Last week, @BLIPLabNTU shared results from a huge study of child-directed speech at the Workshop on Infant Language Development. The study led by @suzyjstyles looks at how parents use all of their languages when narrating a story
#MultilingualIsNormal
In today’s poster we’re presenting how all 143 parents use their different languages while they are telling the wordless picture story. Here is the full dataset 😍
Some parents talk for only about 2 mins, others for 20. Here in SG Almost all parents are translanguaging 🎉
@BLIPLabNTU is back at @Infantstudies for #ICIS2024 with data from our Talk Together Study: 146 parents narrated a storybook to their children. Here, every row is one parental narration (8m/o at bottom, 40m/o at top). Something changes with child age, but what is it?
(P1-E-94)
Our Bilingual Futures - an info session for parents kicked off at 9am this morning 🎉
Talks by @suzyjstyles, Dr Beth O’Brian and @DrAnnabelChen, plus a preview screening of a short film by @yvonnecrepaldi. Our team is ready to chat with parents, and our open Q&A will begin soon!
Ever seen auto-generated subtitles turn to mush because they couldn’t handle a speaker’s accent or figure out what language they’re speaking after a switch?
Join the #MerlionChallenge at #Interspeech23 or RT to help us build more robust speech systems for multilingualism 💪🏼💪🏽💪🏿
It has taken @BLIPLabNTU YEARS, but it has been worth it - the labelled data are beautiful 🤩
👉All audio transcribed/checked by TWO multilingual coders
👉Precise timestamps for speaker and language
👉Every segment has a single language label (Code switching boundaries marked)
#Multilingual#LangDev + #SpeechProc friends!
We’ve prepped >30hrs of our English/Mandarin code-switched speech for the @ISCAInterspeech#merlionchallenge
>300 files, >100 voices 🙀
Can your speech system figure out which language is spoken when?
https://t.co/AicZHoEooR