Abstract:

There are 16 ethnic groups and 42 languages in Taiwan's indigenous peoples. However, these languages are severely under-resourced and lack standardized writing systems. This study first built the first complete corpus of ethnic languages (ethnic speech, ethnic Chinese characters, Chinese speech, and Chinese characters) in Taiwan, and developed the "Indigenous language education artificial intelligence technology platform". The platform focused on artificial intelligence natural language processing. The core AI module was introduced into the interactive technology platform of the AI test and self-learning service system of the Chinese language literacy ability, which actively meets and serves the needs of the indigenous people in language, education, learning, and certification, and helps the sustainable preservation and development of Taiwan's indigenous languages. Furthermore, the interactive technology platform of the AI self-learning service system was used to optimize and promote the bi-lingual translation and literacy ability of ethnic Chinese, Chinese, and English to achieve the following research objectives: 1. Establishment of a Chinese language corpus with historical value; 2. Development of an AI model for bi-directional translation of Chinese and Chinese; 3. Development of an AI interactive platform for indigenous language education, 4. Development of an AI adaptive test for indigenous language literacy based on an NLP database, 5. Building a self-learning service system for Indigenous language literacy based on the NLP language database, 6. Building a multi-language translation platform based on NLP database for educational application promotion, 7. In order to integrate Taiwan's indigenous languages with the rest of the world, the AI model should be used for two-way translation and analysis in ethnic Chinese, Chinese, and English, and the development and application of technology should be guided by social inclusion, preserving indigenous culture, and promoting the diffusion of research work models.

Keywords:

Adaptive test, artificial intelligence, Indigenous languages, natural language processing, self-learning system.