Review stimuli

[cesine]

In an initial analysis of the stimuli http://batumi.github.io/labphon2018/ we found some of them were not grammatical and not producible by participants

• Review web search stimuli https://github.com/batumi/KartuliSpeechRecognition/blob/master/data/web_search_corpus.txt
• Review legal search stimuli https://github.com/batumi/KartuliSpeechRecognition/blob/master/data/legal_search_corpus.txt

Other corpora

• https://targm.net/tag/%E1%83%95%E1%83%A3%E1%83%9A%E1%83%92%E1%83%90%E1%83%A0%E1%83%A3%E1%83%9A%E1%83%98/
• https://www.ted.com/talks?language=ka (from https://targm.net/2013/01/12/ted-com/)
• http://corpora.iliauni.edu.ge/

How to avoid suggesting training vulgar words in sms?

• Movie subtitles