Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Tibetan Language and AI: A Comprehensive Survey of Resources, Methods and Challenges

Cheng Huang, Nyima Tashi, Fan Gao, Yutong Liu, Jiahao Li, Hao Tian, Siyang Jiang, Thupten Tsering, Ban Ma-bao, Renzeg Duojie, Gadeng Luosang, Rinchen Dongrub, Dorje Tashi, Jin Zhang, Xiao Feng, Hao Wang, Jie Tang, Guojie Tang, Xiangxiang Wang, Jia Zhang, Tsengdar Lee, Yongbin Yu

TL;DR

The survey addresses the challenge of advancing AI for Tibetan, a representative low-resource language, by systematically cataloging textual and speech resources, datasets, benchmarks, and existing models. It argues for a shift from fragmented, task-specific approaches to unified, instruction-tuned, Tibetan-centric LLMs, while highlighting data scarcity, orthographic variation, and evaluation gaps as core bottlenecks. The authors propose a data- and benchmark-centric roadmap, advocate community-driven resource creation, and explore hardware-efficient avenues such as memristor-based neuromorphic computing to enable edge deployment. Collectively, the paper provides a foundational reference for building an inclusive Tibetan AI ecosystem, with potential lessons applicable to other low-resource languages in multilingual AI research.

Abstract

Tibetan, one of the major low-resource languages in Asia, presents unique linguistic and sociocultural characteristics that pose both challenges and opportunities for AI research. Despite increasing interest in developing AI systems for underrepresented languages, Tibetan has received limited attention due to a lack of accessible data resources, standardized benchmarks, and dedicated tools. This paper provides a comprehensive survey of the current state of Tibetan AI in the AI domain, covering textual and speech data resources, NLP tasks, machine translation, speech recognition, and recent developments in LLMs. We systematically categorize existing datasets and tools, evaluate methods used across different tasks, and compare performance where possible. We also identify persistent bottlenecks such as data sparsity, orthographic variation, and the lack of unified evaluation metrics. Additionally, we discuss the potential of cross-lingual transfer, multi-modal learning, and community-driven resource creation. This survey aims to serve as a foundational reference for future work on Tibetan AI research and encourages collaborative efforts to build an inclusive and sustainable AI ecosystem for low-resource languages.

Tibetan Language and AI: A Comprehensive Survey of Resources, Methods and Challenges

TL;DR

The survey addresses the challenge of advancing AI for Tibetan, a representative low-resource language, by systematically cataloging textual and speech resources, datasets, benchmarks, and existing models. It argues for a shift from fragmented, task-specific approaches to unified, instruction-tuned, Tibetan-centric LLMs, while highlighting data scarcity, orthographic variation, and evaluation gaps as core bottlenecks. The authors propose a data- and benchmark-centric roadmap, advocate community-driven resource creation, and explore hardware-efficient avenues such as memristor-based neuromorphic computing to enable edge deployment. Collectively, the paper provides a foundational reference for building an inclusive Tibetan AI ecosystem, with potential lessons applicable to other low-resource languages in multilingual AI research.

Abstract

Tibetan, one of the major low-resource languages in Asia, presents unique linguistic and sociocultural characteristics that pose both challenges and opportunities for AI research. Despite increasing interest in developing AI systems for underrepresented languages, Tibetan has received limited attention due to a lack of accessible data resources, standardized benchmarks, and dedicated tools. This paper provides a comprehensive survey of the current state of Tibetan AI in the AI domain, covering textual and speech data resources, NLP tasks, machine translation, speech recognition, and recent developments in LLMs. We systematically categorize existing datasets and tools, evaluate methods used across different tasks, and compare performance where possible. We also identify persistent bottlenecks such as data sparsity, orthographic variation, and the lack of unified evaluation metrics. Additionally, we discuss the potential of cross-lingual transfer, multi-modal learning, and community-driven resource creation. This survey aims to serve as a foundational reference for future work on Tibetan AI research and encourages collaborative efforts to build an inclusive and sustainable AI ecosystem for low-resource languages.
Paper Structure (84 sections, 5 figures, 3 tables)

This paper contains 84 sections, 5 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Tibetan Resource
  3. Public Corpus
  4. Annotated Dataset
  5. Benchmark
  6. Specialized Benchmark
  7. Comprehensive Multilingual Benchmark
  8. Others
  9. Task Taxonomy of Tibetan AI
  10. Fundamental Task
  11. Data Collection
  12. Annotation Standardization
  13. Discriminative Task
  14. Low-level Lexical Processing
  15. Information Extraction
  16. ...and 69 more sections

Figures (5)

  • Figure 1: Overview of the Practical Guide for current AI Trend on Tibetan
  • Figure 3: Task Taxonomy of Tibetan AI
  • Figure 4: Future Directions of AI in Tibetan in terms of both Data and Models
  • Figure 5: Overview of the TLUE Benchmark
  • Figure 6: Statistical Category of the Ti-MMLU Benchmark