Subventions et des contributions :
Subvention ou bourse octroyée s'appliquant à plus d'un exercice financier (2017-2018 à 2018-2019).
Our partner Fluent.AI develops Automatic Speech Recognition (ASR) systems based on neural networks. Thex000D
performance of ASR systems degrades significantly in the presence of noise. The types of noises that arex000D
relevant occur e.g. on urban sidewalks, construction sites, factory floors or in cars (e.g. radio music orx000D
background conversations). The performance degradation of ASR systems is partly caused by a lack ofx000D
robustness to noise of the data representations used in ASR. To improve the performance of our partner's ASRx000D
systems, we propose to develop a pre-processing method that reduces noises in the speech signal before it isx000D
sent to the ASR engine. Our method is based on independent component analysis.