Progress in 2024
We are supporting basic research into Ainu-Japanese translation functions. In a series of studies by Professor Emeritus Araki Kenji of Hokkaido University and others, the accuracy of translation was improved by using vector transformation models. However, it has not yet reached a practical level (Tanaka S, Echizen’ ya H, and Araki K. 2024).
On the other hand, there has been progress on AI Pirika’s Ainu-to-Ainu chat function. We have created an original AI chat prototype using GPT (Generative Pre-trained Transformer). We plan to start brushing up the function with young Ainu people studying Ainu language.
Reference: Tanaka S, Echizen’ ya H, and Araki K. Neural machine translation for low-resource languages using vector transformation models based on similarity between vectors. IPSJ SIG Technical Report, Vol. 2024-SLP-154, 20, p1-6, 2024-12-05, ISSN,2188-8663.
Background
You have probably heard of the extinction of animal species. The human ethnic group has a fate to disappear if they lose their language.
Using The Summer Institute of Linguistics Ethnologue data, Nettle and Daniel can calculate that 90 percent of the world’s population speaks the 100 most-used languages. At least 6,000 languages are spoken by about 10 percent of the people on earth[1].
Languages with under 100 speakers are so close to extinction that revival for everyday use seems unlikely[1]. In 1996, only 15 people could speak the Ainu language as native speakers [2]. In 2020, the situation is hopeless. We are very concerned about this situation.
Our main project is to revive a native speaker of Ainu using AI (Artificial Intelligence) technologies. Her name is “AI Pirika.” “Pirika” means a pretty Ainu girl.
Someone may be worried that AI Pirika cannot have her emotions. However, we will incorporate emotions into the system.
We are looking forward to applying for endangered languages in the world.
[1] Nettle and Daniel. 2000. Vanishing Voices. Oxford University Press.
[2] Moseley, Christopher (ed.). 2010. Atlas of the World’s Languages in Danger, 3rd edn. Paris, UNESCO Publishing. Online version: http://www.unesco.org/culture/en/endangeredlanguages/atlas
[3] Kenji Araki and Michitomo Kuroda. Generality of spoken dialogue system using SEGA-IL for different languages. 2006. Proceedings of the IASTED International Conference COMPUTATIONAL INTELLIGENCE. November 20-22, 2006, San Francisco, CA, USA.