Szczegóły publikacji
Opis bibliograficzny
BUT CHiME-7 system description / Martin Karafiát, Karel Veselý, Igor Szőke, Ladislav Mošner, Karel Beneš, Marcin WITKOWSKI, Germán Barchi, Leonardo Pepino // W: ChiME 2023 [Dokument elektroniczny] : 7th international workshop on Speech processing in everyday environments : Dublin, Ireland, 25 August 2023. — Wersja do Windows. — Dane tekstowe. — [Germany] : ISCA Archive, [2023]. — S. 67–72. — Wymagania systemowe: Adobe Reader. — Tryb dostępu: https://www.isca-archive.org/chime_2023/karafiat23_chime.pdf [2024-10-04]. — Bibliogr. s. 71–72, Abstr.
Autorzy (8)
- Karafiát Martin
- Veselý Karel
- Szőke Igor
- Mošner Ladislav
- Beneš Karel
- AGHWitkowski Marcin
- Barchi Germán
- Pepino Leonardo
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 155775 |
|---|---|
| Data dodania do BaDAP | 2024-10-29 |
| Rok publikacji | 2023 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
Abstract
This paper describes the joint effort of Brno University of Technology (BUT), AGH University of Krakow and University of Buenos Aires on the development of Automatic Speech Recognition systems for the CHiME-7 Challenge. We train and evaluate various end-to-end models with several toolkits. We heavily relied on Guided Source Separation (GSS) to convert multi-channel audio to single channel. The ASR is leveraging speech representations from models pre-trained by selfsupervised learning, and we do a fusion of several ASR systems. In addition, we modified external data from the LibriSpeech corpus to become a close domain and added it to the training. Our efforts were focused on the far-field acoustic robustness sub-track of Task 1 - Distant Automatic Speech Recognition (DASR), our systems use oracle segmentation
