ZHAW-InIT at GermEval 2020 task 4 : low-resource speech-to-text

Büchi, Matthias; Ulasik, Malgorzata Anna; Hürlimann, Manuela; Benites de Azevedo e Souza, Fernando; von Däniken, Pius; Cieliebak, Mark

doi:10.21256/zhaw-21550

Please use this identifier to cite or link to this item: https://doi.org/10.21256/zhaw-21550

Publication type:	Conference paper
Type of review:	Peer review (abstract)
Title:	ZHAW-InIT at GermEval 2020 task 4 : low-resource speech-to-text
Authors:	Büchi, Matthias Ulasik, Malgorzata Anna Hürlimann, Manuela Benites de Azevedo e Souza, Fernando von Däniken, Pius Cieliebak, Mark
et. al:	No
DOI:	10.21256/zhaw-21550
Proceedings:	Proceedings of the 5th Swiss Text Analytics Conference (SwissText) & 16th Conference on Natural Language Processing (KONVENS)
Editors of the parent work:	Ebling, Sarah Tuggener, Don Hürlimann, Manuela Cieliebak, Mark Volk, Martin
Conference details:	5th SwissText & 16th KONVENS Joint Conference, Zurich (online), 24-25 June 2020
Issue Date:	Jun-2020
Publisher / Ed. Institution:	CEUR Workshop Proceedings
ISSN:	1613-0073
Language:	English
Subject (DDC):	410.285: Computational linguistics
Abstract:	This paper presents the contribution of ZHAW-InIT to Task 4 ”Low-Resource STT” at GermEval 2020. The goal of the task is to develop a system for translating Swiss German dialect speech into Standard German text in the domain of parliamentary debates. Our approach is based on Jasper, a CNN Acoustic Model, which we fine-tune on the task data. We enhance the base system with an extended Language Model containing in-domain data and speed perturbation and run further experiments with post-processing. Our submission achieved first place with a final Word Error Rate of 40.29%.
URI:	https://digitalcollection.zhaw.ch/handle/11475/21550
Fulltext version:	Published version
License (according to publishing contract):	CC BY 4.0: Attribution 4.0 International
Departement:	School of Engineering
Organisational Unit:	Institute of Computer Science (InIT)
Appears in collections:	Publikationen School of Engineering

Files in This Item:

File	Description	Size	Format
2020_Buechi_etal_ZHAW-InIT-at-GermEval-2020.pdf		458.44 kB	Adobe PDF	View/Open

Show full item record

Büchi, M., Ulasik, M. A., Hürlimann, M., Benites de Azevedo e Souza, F., von Däniken, P., & Cieliebak, M. (2020). ZHAW-InIT at GermEval 2020 task 4 : low-resource speech-to-text [Conference paper]. In S. Ebling, D. Tuggener, M. Hürlimann, M. Cieliebak, & M. Volk (Eds.), Proceedings of the 5th Swiss Text Analytics Conference (SwissText) & 16th Conference on Natural Language Processing (KONVENS). CEUR Workshop Proceedings. https://doi.org/10.21256/zhaw-21550

Büchi, M. et al. (2020) ‘ZHAW-InIT at GermEval 2020 task 4 : low-resource speech-to-text’, in S. Ebling et al. (eds) Proceedings of the 5th Swiss Text Analytics Conference (SwissText) & 16th Conference on Natural Language Processing (KONVENS). CEUR Workshop Proceedings. Available at: https://doi.org/10.21256/zhaw-21550.

M. Büchi, M. A. Ulasik, M. Hürlimann, F. Benites de Azevedo e Souza, P. von Däniken, and M. Cieliebak, “ZHAW-InIT at GermEval 2020 task 4 : low-resource speech-to-text,” in Proceedings of the 5th Swiss Text Analytics Conference (SwissText) & 16th Conference on Natural Language Processing (KONVENS), Jun. 2020. doi: 10.21256/zhaw-21550.

BÜCHI, Matthias, Malgorzata Anna ULASIK, Manuela HÜRLIMANN, Fernando BENITES DE AZEVEDO E SOUZA, Pius VON DÄNIKEN und Mark CIELIEBAK, 2020. ZHAW-InIT at GermEval 2020 task 4 : low-resource speech-to-text. In: Sarah EBLING, Don TUGGENER, Manuela HÜRLIMANN, Mark CIELIEBAK und Martin VOLK (Hrsg.), Proceedings of the 5th Swiss Text Analytics Conference (SwissText) & 16th Conference on Natural Language Processing (KONVENS). Conference paper. CEUR Workshop Proceedings. Juni 2020

Büchi, Matthias, Malgorzata Anna Ulasik, Manuela Hürlimann, Fernando Benites de Azevedo e Souza, Pius von Däniken, and Mark Cieliebak. 2020. “ZHAW-InIT at GermEval 2020 Task 4 : Low-Resource Speech-to-Text.” Conference paper. In Proceedings of the 5th Swiss Text Analytics Conference (SwissText) & 16th Conference on Natural Language Processing (KONVENS), edited by Sarah Ebling, Don Tuggener, Manuela Hürlimann, Mark Cieliebak, and Martin Volk. CEUR Workshop Proceedings. https://doi.org/10.21256/zhaw-21550.

Büchi, Matthias, et al. “ZHAW-InIT at GermEval 2020 Task 4 : Low-Resource Speech-to-Text.” Proceedings of the 5th Swiss Text Analytics Conference (SwissText) & 16th Conference on Natural Language Processing (KONVENS), edited by Sarah Ebling et al., CEUR Workshop Proceedings, 2020, https://doi.org/10.21256/zhaw-21550.