FLEURS: FEW-Shot Learning Evaluation of Universal Representations of Speech
2022 IEEE Spoken Language Technology Workshop (SLT)2023pp. 798–805
Citations Over TimeTop 1% of 2023 papers
Alexis Conneau, Min Ma, Simran Khanuja, Yu Zhang, Vera Axelrod, Siddharth Dalmia, Jason Riesa, Clara E. Rivera, Ankur Bapna
Abstract
We introduce FLEURS, the Few-shot Learning Evaluation of Universal Representations of Speech benchmark. FLEURS is an n-way parallel speech dataset in 102 languages built on top of the machine translation FLoRes-101 benchmark, with approximately 12 hours of speech supervision per language. FLEURS can be used for a variety of speech tasks, including Automatic Speech Recognition (ASR), Speech Language Identification (Speech LangID), Speech-Text Retrieval. In this paper, we provide baselines for the tasks based on multilingual pre-trained models like speech-only w2v-BERT [1] and speech-text multimodal mSLAM [2]. The goal of FLEURS is to enable speech technology in more languages and catalyze research in low-resource speech understanding. 1 .
Related Papers
- Ogmios: The UPC Text-to-Speech synthesis system for Spoken Translation(2006)
- → Design issues in developing speech corpus for Indian languages — A survey(2012)9 cited
- → Digital Speech Technology(2010)1 cited
- → A novel intonation model to improve the quality of tamil text-to-speech synthesis system(2014)1 cited
- → A System Design of English Speech Synthesis(2021)