|21 Apr 2023
|15:00 - 16:30
|Room S1, Alison Richard Building, 7 West Road, Cambridge
Rolando Coto-Solano (Dartmouth College)
Artificial intelligence techniques can help us accelerate our documentation work even in languages with very few resources. We will focus on breaking the transcription bottleneck by discussing how current cross-lingual speech recognition, bootstrapping off of large languages like English, can facilitate the task of transcription in the field. We will examine how techniques such as automated parsing and machine translation can help compile language corpora and also look at the impact that technologies can have in language revitalisation and how even simple tools like predictive keyboards can help in the creation of language resources. The talk will present two test cases, the languages Bribri from Costa Rica and Cook Islands Māori from Polynesia.
About the Speakers
- Coto-Solano is the Assistant Professor of Linguistics, Adjunct Assistant Professor of Computer Science and Affiliate Professor of Quantitative Social Science Program (QSS) at Dartmouth College. His interests are Natural Language Processing for Indigenous and Under-Resourced Languages, including Speech Recognition, Automated parsing, Language documentation, revitalisation and technology.
- Hugo Leal, Teaching Associate at CDH and Research Associate at the Minderoo Centre for Technology and Democracy, will host this event which is convened in association with the Automatic Speech Recognition (ASR) for Endangered Languages Project, for which a CDH Incubation award has been granted to Marieke Meelen, Theoretical & Applied Linguistics, MMLL Faculty.