Exchanges for SPEech ReseArch aNd TechnOlogies

Risultati finali

Organise communication activities, workshops, courses and other knowledge share actions. Ensure that the dissemination is compliant with open access requirements.

Description of explainability for speech

This deliverable will provide a list and description of criteria for explainability in the context of speech processing

Scientific dissemination guidelines

Identify partners specificities and needs Writing guidelines to the attention of partners to ensure an active and coherent communication activity among countries

Project website and visual identity

Setting up a visual identity and a website presenting to the general public the objectives of the project, its partners and the main tools employed. The website will also be a resource for the partners for all that concerns good practices.

SDK for human assisted learning speech processing

Software library developed in open-source licence to tackle to issue of human assisted learning for various tasks of speech processing such as speaker diarization, speaker verification, speech translation of speech recognition.

Models that learn from small data

To deal with under-resource task, this deliverable will implement different approaches dealing with low or zero resource learning such as transfer learning and modeling based on the same information shared between languages or tasks, approaches that do not need annotated data, use of expert knowledge in empirical systems, systems using multimodal data for semantic supervision.

Pubblicazioni

Cross-Corpus Speech Emotion Recognition with HuBERT Self-Supervised Representation

Autori: Miguel Pastor; Dayana Ribas; Alfonso Ortega; Antonio Miguel; Eduardo Lleida
Pubblicato in: Proc. IberSPEECH, Numero 12, 2022
Editore: ISCA
DOI: 10.21437/iberspeech.2022-16

Multi-Channel Speaker Verification with Conv-Tasnet Based Beamformer

Autori: Mošner Ladislav, Plchot Oldřich, Burget Lukáš, Černocký Jan
Pubblicato in: Proceedings of ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2022, Pagina/e 7982-7986, ISBN 978-1-6654-0540-9
Editore: IEEE Signal Processing Society
DOI: 10.1109/icassp43922.2022.9747771

A transfer learning based approach for pronunciation scoring

Autori: Sancinetti, Marcelo; Vidal, Jazmin; Bonomi, Cyntia; Ferrer, Luciana
Pubblicato in: ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Numero 4, 2022, Pagina/e 6812-6816, ISBN 978-1-6654-0540-9
Editore: IEEE
DOI: 10.48550/arxiv.2111.00976

Microphone Array Channel Combination Algorithms for Overlapped Speech Detection

Autori: Theo Mariotte; Anthony Larcher; Silvio Montrésor; Jean-Hugh Thomas
Pubblicato in: Interspeech 2022 Human and Humanizing Speech Technology, Numero 12, 2022
Editore: ISCA
DOI: 10.21437/interspeech.2022-10758

Log-Likelihood-Ratio Cost Function as Objective Loss for Speaker Verification Systems

Autori: Victoria Mingote, Antonio Miguel, Alfonso Ortega, Eduardo Lleida
Pubblicato in: Interspeech 2021, 2021, Pagina/e 2361-2365
Editore: ISCA
DOI: 10.21437/interspeech.2021-1085

Extracting Speaker and Emotion Information from Self-Supervised Speech Models via Channel-Wise Correlations

Autori: Themos Stafylakis; Ladislav Mosner; Sofoklis Kakouros; Oldrich Plchot; Lukas Burget; Jan Cernocky
Pubblicato in: 2022 IEEE Spoken Language Technology Workshop (SLT), Numero 22, 2023, ISBN 979-8-3503-9690-4
Editore: IEEE
DOI: 10.1109/slt54892.2023.10023345

Learnable Sparse Filterbank for Speaker Verification

Autori: PENG Junyi, GU Rongzhi, MOŠNER Ladislav, PLCHOT Oldřich, BURGET Lukáš and ČERNOCKÝ Jan.
Pubblicato in: Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, 2022, Pagina/e 5110-5114
Editore: International Speech Communication Association (ISCA)
DOI: 10.21437/interspeech.2022-11309

On the Use of Semantically-Aligned Speech Representations for Spoken Language Understanding

Autori: Gaelle Laperriere; Valentin Pelloin; Mickael Rouvier; Themos Stafylakis; Yannick Esteve
Pubblicato in: 2022 IEEE Spoken Language Technology Workshop (SLT), Numero 24, 2023
Editore: IEEE
DOI: 10.1109/slt54892.2023.10023013

Multisv: Dataset for Far-Field Multi-Channel Speaker Verification

Autori: Mošner Ladislav, Plchot Oldřich, Burget Lukáš, Černocký Jan
Pubblicato in: Proceedings of ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2022, Pagina/e 7977-7981, ISBN 978-1-6654-0540-9
Editore: IEEE Signal Processing Society
DOI: 10.1109/icassp43922.2022.9746833

Probabilistic Spherical Discriminant Analysis: An Alternative to PLDA for length-normalized embeddings.

Autori: BRUMMER Johan Nikolaas Langenhoven, SWART Albert du Preez, MOŠNER Ladislav, SILNOVA Anna, PLCHOT Oldřich, STAFYLAKIS Themos and BURGET Lukáš.
Pubblicato in: Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH., 2022, Pagina/e 1446-1450
Editore: International Speech Communication Association (ISCA)
DOI: 10.21437/interspeech.2022-731

Speaker Embeddings for Diarization of Broadcast Data In The Allies Challenge

Autori: Anthony Larcher; Ambuj Mehrish; Marie Tahon; Sylvain Meignier; Jean Carrive; David Doukhan; Olivier Galibert; Nicholas Evans
Pubblicato in: ICASSP, Numero 1, 2021
Editore: IEEE
DOI: 10.1109/icassp39728.2021.9414215

Semantic Enrichment Towards Efficient Speech Representations

Autori: Gaëlle Laperrière; Ha Nguyen; Sahar Ghannay; Bassam Jabaian; Yannick Estève
Pubblicato in: Proc. INTERSPEECH 2023, 2023, Pagina/e 705-709, ISSN 1990-9772
Editore: ISCA
DOI: 10.21437/interspeech.2023-2234

Speech-Based Emotion Recognition with Self-Supervised Models Using Attentive Channel-Wise Correlations and Label Smoothing

Autori: Sofoklis Kakouros; Themos Stafylakis; Ladislav Mošner; Lukáš Burget
Pubblicato in: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Numero 19, 2023, Pagina/e pp. 1-5, ISBN 978-1-7281-6327-7
Editore: IEEE
DOI: 10.1109/icassp49357.2023.10094673

Multi-Channel Speech Separation with Cross-Attention and Beamforming

Autori: "Ladislav Mosner, Oldřich Plchot, Junyi Peng, Lukáš Burget, Jan ""Honza"" Černocký"
Pubblicato in: Proc. INTERSPEECH 2023, 2023, Pagina/e 693-1697, ISSN 1990-9772
Editore: International Speech Communication Association
DOI: 10.21437/interspeech.2023-2537

Speaker Embeddings by Modeling Channel-Wise Correlations

Autori: Themos Stafylakis, Johan Rohdin, Lukáš Burget
Pubblicato in: Interspeech 2021, 2021, Pagina/e 501-505
Editore: ISCA
DOI: 10.21437/interspeech.2021-1442

A Study on the Use of wav2vec Representations for Multiclass Audio Segmentation

Autori: Pablo Gimeno; Alfonso Ortega; Antonio Miguel; Eduardo Lleida
Pubblicato in: Proc. IberSPEECH2022, Numero 8, 2022, Pagina/e 56-60
Editore: ISCA
DOI: 10.21437/iberspeech.2022-12

End-to-End Speech Translation of Arabic to English Broadcast News

Autori: Fethi Bougares; Salim Jouili
Pubblicato in: WANLP@ACL 2022: Abu Dhabi, United Arab Emirates, Numero 5, 2022, Pagina/e 312–319, ISBN 978-1-959429-27-2
Editore: Association for Computational Linguistics
DOI: 10.18653/v1/2022.wanlp-1.29

Improving Speaker Verification with Self-Pretrained Transformer Models

Autori: "Junyi Peng, Oldřich Plchot, Themos Stafylakis, Ladislav Mosner, Lukáš Burget, Jan ""Honza"" Černocký"
Pubblicato in: Proc. INTERSPEECH 2023, 2023, Pagina/e 5361-5365, ISSN 1990-9772
Editore: International Speech Communication Association
DOI: 10.21437/interspeech.2023-453

Training Speaker Embedding Extractors Using Multi-Speaker Audio with Unknown Speaker Boundaries.

Autori: STAFYLAKIS Themos, MOŠNER Ladislav, PLCHOT Oldřich, ROHDIN Johan A., SILNOVA Anna, BURGET Lukáš and ČERNOCKÝ Jan.
Pubblicato in: Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, 2022, Pagina/e 605-609
Editore: International Speech Communication Association (ISCA)
DOI: 10.21437/interspeech.2022-10165

Improving Speaker Diarization for Low-Resourced Sarawak Malay Language Conversational Speech Corpus

Autori: Mohd Zulhafiz Rahim; Sarah Samson Juan; Fitri Suraya Mohamad
Pubblicato in: 2023 International Conference on Asian Language Processing (IALP), 2023, Pagina/e 228-233
Editore: IEEE
DOI: 10.1109/ialp61005.2023.10337314

Multi-Speaker and Wide-Band Simulated Conversations as Training Data for End-to-End Neural Diarization

Autori: Federico Landini; Mireia Diez; Alicia Lozano-Diez; Lukáš Burget
Pubblicato in: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Numero 6, 2023, Pagina/e 1-5, ISBN 978-1-7281-6327-7
Editore: IEEE
DOI: 10.1109/icassp49357.2023.10097049

Description and Analysis of ABC Submission to NIST LRE 2022

Autori: Pavel Matejka, Anna Silnova, Josef Slavíček, Ladislav Mosner, Oldřich Plchot, Michal Klčo, Junyi Peng, Themos Stafylakis, Lukáš Burget
Pubblicato in: Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, 2023, Pagina/e 511-515, ISSN 1990-9772
Editore: International Speech Communication Association
DOI: 10.21437/interspeech.2023-1529

An Attention-Based Backend Allowing Efficient Fine-Tuning of Transformer Models for Speaker Verification

Autori: Junyi Peng; Oldrich Plchot; Themos Stafylakis; Ladislav Mosner; Lukas Burget; Jan Cernocky
Pubblicato in: 2022 IEEE Spoken Language Technology Workshop (SLT), Numero 5, 2023, Pagina/e 555-562, ISBN 979-8-3503-9690-4
Editore: IEEE
DOI: 10.1109/slt54892.2023.10022775

Active Correction for Incremental Speaker Diarization of a Collection with Human in the Loop

Autori: Larcher, Yevhenii Prokopalo; Meysam Shamsi; Loïc Barrault; Sylvain Meignier; Anthony
Pubblicato in: Applied Sciences; Volume 12; Numero 4;, Numero 20, 2022, Pagina/e Pages: 1782, ISSN 2076-3417
Editore: MDPI
DOI: 10.3390/app12041782

Unsupervised Representation Learning for Speech Activity Detection in the Fearless Steps Challenge 2021

Autori: Pablo Gimeno, Alfonso Ortega, Antonio Miguel, Eduardo Lleida
Pubblicato in: Interspeech 2021, 2021, Pagina/e 4359-4363
Editore: ISCA
DOI: 10.21437/interspeech.2021-309

An Explainable Proxy Model for Multilabel Audio Segmentation

Autori: Mariotte, Théo; Almudévar, Antonio; Tahon, Marie; Ortega, Alfonso
Pubblicato in: International Conference on Acoustics Speech and Signal Processing, IEEE, Apr 2024, Seoul (Korea), Numero 5, 2024
Editore: IEEE Signal Processing Society
DOI: 10.1109/icassp48485.2024.10446648

Analyzing speaker verification embedding extractors and back-ends under language and channel mismatch

Autori: SILNOVA Anna, STAFYLAKIS Themos, MOŠNER Ladislav, PLCHOT Oldřich, ROHDIN Johan A., MATĚJKA Pavel, BURGET Lukáš, GLEMBEK Ondřej a BRUMMER Johan Nikolaas Langenhoven.
Pubblicato in: Proceedings of The Speaker and Language Recognition Workshop (Odyssey 2022), 2022, Pagina/e 9-16
Editore: International Speech Communication Association (ISCA)
DOI: 10.21437/odyssey.2022-2

Toroidal Probabilistic Spherical Discriminant Analysis

Autori: Anna Silnova; Niko Brümmer; Albert Swart; Lukáš Burget
Pubblicato in: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Numero 2, 2023, Pagina/e 1-5, ISBN 978-1-7281-6327-7
Editore: IEEE
DOI: 10.1109/icassp49357.2023.10095580

Sur la vérification du locuteur à partir de traces d’exécution de modèles acoustiques personnalisés

Autori: Tomashenko, Natalia; Mdhaffar, Salima; Tommasi, Marc; Estève, Yannick; Bonastre, Jean-François
Pubblicato in: Journées d'Études sur la Parole - JEP2022, Jun 2022, Île de Noirmoutier, France, Numero 13, 2022
Editore: ISCA

ON-TRAC Consortium Systems for the IWSLT 2023 Dialectal and Low-resource Speech Translation Tasks

Autori: Antoine Laurent, Souhir Gahbiche, Ha Nguyen, Haroun Elleuch, Fethi Bougares, Antoine Thiol, Hugo Riguidel, Salima Mdhaffar, Gaëlle Laperrière, Lucas Maison, Sameer Khurana, Yannick Estève
Pubblicato in: Proceedings of the 20th International Conference on Spoken Language Translation (IWSLT 2023), 2023, Pagina/e 219–226, ISBN 2-9517408-4-0
Editore: Association for Computational Linguistics
DOI: 10.18653/v1/2023.iwslt-1.18

Strategies for Improving Low Resource Speech to Text Translation Relying on Pre-trained ASR Models

Autori: Santosh Kesiraju; Marek Sarvaš; Tomáš Pavlíček; Cécile Macaire; Alejandro Ciuba
Pubblicato in: Proc. INTERSPEECH 2023, Numero 14, 2023, Pagina/e 2148--2152, ISSN 1990-9772
Editore: ISCA
DOI: 10.21437/interspeech.2023-2506

Discriminative Training of VBx Diarization

Autori: Klement, Dominik; Diez, Mireia; Landini, Federico; Burget, Lukáš; Silnova, Anna; Delcroix, Marc; Tawara, Naohiro
Pubblicato in: Crossref, Numero 5, 2024
Editore: IEEE
DOI: 10.1109/icassp48485.2024.10446119

BUT CHiME-7 system description

Autori: Karafiát, Martin; Veselý, Karel; Szöke, Igor; Mošner, Ladislav; Beneš, Karel; Witkowski, Marcin; Barchi, Germán; Pepino, Leonardo
Pubblicato in: CHiME-7 proceedings, Numero 8, 2023
Editore: arxiv
DOI: 10.48550/arxiv.2310.11921

Development of ABC systems for the 2021 edition of NIST Speaker Recognition Evaluation

Autori: ALAM Jahangir, BURGET Lukáš, GLEMBEK Ondřej, MATĚJKA Pavel, MOŠNER Ladislav, PLCHOT Oldřich, ROHDIN Johan A., SILNOVA Anna and STAFYLAKIS Themos et al.
Pubblicato in: Proceedings of The Speaker and Language Recognition Workshop (Odyssey 2022), 2022, Pagina/e 346-353
Editore: International Speech Communication Association (ISCA)
DOI: 10.21437/odyssey.2022-48

Parameter-Efficient Transfer Learning of Pre-Trained Transformer Models for Speaker Verification Using Adapters

Autori: Junyi Peng; Themos Stafylakis; Rongzhi Gu; Oldřich Plchot; Ladislav Mošner; Lukáš Burget; Jan Černocký
Pubblicato in: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Numero 6, 2023, Pagina/e 1-5, ISBN 978-1-7281-6327-7
Editore: IEEE
DOI: 10.1109/icassp49357.2023.10094795

Generalizing AUC Optimization to Multiclass Classification for Audio Segmentation With Limited Training Data

Autori: Pablo Gimeno; Victoria Mingote; Alfonso Ortega; Antonio Miguel; Eduardo Lleida
Pubblicato in: IEEE Signal Processing Letters, Numero 4, 2021, Pagina/e 1135 - 1139, ISSN 1070-9908
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/lsp.2021.3084501

YembaTones: a syllable-tone annotated dataset for speech recognition and prosodic analysis of the Yemba language.

Autori: Kenfack Jeuguim Marc Sturm; Paulin Melatagia Yonta; Sandembouo Etienne
Pubblicato in: Data in Brief, Numero 11, 2023, ISSN 2352-3409
Editore: Elsevier BV
DOI: 10.1016/j.dib.2023.109860

Towards Lifelong Human Assisted Speaker Diarization

Autori: Meysam Shamsi; Anthony Larcher; Loic Barrault; Sylvain Meignier; Yevheni Prokopalo; Marie Tahon; Ambuj Mehrish; Simon Petitrenaud; Olivier Galibert; Samuel Gaist; André Anjos; Sebastien Marcel; Marta R. Costa-jussà
Pubblicato in: Computer Speech & Language, Numero 4, 2023, ISSN 0885-2308
Editore: Academic Press
DOI: 10.1016/j.csl.2022.101437

aDCF Loss Function for Deep Metric Learning in End-to-End Text-Dependent Speaker Verification Systems

Autori: Victoria Mingote; Antonio Miguel; Dayana Ribas; Alfonso Ortega; Eduardo Lleida
Pubblicato in: IEEE/ACM Transactions on Audio, Speech, and Language Processing, Numero 25 January 2022, 2022, Pagina/e 772-784, ISSN 2329-9290
Editore: IEEE Advancing Technology for Humanity
DOI: 10.1109/taslp.2022.3145307

Direct Text to Speech Translation System Using Acoustic Units

Autori: Victoria Mingote, Pablo Gimeno, Luis Vicente, Sameer Khurana, Antoine Laurent and Jarod Duret
Pubblicato in: IEEE Signal Processing Letters, Numero vol. 30, 2023, Pagina/e 1262-1266, ISSN 1070-9908
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/lsp.2023.3313513

An Overview of the IberSpeech-RTVE 2022 Challenges on Speech Technologies

Autori: Eduardo Lleida; Luis Javier Rodriguez-Fuentes; Javier Tejedor; Alfonso Ortega; Antonio Miguel; Virginia Bazán; Carmen Pérez; Alberto de Prada; Mikel Penagarikano; Amparo Varona; Germán Bordel; Doroteo Torre-Toledano; Aitor Álvarez; Haritz Arzelus
Pubblicato in: Applied Sciences; Volume 13; Numero 15;, Numero 5, 2023, Pagina/e Pages: 8577, ISSN 2076-3417
Editore: MDPI
DOI: 10.3390/app13158577

Class token and knowledge distillation for multi-head self-attention speaker verification systems

Autori: Victoria Mingote; Antonio Miguel; Alfonso Ortega; Eduardo Lleida
Pubblicato in: Digital Signal Processing, Numero 6, 2023, ISSN 1051-2004
Editore: Academic Press
DOI: 10.1016/j.dsp.2022.103859

Multimodal Diarization Systems by Training Enrollment Models as Identity Representations

Autori: Victoria Mingote; Ignacio Viñals; Pablo Gimeno; Antonio Miguel; Alfonso Ortega; Eduardo Lleida
Pubblicato in: Applied Sciences, Numero 12-3, 2022, Pagina/e 1141, ISSN 2076-3417
Editore: MDPI
DOI: 10.3390/app12031141

Wiener Filter and Deep Neural Networks: A Well-Balanced Pair for Speech Enhancement

Autori: Dayana Ribas; Antonio Miguel; Alfonso Ortega; Eduardo Lleida
Pubblicato in: Applied Sciences, Numero 3 ; Volume 12; Numero 18;, 2022, Pagina/e Pages: 9000, ISSN 2076-3417
Editore: MDPI
DOI: 10.3390/app12189000

Automatic Voice Disorder Detection Using Self-Supervised Representations

Autori: Dayana Ribas; Miguel A. Pastor; Antonio Miguel; David Martinez; Alfonso Ortega; Eduardo Lleida
Pubblicato in: IEEE Access, Numero 6, 2023, Pagina/e 14915-14927,, ISSN 2169-3536
Editore: Institute of Electrical and Electronics Engineers Inc.
DOI: 10.1109/access.2023.3243986

Cross-Corpus Training Strategy for Speech Emotion Recognition Using Self-Supervised Representations

Autori: Miguel A. Pastor; Dayana Ribas; Alfonso Ortega; Antonio Miguel; Eduardo Lleida
Pubblicato in: Applied Sciences, Numero 13, 2023, Pagina/e 9062, ISSN 2076-3417
Editore: MDPI
DOI: 10.3390/app13169062

Unsupervised Adaptation of Deep Speech Activity Detection Models to Unseen Domains

Autori: Pablo Gimeno; Dayana Ribas; Alfonso Ortega; Antonio Miguel; Eduardo Lleida
Pubblicato in: Applied Sciences, Numero 3, 2022, ISSN 2076-3417
Editore: MDPI
DOI: 10.3390/app12041832

The Domain Mismatch Problem in the Broadcast Speaker Attribution Task

Autori: Ignacio Viñals, Alfonso Ortega, Antonio Miguel, Eduardo Lleida
Pubblicato in: Applied Sciences, Numero 11/18, 2021, Pagina/e 8521, ISSN 2076-3417
Editore: MDPI
DOI: 10.3390/app11188521

Risultati finali

Pubblicazioni

Condividi questa pagina

Scarica