A Speech Test Set of Practice Business Presentations with Additional Relevant Texts

  Autoři

Macháček, Dominik ; Kratochvíl, Jonáš ; Vojtěchová, Tereza and Bojar, Ondřej

  Identifikátor

http://hdl.handle.net/11234/1-3023

 Odkaz na záznam v

https://doi.org/10.1007/978-3-030-31372-2_13

 Datum vydání

2019-07-13

 Typ

audio, corpus

 Velikost

59 minutes, 39 entries

 Jazyky

English

 Popis

We present a test corpus of audio recordings and transcriptions of presentations of students' enterprises together with their slides and web-pages. The corpus is intended for evaluation of automatic speech recognition (ASR) systems, especially in conditions where the prior availability of in-domain vocabulary and named entities is benefitable. The corpus consists of 39 presentations in English, each up to 90 seconds long, and slides and web-pages in Czech, Slovak, English, German, Romanian, Italian or Spanish. The speakers are high school students from European countries with English as their second language. We benchmark three baseline ASR systems on the corpus and show their imperfection.

 Nakladatel

Charles University, Faculty of Mathematics and Physics, Institute of Formal and Applied Linguistics (UFAL)

 Sponzoři

European Union

Kód projektu: H2020-ICT-2018-2-825460

Jméno projektu: ELITR - European Live Translator

 

Czech Science Foundation

Kód projektu: 19-26934X

Jméno projektu: Neural Representations in Multi-modal and Multi-lingual Modelling

 Klíčová slova

ASR ASR evaluation speech corpus non-native English speech recognition speech recognition evaluation speech and relevant texts European non-native English

 Kolekce

LINDAT / CLARIAH-CZ Data & Tools

Zobrazit celý záznam