The Chinese technology firm Choknor is enhancing its DeepZang large language model to improve support for Tibetan dialects and enable offline functionality. Unveiled at a recent seminar in Hohhot, these updates aim to refine real-time translation between Tibetan and Mandarin. Since its March debut, the platform has gained nearly 400,000 users across regions with significant Tibetan populations. The AI has already streamlined administrative tasks, medical consultations, and commercial documentation. Trained on extensive linguistic datasets covering major dialect groups, the model is being integrated into various sectors, including healthcare and local governance. Choknor plans to introduce offline language packages and test simultaneous interpretation hardware by the end of the year. The initiative seeks to bridge communication gaps and provide equitable access to digital tools for Tibetan speakers, while supporting regional integration and information accuracy through advanced speech and text processing technologies.
Source : [Sixth Tone](https://www.sixthtone.com/news/1019100/China’s DeepZang AI Aims to Improve Tibetan Translation)
Photo : Sixth Tone
