Tibetan Language and AI: A Comprehensive Survey of Resources, Methods and Challenges
Not on the map. Why?
Abstract (arXiv)
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.
Key facts. Cited by 1 paper in the map (data through 13 Oct 2026). Builds on 56 of its 172 references that are in the map. First posted on arXiv 22 Oct 2025.
Builds on
56 of 172 references.
References · 56
- 2015Deep Residual Learning for Image Recognitioncited by 51,950
- 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,588
- 2023GPT-4 Technical Reportcited by 18,278
- 2022Training language models to follow instructions with human feedbackcited by 15,856
- 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,322
- 2025Qwen3 Technical Reportcited by 8,510
- 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,986
- 2020Measuring Massive Multitask Language Understandingcited by 7,876
- 2023Visual Instruction Tuningcited by 7,827
- 2023BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Modelscited by 6,103
- 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,884
- 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,143
38 more
- 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,880
- 2023Mamba: Linear-Time Sequence Modeling with Selective State Spacescited by 4,658
- 2014Neural Machine Translation by Jointly Learning to Align and Translatecited by 4,539
- 2024GPT-4o System Cardcited by 4,536
- 2022Flamingo: a Visual Language Model for Few-Shot Learningcited by 4,363
- 2022BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generationcited by 4,314
- 2014Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translationcited by 4,024
- 2019Unsupervised Cross-lingual Representation Learning at Scalecited by 3,916
- 2021Finetuned Language Models Are Zero-Shot Learnerscited by 3,765
- 2023QLoRA: Efficient Finetuning of Quantized LLMscited by 3,388
- 2024Gemini 1.5: Unlocking multimodal understanding across millions of tokens of contextcited by 3,197
- 2014Sequence to Sequence Learning with Neural Networkscited by 2,981
- 2015Neural Machine Translation of Rare Words with Subword Unitscited by 2,910
- 2019XLNet: Generalized Autoregressive Pretraining for Language Understandingcited by 2,810
- 2019ALBERT: A Lite BERT for Self-supervised Learning of Language Representationscited by 2,779
- 2019ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Language Taskscited by 2,220
- 2024Qwen2 Technical Reportcited by 2,102
- 2018SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processingcited by 1,979
- 2022BLOOM: A 176B-Parameter Open-Access Multilingual Language Modelcited by 1,978
- 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,844
- 2020mT5: A massively multilingual pre-trained text-to-text transformercited by 1,815
- 2016WaveNet: A Generative Model for Raw Audiocited by 1,717
- 2016Enriching Word Vectors with Subword Informationcited by 1,581
- 2022No Language Left Behind: Scaling Human-Centered Machine Translationcited by 1,302
- 2019VisualBERT: A Simple and Performant Baseline for Vision and Languagecited by 1,294
- 2020HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesiscited by 1,148
- 2019PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarizationcited by 1,100
- 2020XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalizationcited by 741
- 2016Neural Architectures for Named Entity Recognitioncited by 549
- 2015Bidirectional LSTM-CRF Models for Sequence Taggingcited by 545
- 2015An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognitioncited by 520
- 2023CMMLU: Measuring massive multitask language understanding in Chinesecited by 439
- 2014Deep Speech: Scaling up end-to-end speech recognitioncited by 354
- 2023SafetyBench: Evaluating the Safety of Large Language Modelscited by 197
- 2025TLUE: A Tibetan Language Understanding Evaluation Benchmarkcited by 5
- 2022TiBERT: Tibetan Pre-trained Language Modelcited by 3
- 2025TFD: A Comprehensive Structured Tibetan Foundation Dataset for Low-Resource Language Processing and Large-Scale Modelingcited by 3
- 2025FMSD-TTS: Few-shot Multi-Speaker Multi-Dialect Text-to-Speech Synthesis for "U-Tsang, Amdo and Kham Speech Dataset Generationcited by 3
50 most cited of 56 All 56 (JSON)
Built on by
Citing papers per year: 2026: 1.
1 paper.