Joseph Dureau


2024

pdf
Sonos Voice Control Bias Assessment Dataset: A Methodology for Demographic Bias Assessment in Voice Assistants
Chloe Sekkat | Fanny Leroy | Salima Mdhaffar | Blake Perry Smith | Yannick Estève | Joseph Dureau | Alice Coucke
Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

Recent works demonstrate that voice assistants do not perform equally well for everyone, but research on demographic robustness of speech technologies is still scarce. This is mainly due to the rarity of large datasets with controlled demographic tags. This paper introduces the Sonos Voice Control Bias Assessment Dataset, an open dataset composed of voice assistant requests for North American English in the music domain (1,038 speakers, 166 hours, 170k audio samples, with 9,040 unique labelled transcripts) with a controlled demographic diversity (gender, age, dialectal region and ethnicity). We also release a statistical demographic bias assessment methodology, at the univariate and multivariate levels, tailored to this specific use case and leveraging spoken language understanding metrics rather than transcription accuracy, which we believe is a better proxy for user experience. To demonstrate the capabilities of this dataset and statistical method to detect demographic bias, we consider a pair of state-of-the-art Automatic Speech Recognition and Spoken Language Understanding models. Results show statistically significant differences in performance across age, dialectal region and ethnicity. Multivariate tests are crucial to shed light on mixed effects between dialectal region, gender and age.

2016

pdf
Identification de lieux dans les messageries mobiles (Place extraction from smartphone messaging applications)
Clément Doumouro | Adrien Ball | Joseph Dureau | Sylvain Raybaud | Ramzi Ben Yahya
Actes de la conférence conjointe JEP-TALN-RECITAL 2016. volume 5 : Démonstrations

Nous présentons un système d’identification de lieux dans les messageries typiquement utilisées sur smartphone. L’implémentation sur mobile et son cortège de contraintes, ainsi que la faible quantité de ressources disponibles pour le type de langage utilisé rendent la tâche particulièrement délicate. Ce système, implémenté sur Android, atteint une précision de 30% et un rappel de 72%.