Skeletal and Signer Noise Reduction in Sign Language Production via Quaternion-Based Pose Encoding and Contrastive LearningSLTAT 2025: 9th Workshop on Sign Language Translation and Avatar Technologies, Sep 2025, Berlin, Germany. pp.1-9, ⟨10.1145/3742886.3756728⟩ |
|
|
Joint Beamforming and Speaker-Attributed ASR for Real Distant-Microphone Meeting TranscriptionEuropean Signal Processing Conference (EUSIPCO 2025), Sep 2025, Palermo, Italy |
|
|
A weighted-variance variational autoencoder model for speech enhancementICASSP 2024 - International Conference on Acoustics Speech and Signal Processing, IEEE, Apr 2024, Seoul (Korea), South Korea. pp.1-5, ⟨10.1109/ICASSP48485.2024.10446294⟩ |
|
|
Unsupervised speech enhancement with diffusion-based generative modelsInternational Conference on Acoustics Speech and Signal Processing (ICASSP), IEEE, Apr 2024, Seoul (Korea), South Korea. ⟨10.48550/arXiv.2309.10450⟩ |
|
|
Posterior sampling algorithms for unsupervised speech enhancement with recurrent variational autoencoderInternational Conference on Acoustics Speech and Signal Processing (ICASSP), IEEE, Apr 2024, Seoul (Korea), South Korea. ⟨10.48550/arXiv.2309.10439⟩ |
|
|
Improving Speaker Assignment in Speaker-Attributed ASR for Real Meeting ApplicationsThe Speaker and Language Recognition Workshop Odyssey 2024, Jun 2024, Quebec, Canada |
|
|
Diffusion-based speech enhancement with a weighted generative-supervised learning lossInternational Conference on Acoustics Speech and Signal Processing (ICASSP), IEEE, Apr 2024, Seoul (Korea), South Korea. ⟨10.48550/arXiv.2309.10457⟩ |
|
|
Fast and efficient speech enhancement with variational autoencodersInternational Conference on Acoustics Speech and Signal Processing (ICASSP), IEEE, Jun 2023, Rhodes island, Greece |
|
|
End-to-end Multichannel Speaker-Attributed ASR: Speaker Guided Decoder and Input Feature Analysis2023 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU 2023), Dec 2023, Taipei, Taiwan. ⟨10.1109/ASRU57964.2023.10389729⟩ |
|
|
The CHiME-7 UDASE task: Unsupervised domain adaptation for conversational speech enhancement7th International Workshop on Speech Processing in Everyday Environments (CHiME), Aug 2023, Dublin, Ireland. ⟨10.21437/CHiME.2023-2⟩ |
|
|
Audio-visual speech enhancement with a deep kalman filter generative modelInternational Conference on Acoustics Speech and Signal Processing (ICASSP), IEEE, Jun 2023, Rhodes island, Greece |
|
|
A Sparsity-promoting Dictionary Model for Variational AutoencodersINTERSPEECH 2022, Sep 2022, Incheon, South Korea |
|
|
The Impact of Removing Head Movements on Audio-visual Speech EnhancementICASSP 2022 - IEEE International Conference on Acoustics, Speech and Signal Processing, IEEE Signal Processing Society, May 2022, Singapore, Singapore. pp.1-5, ⟨10.1109/ICASSP43922.2022.9746401⟩ |
|
|
Switching Variational Auto-Encoders for Noise-Agnostic Audio-visual Speech EnhancementICASSP 2021 - 46th International Conference on Acoustics, Speech, and Signal Processing, Jun 2021, Toronto / Virtual, Canada. pp.1-5, ⟨10.1109/ICASSP39728.2021.9414097⟩ |
|
|
Robust Face Frontalization For Visual Speech RecognitionICCVW 2021 - International Conference on Computer Vision Workshops, IEEE, Oct 2021, Montreal - Virtual, Canada. pp.2485-2495, ⟨10.1109/ICCVW54120.2021.00281⟩ |
|
|
Deep Variational Generative Models for Audio-visual Speech SeparationMLSP 2021 - IEEE International Workshop on Machine Learning for Signal Processing, Oct 2021, Gold Coast, Australia. pp.1-6, ⟨10.1109/MLSP52302.2021.9596406⟩ |
|
|
Robust Unsupervised Audio-visual Speech Enhancement Using a Mixture of Variational AutoencodersIEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), May 2020, Barcelona, Spain. pp.7534-7538, ⟨10.1109/ICASSP40776.2020.9053730⟩ |
|
|
Low Mutual and Average Coherence Dictionary Learning Using Convex ApproximationICASSP 2020 - IEEE International Conference on Acoustics, Speech and Signal Processing, IEEE, May 2020, Barcelone (virtual), Spain. pp.3417-3421, ⟨10.1109/ICASSP40776.2020.9052901⟩ |
|
|
Accelerated Dictionary Learning for Sparse Signal RepresentationLVA/ICA 2017 - 13th International Conference on Latent Variable Analysis and Signal Separation, Olivier Michel; Nadège Thirion-Moreau, Feb 2017, Grenoble, France. pp.531 - 541, ⟨10.1007/978-3-319-53547-0_50⟩ |
|
|
Learning overcomplete dictionaries based on parallel atom-updatingMLSP 2013 - IEEE 23rd International Workshop on Machine Learning for Signal Processing, Sep 2013, Southampton, United Kingdom. 5 p |
|
|
Sequential subspace finding: a new algorithm for learning low-dimensional linear subspacesEUSIPCO 2013 - 21th European Signal Processing Conference, Sep 2013, Marrakech, Morocco. pp.EUSIPCO 2013 1569746207 |
|
|
A new algorithm for learning overcomplete dictionariesEUSIPCO 2013 - 21th European Signal Processing Conference, Sep 2013, Marrakech, Morocco. pp.EUSIPCO 2013 1569746047 |
|
|
End-to-end Multichannel Speaker-Attributed ASR: Speaker Guided Decoder and Input Feature AnalysisRencontre des Jeunes Chercheurs en Parole 2023 - 10E Edition, Nov 2023, Grenoble, France |
|
|
Face Frontalization Based on Robustly Fitting a Deformable Shape Model to 3D Landmarks2020 |
|
|
Unsupervised Performance Analysis of 3D Face Alignment2020 |