2023-07-29, 09:30–10:00 (Asia/Taipei), TR 509
Language: English
Lingua Libre by Wikimedia France is an open source webapp to create large and clean audio corpora, best suited for e-dictionaries and text2speech machine learning.
Outline
Demo of the tool (10mins)
Stated objectives
Current progresses
Limits & biases
The diversity of world languages, their words, expressions, voices, are poorly documented and accessible. Lingualibre.org allows us to record languages vocabulary and audio dictionaries at large scale, in an easy and quick fashion (800 audio/hour) .
After 5 year of action and 800,000+ recordings, we would like to share past progresses and discuss future actions.
Beginner
Target Audience –語言,詞彙,詞典,台語,方言,dictionary,
E-learning professional, Wikimedian in Residence at Université de Toulouse, Hugo has been promoting free online education via Wikipedia, MOOCs and Lingualibre for two decades. His strongest action and expertise is related to languages contents, as a part of the Wikimedia Languages Diversity Hub.