The indigenous languages technology project at NRC: An empowerment approach to language
summary
Three‑year, $6M NRC project builds practical tooling for Indigenous language work rather than abstract research: a usable read-along studio that aligns audiobook text and audio and is already deployed in about five communities; a scalable WordWeaver verb conjugator for polysynthetic languages (Mohawk live in two immersion schools, being extended to Michif); a released Inuktut–English parallel corpus used to train MT; ASR preprocessing and speaker‑dependent transcription pipelines; most code and all community data ownership go back to communities. Funding renewal is pending.
AI-written summary · about
chapters (26)
- 0:00Introduction
- 0:22Challenges
- 9:13Language Families
- 9:57Statistics
- 11:25History of Indigenous Languages
- 13:23Team
- 14:23Structure
- 15:50Speech
- 16:17Readalong Studio
- 19:05Predictive Text
- 19:49Word Count
- 20:09Centrosome
- 20:28Wordweaver
- 21:13Verb wheels
- 22:05Wordweaver software
- 23:23Mohawk
- 24:51Office tools
- 25:17The gister
- 25:34The spell checker
- 26:02The transcription bottleneck
- 28:53Funding
- 29:18Audience Questions
- 35:07Community engagement
- 36:18Contact information
- 36:59Texttospeech
- 39:02Takeaways
related talks
21mThe indigenous languages technology (ILT) project at NRC in 2021
29mIndian Local Language Enabling: Are We Ready? Learn How to Get Started!
31mNLP for Preserving the Quechua Language
26mMachine Translation into Learning
26mResponsible technology under digital colonialism
45mTranscend Language Barriers with Custom Machine Learning APIs: um...
44mSome tentative first steps towards a Star Trek universal communicator
16mThe End of Awkward AI Transcriptions