Towards the Roof of the World: NLP Progress in Tibetan Languages

# Towards the Roof of the World: NLP Progress inTibetan Languages
Source: [https://arxiv.org/html/2610.13398](https://arxiv.org/html/2610.13398)
rmTeXGyreTermesX \[tibetan\]rmNotoSerifTibetan\.ttf

## ![[Uncaptioned image]](https://arxiv.org/html/2610.13398v1/logo.png)Towards theRoof of the World: NLP Progress in Tibetan LanguagesThanks:🖂 Corresponding Author

Xi Cao,Yijie Li,Yuan Sun,Affiliation:School of Philosophy and Religious Studies, Minzu University of ChinaAffiliation:School of Information Engineering, Minzu University of ChinaAffiliation:Institute of National Security, Minzu University of ChinaAffiliation:National Language Resource Monitoring & Research Center — Minority Languages BranchShiwei Dong,Zeguang Huang,Nuo QunAffiliation:School of Information Science and Technology, Xizang UniversityAffiliation:The State Key Laboratory of Tibetan Intelligencecaoxi@muc\.edu\.cn,sunyuan@muc\.edu\.cn

###### Abstract

Tibetan, spoken by some seven million people on and around the Qinghai\-Xizang Plateau, remains underrepresented in natural language processing \(NLP\)\. We present, to our knowledge, the first peer\-reviewed systematic bibliometric survey of Tibetan NLP: 140 publications from 2004 to 2026 in the ACL Anthology and DBLP, annotated for venue type, research area, first\-affiliation country, and resource release\. Output has grown rapidly, with 84% of the publications appearing since 2018, and has expanded bottom\-up along the NLP pipeline, from corpora and word segmentation through script, speech, and applications to large language models \(LLMs\)\. Researchers based in China first\-authored 87% of the publications, which span all 25 represented research areas, whereas work from other countries concentrates on Classical Tibetan\. We argue that this full\-pipeline paradigm, realized under the compounded constraints of a complex script, dialectal diversity, and geographic remoteness, offers a reference model for other low\-resource languages of the Global South\. We also identify gaps in dialect coverage, responsible NLP, native and comparable evaluation, cross\-community visibility, and resource release \(only 29% of the publications release data, code, or models\), and recommend open science, shared standards, and community participation for the LLM era\.

## 1Introduction

The benefits of NLP remain unevenly distributed: most of the world’s languages have little or no labelled data\([Joshi et al\., 2020](https://arxiv.org/html/2610.13398#bib.bib150)\), and LLMs, whose capabilities depend on each language’s digital text, threaten to widen this gap\. The risk is compounded in remote highlands, where geographic isolation, dialectal fragmentation, and limited digital infrastructure add to data scarcity; the Qinghai\-Xizang Plateau, the world’s highest, known as theRoof of the Worldand theThird Pole, is an extreme instance\.

Tibetan, the principal language of the plateau, is spoken by some seven million people in the Xizang Autonomous Region and the provinces of Qinghai, Sichuan, Gansu, and Yunnan in China, and by communities in Pakistan, India, Nepal, and Bhutan\. Its script does not mark word boundaries, its spoken varieties differ substantially, and its digital text is scarce \(§[2\.1](https://arxiv.org/html/2610.13398#S2.SS1)\); in the taxonomy of[Joshi et al\. \(2020\)](https://arxiv.org/html/2610.13398#bib.bib150), it belongs to Class 1 \(“The Scraping\-Bys”\), with some unlabelled but almost no labelled data, a class that, the authors argue, needs “a solid, organized movement” to collect labelled data\. Tibetan thus tests whether such a movement can bring a low\-resource language into the LLM era\.

Although Tibetan NLP has been studied for more than two decades, its literature is fragmented across communities and venues: 103 of the 140 publications we identify \(74%\) lie outside the ACL Anthology, the sole source of a recent survey of Sino\-Tibetan languages\([Liu and Best, 2025](https://arxiv.org/html/2610.13398#bib.bib141)\), and the closest Tibetan\-specific overview is a brief editorial\([Long and Hill, 2021](https://arxiv.org/html/2610.13398#bib.bib142)\)\. We address this gap with a bibliometric survey of these publications \(§[3](https://arxiv.org/html/2610.13398#S3)\), guided by four research questions:

#### RQ1\.

How has Tibetan NLP output evolved over time and across venues? \(§[4\.1](https://arxiv.org/html/2610.13398#S4.SS1)\)

#### RQ2\.

Which research areas have received the most and least attention, and how have priorities shifted? \(§[4\.2](https://arxiv.org/html/2610.13398#S4.SS2)\)

#### RQ3\.

Which countries have contributed, and what role has China played? \(§[4\.3](https://arxiv.org/html/2610.13398#S4.SS3)\)

#### RQ4\.

What are the key open challenges, and what could accelerate progress? \(§§[5](https://arxiv.org/html/2610.13398#S5)–[6](https://arxiv.org/html/2610.13398#S6)\)

Our central finding is that Tibetan NLP exemplifies what we call afull\-pipeline paradigm: starting from annotated corpora, lexicography, and word segmentation, the field has successively added script and speech processing, applications, and, since 2022, pre\-trained and large language models, while continuing to build resources\. We argue that this paradigm offers a reference model for other low\-resource languages of the Global South, identify the gaps to be closed in the LLM era, and release our annotated bibliography with an audit of publicly released resources\.

## 2Background

### 2\.1The Tibetan Language

Tibetan belongs to the Tibeto\-Burman branch of the Sino\-Tibetan language family\. Written continuously since the seventh century, its literary language, rooted in Classical Tibetan, serves as a shared written standard across three main dialect groups, Ü\-Tsang, Kham, and Amdo, whose spoken forms differ so much in phonology that mutual intelligibility is limited\([Schmidt, 2020](https://arxiv.org/html/2610.13398#bib.bib118);[Ren et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib72)\); even lexical tone varies, from tonal Ü\-Tsang through Kham to largely atonal Amdo\([Liang and Zerong, 2026](https://arxiv.org/html/2610.13398#bib.bib8)\)\. We useTibetancollectively for the classical and modern written language and its spoken varieties, often classified as distinct Tibetic languages \(hence the plural in our title\)\. This diglossia challenges NLP: models trained on written text may transfer poorly to speech, and variety\-specific systems generalize poorly\.

As Figure[1](https://arxiv.org/html/2610.13398#S2.F1)shows, Tibetan script is an abugida in which 30 consonant letters and 4 vowel signs\([Liu et al\., 2014](https://arxiv.org/html/2610.13398#bib.bib30)\)combine under strict orthographic rules to form syllables\. Syllables are delimited by thetsheg, but word boundaries are not marked, so word segmentation has long come first in text processing\([Liu et al\., 2011](https://arxiv.org/html/2610.13398#bib.bib33)\), although neural models can operate directly on syllables\([Lai et al\., 2018](https://arxiv.org/html/2610.13398#bib.bib28)\)\. Even syllable boundaries can be implicit: several case particles fuse with the preceding syllable and resemble ordinary final letters, a major source of errors in segmentation, syllable analysis, and subword machine translation \(MT\)\([Li and Yu, 2013](https://arxiv.org/html/2610.13398#bib.bib64);[Zhu et al\., 2019](https://arxiv.org/html/2610.13398#bib.bib52);[Gyatso et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib89)\)\. Encoding raises further problems: unlike Unicode, Chinese national extension standards encode whole stacks, 7,198 in all, as single characters\([Liu et al\., 2014](https://arxiv.org/html/2610.13398#bib.bib30);[Wang et al\., 2019](https://arxiv.org/html/2610.13398#bib.bib133)\); stacks may be stored in variant code\-point orders\([Faggionato and Meelen, 2019](https://arxiv.org/html/2610.13398#bib.bib27)\); and replacing thetshegwith spaces degrades fine\-tuned models\([Cao et al\., 2023](https://arxiv.org/html/2610.13398#bib.bib20)\)\. Tibetan is verb\-final with ergative marking\([An and Long, 2021](https://arxiv.org/html/2610.13398#bib.bib108)\), case\-particle forms depend on the preceding syllable’s final letter\([Khysru et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib117)\), and verbs have up to four stems\([Khysru et al\., 2019](https://arxiv.org/html/2610.13398#bib.bib126)\); its conservative orthography retains prefixed, superscribed, and subscribed letters and leaves Lhasa tones unwritten\([Yang et al\., 2015](https://arxiv.org/html/2610.13398#bib.bib139)\), both of which grapheme\-to\-phoneme conversion must model\. Together with scarce digital text, each property has spawned a sustained line of research \(§[4\.2](https://arxiv.org/html/2610.13398#S4.SS2)\)\.

Figure 1:The Tibetan word བསྒྲིགས་བཟོ་ \(programming\), illustrating the three\-level structure of Tibetan script: letters \(30 consonants and 4 vowel signs, each encoded by a Unicode code point, shown at the bottom\) combine into syllables according to strict orthographic rules, and syllables, each followed by the inter\-syllable marktsheg\(་\), form words whose boundaries are not marked\.
### 2\.2Related Surveys

Surveys of geographically defined, under\-resourced language groups, such as African languages\([Adebara and Abdul\-Mageed, 2022](https://arxiv.org/html/2610.13398#bib.bib144);[Alabi et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib143)\), the indigenous languages of Latin America\([Tonja et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib145)\), and the languages of India\([Kumar et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib146)\), inform our design\. Our survey differs from the two closest overviews of Tibetan NLP in scope, coverage of the LLM era, sources, and annotation \(Table[1](https://arxiv.org/html/2610.13398#S2.T1)\)\.

FeatureOur SurveyNAACL Survey\([Liu and Best, 2025](https://arxiv.org/html/2610.13398#bib.bib141)\)TALLIP Editorial\([Long and Hill, 2021](https://arxiv.org/html/2610.13398#bib.bib142)\)ScopeTibetan onlySino\-Tibetan family \(49 low\-resource languages\)Tibetan onlyCoverage Period2004–2026: from the earliest identified paper through 00:00 UTC, September 25, 20261952–2022: the ACL OCL corpus\([Rohatgi et al\., 2023](https://arxiv.org/html/2610.13398#bib.bib149)\); venue trends for 2000–20212016–2021: editorial for a special issue ofACM TALLIPCorpus140 publications, annotated for venue type, research area, country, and resource release253 papers \(274 language–paper matches for 23 of the 49 languages\), annotated for topicNo fixed count: narrative overview of key developmentsSourcesHybrid: ACL Anthology BibTeX and DBLP Search APIACL Anthology only, via the ACL OCL corpusSelected by the special\-issue editorsTable 1:Comparison of our survey with the two most closely related overviews of Tibetan NLP\.

## 3Methodology

#### Scope and inclusion criteria\.

We include peer\-reviewed publications in which Tibetan, modern or classical, is the primary language of study or a principal experimental object, and whose contribution concerns computational processing of Tibetan text, speech, or script; multilingual studies qualify when Tibetan is among their principal experimental languages\. Coverage runs from 2004, the year of the earliest Tibetan NLP paper in our sources, through 00:00 UTC, September 25, 2026\.

Figure 2:Two\-step construction of the survey corpus\.Step 1 \(automated retrieval\):entries withTibetanin the title or abstract are retrieved from the full ACL Anthology BibTeX file, and publications matching the keywordTibetanare retrieved through the DBLP Search API, retaining journal articles verified by Web of Science and conference papers from venues recommended in the 7th edition of the CCF catalogue\.Step 2 \(manual screening\):the merged and deduplicated candidate pool is reviewed for relevance, yielding the final set of 140 publications\.
#### Retrieval and screening\.

We proceeded in two stages \(Figure[2](https://arxiv.org/html/2610.13398#S3.F2)\)\. First, we retrieved all entries with the keywordTibetanin the title or abstract from the ACL Anthology’s full BibTeX export111[https://aclanthology\.org/faq/bib/](https://aclanthology.org/faq/bib/)\. To cover research outside it, we queried the DBLP Search API222[https://dblp\.org/faq/13501473\.html](https://dblp.org/faq/13501473.html)with the same keyword and, to restrict results to established peer\-reviewed venues, retained only Web of Science333[https://www\.webofscience\.com](https://www.webofscience.com/)indexed journal articles and papers from international conferences recommended in the 7th edition of the China Computer Federation \(CCF\) catalogue444[https://www\.ccf\.org\.cn/Academic\_Evaluation/By\_category/](https://www.ccf.org.cn/Academic_Evaluation/By_category/)\. Second, four researchers independently screened the deduplicated candidates against the inclusion criteria, considering only English\-language papers and resolving disagreements by discussion\. The final corpus has 140 publications: 37 from the ACL Anthology \(25 conference and 12 workshop papers; Appendix[A](https://arxiv.org/html/2610.13398#A1)\) and 103 from DBLP \(29 conference papers, 74 journal articles; Appendix[B](https://arxiv.org/html/2610.13398#A2)\)\.

#### Annotation\.

Each publication was annotated along three dimensions: \(1\)venue type\(conference, workshop, or journal\); \(2\)countryof the first author’s primary affiliation \(Appendix[C](https://arxiv.org/html/2610.13398#A3)\); and \(3\)research area: up to two of the 29 areas of the ACL Rolling Review \(ARR\) taxonomy555[https://aclrollingreview\.org/areas](https://aclrollingreview.org/areas)\(Appendix[D](https://arxiv.org/html/2610.13398#A4)\)\. In practice, 115 publications received two areas and 25 received one, yielding 255 area assignments, so area counts sum to more than 140\. We also recorded whether the authors release code, data, or models at a public URL \(Appendix[F](https://arxiv.org/html/2610.13398#A6)\)\.

## 4Results

### 4\.1Publication Trends by Year and Venue

Annual output \(Figure[3](https://arxiv.org/html/2610.13398#S4.F3)\) was sporadic for nearly two decades, with three papers in 2004–2006, none in 2007–2009, and 19 in 2010–2017, before rising sharply: 118 publications \(84\.3%\) date from 2018 onward, and 61 \(43\.6%\) from 2024–2026 alone, although 2026 is only partially covered\. This growth coincides with the wider field’s adoption of deep learning and, later, of pre\-trained Transformers and LLMs\. Several annual peaks, however, reflect clusters in single venues: five of the nine papers of 2018 appeared at PRCV 2018, seven of the 13 of 2019 inIEEE Access, eight of the ten of 2021 in TALLIP, six of them in the special issue introduced by[Long and Hill \(2021\)](https://arxiv.org/html/2610.13398#bib.bib142), and five of the 21 of 2026 in one issue ofData Intelligence\. Dedicated venues have thus shaped growth, and yearly fluctuations should not be over\-interpreted\.

Figure 3:Annual number of Tibetan NLP publications by venue type \(left; stacked bars\) and overall share of each venue type \(right; counts in the legend\), 2004–2026\. The year 2026 is only partially covered\. The full\-scale version is Figure[7](https://arxiv.org/html/2610.13398#A7.F7)in Appendix[E](https://arxiv.org/html/2610.13398#A5)\.Journal articles form the largest share of the corpus \(74; 52\.9%\), followed by conference papers \(54; 38\.6%\) and workshop papers \(12; 8\.6%\), dispersed across 10 workshops, 22 conferences, and 25 journals, led by TALLIP \(19 articles\) andIEEE Access\(14 articles\) \(Figure[4](https://arxiv.org/html/2610.13398#S4.F4)\)\. Only 37 publications \(26\.4%\) appear in the ACL Anthology, but 17 of these date from 2024–2026, including papers at ACL, EMNLP, AACL, COLING, and LREC: Tibetan NLP is entering the core NLP venues, although most of its literature circulates elsewhere\.

Figure 4:Number of publications per conference \(top left\), workshop \(top right\), and journal \(bottom\)\. Papers at joint conferences \(ACL\-IJCNLP 2015, IJCNLP\-AACL 2025\) are counted under each constituent venue, so the conference counts sum to 57 for 54 papers\. The full\-scale version is Figure[8](https://arxiv.org/html/2610.13398#A7.F8)in Appendix[E](https://arxiv.org/html/2610.13398#A5)\.
### 4\.2Research Areas

The 255 area assignments \(Figures[5](https://arxiv.org/html/2610.13398#S4.F5), left, and[6](https://arxiv.org/html/2610.13398#S4.F6)\) cover 25 of the 29 ARR areas but are highly skewed: Multilingualism and Cross\-Lingual NLP \(41 publications\) and Resources and Evaluation \(40\) account for 31\.8% of all assignments and the five largest areas for 56\.1%, whereas 11 areas have between one and four publications and four \(Dialogue and Interactive Systems; Ethics, Bias, and Fairness; LLM Agents; and NLP and Code Models\) have none\. We discuss them in four groups corresponding broadly to layers of the NLP pipeline\.

Figure 5:Number of publications per ARR research area \(left; descending order\) and per country of the first author’s affiliation \(right; ascending order\)\. As a publication can belong to more than one area, area counts sum to 255\. The full\-scale version is Figure[9](https://arxiv.org/html/2610.13398#A7.F9)in Appendix[E](https://arxiv.org/html/2610.13398#A5)\.![Refer to caption](https://arxiv.org/html/2610.13398v1/fig_heatmap.png)Figure 6:Number of publications per ARR research area and year, 2004–2026; darker cells indicate more publications\. Areas are ordered by total count \(descending\)\. The full\-scale version is Figure[10](https://arxiv.org/html/2610.13398#A7.F10)in Appendix[E](https://arxiv.org/html/2610.13398#A5)\.#### Foundations: resources, segmentation, and syntax\.

Resources and Evaluation is the most persistent area, active in 16 of the 20 years with any Tibetan NLP publication, from a syntactically annotated corpus and lexicographic resources\([Wagner and Zeisler, 2004](https://arxiv.org/html/2610.13398#bib.bib37);[Jiang et al\., 2005](https://arxiv.org/html/2610.13398#bib.bib36);[Jiang, 2006](https://arxiv.org/html/2610.13398#bib.bib35)\)to the training and evaluation data of the LLM era\([Zhang et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib15);[Zhang et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib13);[Gao et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib12);[Zhuang et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib87);[Zhuang and Sun, 2025](https://arxiv.org/html/2610.13398#bib.bib14);[Li et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib1)\)\. Corpus scale has grown from 35M syllables of news text\([Liu et al\., 2012](https://arxiv.org/html/2610.13398#bib.bib32)\)to the 166M\-token diachronic PACTib\([Meelen and Roux, 2020](https://arxiv.org/html/2610.13398#bib.bib25)\)and 72 GB for LLM pre\-training\([Yang et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib2)\), and 17 of its 40 publications appeared in 2025–2026\. Phonology, Morphology, and Word Segmentation \(17\) centres on the absence of word boundaries\. For lack of a public segmented corpus, early work trained a conditional random field segmenter on machine\-segmented text\([Liu et al\., 2011](https://arxiv.org/html/2610.13398#bib.bib33)\), and Chinese national standards for Tibetan word segmentation and part\-of\-speech tags followed only in 2018\([Zhu et al\., 2019](https://arxiv.org/html/2610.13398#bib.bib52);[Li et al\., 2022b](https://arxiv.org/html/2610.13398#bib.bib104)\)\. The area has since moved from syllable tagging\([Liu et al\., 2011](https://arxiv.org/html/2610.13398#bib.bib33);[Li and Yu, 2013](https://arxiv.org/html/2610.13398#bib.bib64)\)and unknown\-word identification\([Nuo et al\., 2015](https://arxiv.org/html/2610.13398#bib.bib29)\)to morphology\-aware language models\([Khysru et al\., 2019](https://arxiv.org/html/2610.13398#bib.bib126)\), subword methods for MT\([Li et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib110);[Gyatso et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib89)\), joint segmentation and part\-of\-speech tagging\([Li et al\., 2022b](https://arxiv.org/html/2610.13398#bib.bib104)\), neural segmentation\([Brown, 2024](https://arxiv.org/html/2610.13398#bib.bib17)\), and segmentation evaluation\([Zhuoma and Zhijie, 2026](https://arxiv.org/html/2610.13398#bib.bib74)\)\. Syntax: Tagging, Chunking and Parsing \(14\) follows the opposite trajectory: anchored in treebanks of Old and Classical Tibetan\([Faggionato and Meelen, 2019](https://arxiv.org/html/2610.13398#bib.bib27);[Meelen and Roux, 2020](https://arxiv.org/html/2610.13398#bib.bib25)\)and in dependency parsing\([An and Long, 2021](https://arxiv.org/html/2610.13398#bib.bib108);[An, 2023b](https://arxiv.org/html/2610.13398#bib.bib18)\), it has no publications after 2023, mirroring its estimated decline across the ACL Anthology from 20% to 2% of papers in 2000–2021\([Rohatgi et al\., 2023](https://arxiv.org/html/2610.13398#bib.bib149)\)\.

#### Script and speech\.

Nearly a third of the corpus \(45 publications; 32\.1%\) processes visual or acoustic input\. Multimodality and Language Grounding \(24\) is dominated by script recognition: online and offline handwriting\([Ma et al\., 2011](https://arxiv.org/html/2610.13398#bib.bib66);[Ma and Wu, 2015](https://arxiv.org/html/2610.13398#bib.bib60);[Cai and Wang, 2018](https://arxiv.org/html/2610.13398#bib.bib53);[Wang et al\., 2019](https://arxiv.org/html/2610.13398#bib.bib133);[Li et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib48);[Tashi et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib95);[Zhou et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib76)\), historical documents in woodblock and manuscript styles\([Wang et al\., 2018b](https://arxiv.org/html/2610.13398#bib.bib56);[Zhao et al\., 2018](https://arxiv.org/html/2610.13398#bib.bib57);[Wang et al\., 2018a](https://arxiv.org/html/2610.13398#bib.bib55);[Li et al\., 2022c](https://arxiv.org/html/2610.13398#bib.bib106);[Zhao et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib98)\), scene text\([Ke et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib83);[Lhundrub et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib77)\), and, recently, vision–language and multimodal optical character recognition \(OCR\) resources\([Xu et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib4);[Jie et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib71)\)\. Speech Recognition, Text\-to\-Speech and Spoken Language Understanding \(21\) has progressed from isolated\-word recognition with at most 20 speakers\([Zhao et al\., 2012](https://arxiv.org/html/2610.13398#bib.bib65)\)to end\-to\-end recognition for the Lhasa and Amdo varieties and in multi\-dialect settings\([Zhao et al\., 2019b](https://arxiv.org/html/2610.13398#bib.bib132);[Qin et al\., 2022b](https://arxiv.org/html/2610.13398#bib.bib50);[Liu et al\., 2023](https://arxiv.org/html/2610.13398#bib.bib49);[Zhu and Huang, 2020](https://arxiv.org/html/2610.13398#bib.bib121);[Liu et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib96)\), alongside speech synthesis\([Yang et al\., 2015](https://arxiv.org/html/2610.13398#bib.bib139);[Zhao et al\., 2019a](https://arxiv.org/html/2610.13398#bib.bib131);[Xu et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib116);[Zhang et al\., 2026b](https://arxiv.org/html/2610.13398#bib.bib68)\), speech translation\([Fan et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib70);[Ren et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib72)\), and word\-level evaluation\([Moore et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib75)\)\. A recurring strategy exploits the compositional structure of the script: decomposing stacks into components reduces 562 handwritten character classes to 120 components\([Ma and Wu, 2013](https://arxiv.org/html/2610.13398#bib.bib62);[Ma and Wu, 2015](https://arxiv.org/html/2610.13398#bib.bib60)\)and aids few\-shot printed\-text recognition\([Han et al\., 2018](https://arxiv.org/html/2610.13398#bib.bib54)\), and sub\-syllable radicals serve as units for language modelling and speech recognition\([Shen et al\., 2017](https://arxiv.org/html/2610.13398#bib.bib58);[Qin et al\., 2022b](https://arxiv.org/html/2610.13398#bib.bib50);[Qin et al\., 2022a](https://arxiv.org/html/2610.13398#bib.bib105)\)\. Both areas’ prominence plausibly reflects a large manuscript and print heritage and spoken varieties poorly served by written resources\.

#### Cross\-lingual transfer and language modelling\.

Multilingualism and Cross\-Lingual NLP \(41\) is the largest area and has gained the most publications since 2022 \(27\)\. Early work uses the Tibetan–Chinese pair to project base noun phrases without a Tibetan tagger or treebank\([Nuo et al\., 2012](https://arxiv.org/html/2610.13398#bib.bib31)\), to gloss keywords in OCR output in Chinese in lieu of unreliable MT\([Su et al\., 2013](https://arxiv.org/html/2610.13398#bib.bib63)\), and for bilingual clustering and cross\-lingual speech synthesis\([Ye et al\., 2015](https://arxiv.org/html/2610.13398#bib.bib140);[Yang et al\., 2015](https://arxiv.org/html/2610.13398#bib.bib139)\), while Sanskrit–Tibetan resources serve the classical tradition\([Nehrdich, 2022](https://arxiv.org/html/2610.13398#bib.bib22)\)\. Since 2024, the area has been shaped by multilingual corpora and benchmarks for China’s minority languages\([Zhang et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib15);[Zhang et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib13);[Li et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib1)\)and by adapting multilingual LLMs\([Gurgurov et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib16);[Yang et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib2)\)\. Language Modeling \(16; 13 since 2022\) recapitulates the wider field’s arc, from recurrent models exploiting radicals and morphology\([Shen et al\., 2017](https://arxiv.org/html/2610.13398#bib.bib58);[Khysru et al\., 2019](https://arxiv.org/html/2610.13398#bib.bib126)\), through BERT\-style models\([Liu et al\., 2022](https://arxiv.org/html/2610.13398#bib.bib51);[Liang et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib99);[Zhang et al\., 2025c](https://arxiv.org/html/2610.13398#bib.bib85)\), to LLaMA\-based LLMs\([Lv et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib86);[Sha et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib80)\), alignment datasets\([Zhuang et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib87);[Chen et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib69)\), and studies of vocabulary size and continual pre\-training\([Han et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib5);[Yang et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib2)\)\. Machine Learning for NLP \(11; nine since 2024\) and Efficient Methods for NLP \(3; all since 2023\) reflect a turn to few\-shot, prompt\-based, and parameter\-efficient adaptation\([Bao et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib97);[Tang et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib81);[Zhou et al\., 2024b](https://arxiv.org/html/2610.13398#bib.bib43)\)\.

#### Applications\.

Computational Social Science and Cultural Analytics \(12\) and Information Retrieval and Text Mining \(6\) emerged with mining Tibetan web and microblog text in 2014–2019\([Ye et al\., 2014](https://arxiv.org/html/2610.13398#bib.bib61);[Qiu et al\., 2017](https://arxiv.org/html/2610.13398#bib.bib138);[Xu et al\., 2018](https://arxiv.org/html/2610.13398#bib.bib136)\), and the former has returned with offensive\-language and rumour detection\([Ren et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib46);[Yu et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib93)\)\. Eight of the nine publications in NLP Applications concern historical or classical materials, such as interlinearized ancient texts\([Long and An, 2023](https://arxiv.org/html/2610.13398#bib.bib19)\)and OCR spelling correction\([Luo and Chuang, 2024](https://arxiv.org/html/2610.13398#bib.bib91)\)\. Machine Translation \(7\), has moved from syllable\- and subword\-based neural MT\([Lai et al\., 2018](https://arxiv.org/html/2610.13398#bib.bib28);[Li et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib110)\)to filtering noisy parallel data with a cross\-lingual model\([Zhou et al\., 2024a](https://arxiv.org/html/2610.13398#bib.bib45)\), speech translation from Amdo or all three dialect groups\([Ren et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib72);[Fan et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib70)\), and comparisons of dedicated MT engines with a prompted LLM\([Wang et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib79)\)\. Question Answering \(5\), Generation \(5\), Information Extraction \(4\), and Summarization \(3\) remain small\([Dan and Sun, 2024](https://arxiv.org/html/2610.13398#bib.bib94);[Xu et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib11);[Zhang et al\., 2025b](https://arxiv.org/html/2610.13398#bib.bib84), e\.g\.,\); all summarization studies use news titles, leads, or LLM\-generated summaries as references\([Ouyang et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib42);[Hao et al\., 2024b](https://arxiv.org/html/2610.13398#bib.bib47);[Wang et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib39)\)\.

#### Persistent gaps\.

Besides the four empty areas, topics central to the responsible deployment of language technology are nearly absent\. Interpretability and Analysis of Models \(3\) is represented only by adversarial\-robustness studies from a single research team\([Cao et al\., 2023](https://arxiv.org/html/2610.13398#bib.bib20);[Cao et al\., 2025b](https://arxiv.org/html/2610.13398#bib.bib10);[Cao et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib38)\), Human\-Centered NLP \(2\) by a readability tool and human\-in\-the\-loop adversarial\-text generation\([Schmidt, 2020](https://arxiv.org/html/2610.13398#bib.bib118);[Cao et al\., 2025b](https://arxiv.org/html/2610.13398#bib.bib10)\), and Clinical and Biomedical Applications \(1\) by medical named\-entity recognition, where F1 falls to 71% from 94% on news\([Zhang et al\., 2025b](https://arxiv.org/html/2610.13398#bib.bib84)\); Discourse and Pragmatics, Linguistic Theories, and Symbolic Reasoning have one each\.

#### A layered trajectory\.

Figure[6](https://arxiv.org/html/2610.13398#S4.F6)and Table[2](https://arxiv.org/html/2610.13398#S4.T2)show that the field has expanded broadly bottom\-up along the NLP pipeline, adding new layers rather than replacing old ones: resources and syntax appeared in 2004; segmentation, script, speech, and cross\-lingual work in 2010–2012; web mining in 2014; language modelling, machine translation, question answering, generation, and information extraction in 2017–2019; and efficient methods, interpretability, summarization, clinical applications, and discourse only in 2023–2025\. Earlier layers persisted: resources remain the second most frequent area in the LLM era, and segmentation research continues\. Only syntax has declined, as across the ACL Anthology, where resources have also waned but dialogue, interpretability, and ethics have grown\([Rohatgi et al\., 2023](https://arxiv.org/html/2610.13398#bib.bib149)\), the very areas in which Tibetan NLP lags\.

PhasePublicationsAreasChina2004–2017221395\.5%Top areas:Resources \(7\), Phon\./Morph\./Seg\. \(5\), Multilingual \(5\), Syntax \(4\), Multimodality \(4\)2018–2021401882\.5%Top areas:Resources \(9\), Multilingual \(9\), Multimodality \(7\), Speech \(7\)2022–2026782387\.2%Top areas:Multilingual \(27\), Resources \(24\), Language Modeling \(13\), Multimodality \(13\)Table 2:Three phases of Tibetan NLP\.Areas: ARR areas represented;China: share first\-authored in China;Top areas: four most frequent areas, with ties\. Boundaries: the 2018 output rise and the first BERT\-style model, TiBERT\([Liu et al\., 2022](https://arxiv.org/html/2610.13398#bib.bib51)\), in 2022\.

### 4\.3Country\-Level Analysis

Research is highly concentrated \(Figure[5](https://arxiv.org/html/2610.13398#S4.F5), right\): 122 publications \(87\.1%\) are first\-authored in China \(hereafter, China\-affiliated publications\), and the remaining 18 in the USA \(5\), the UK \(4\), Germany \(3\), Israel \(3\), and Australia, France, and India \(one each\)\. China’s share has exceeded 80% in every phase \(Table[2](https://arxiv.org/html/2610.13398#S4.T2)\), consistent with its lead in research teams and public funding for Tibetan information processing\([Long and Hill, 2021](https://arxiv.org/html/2610.13398#bib.bib142)\)\.

China’s contribution is broad as well as large: China\-affiliated publications cover all 25 represented areas, 15 of them exclusively, including Language Modeling, Multimodality, Machine Translation, Question Answering, and Generation\. The 18 publications from other countries fall into ten areas, mainly Resources and Evaluation \(9\), Syntax \(5\), NLP Applications \(5\), and Multilingualism \(5\), and most concern Old or Classical Tibetan and Buddhist texts\([Long and Hill, 2021](https://arxiv.org/html/2610.13398#bib.bib142), cf\.\): UK\-based researchers built treebanks and annotated corpora\([Faggionato and Meelen, 2019](https://arxiv.org/html/2610.13398#bib.bib27);[Meelen and Roux, 2020](https://arxiv.org/html/2610.13398#bib.bib25);[Meelen et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib112)\); Israel\-based researchers, with German Tibetologists, developed evaluation resources and methods for Buddhist texts\([Hashiloni et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib9);[Cohen et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib6);[Bilitski et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib7)\); and researchers in Germany, France, and the USA contributed a Sanskrit–Tibetan parallel corpus, e\-text tools, text analytics, and OCR spelling correction\([Nehrdich, 2022](https://arxiv.org/html/2610.13398#bib.bib22);[Trinley et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib115);[Krishna et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib109);[Luo and Chuang, 2024](https://arxiv.org/html/2610.13398#bib.bib91)\)\. US\-based work also covers modern Tibetan segmentation, speech recognition evaluation, and tonogenesis\([Brown, 2024](https://arxiv.org/html/2610.13398#bib.bib17);[Moore et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib75);[Liang and Zerong, 2026](https://arxiv.org/html/2610.13398#bib.bib8)\)\.

The two groups also publish in different venues: two\-thirds of the publications from outside China \(12 of 18\) appear in the ACL Anthology, seven of them at workshops, whereas 97 of the 122 China\-affiliated publications \(79\.5%\) appear in journals or in conferences outside it\. Tibetan NLP thus comprises two partly separate communities that differ in research focus as well as in venue, although exchange is productive where it occurs, as in the 2021 TALLIP special issue with papers from China, France, and the USA\([Long and Hill, 2021](https://arxiv.org/html/2610.13398#bib.bib142)\)\.

## 5Discussion

### 5\.1A Full\-Pipeline Paradigm and Its Limits

Our results characterize Tibetan NLP byconcentration\(87\.1% first\-authored in China\),breadth\(China\-affiliated work covers all 25 represented areas\), andlayering\(new layers added, earlier ones kept\)\. We call this pattern a full\-pipeline paradigm: research spans every layer from script processing to LLMs, though not every task in each layer\. The paradigm also has limits\. First, it is predominantly engineering\-oriented, with little fundamental research\([Long and Hill, 2021](https://arxiv.org/html/2610.13398#bib.bib142), cf\.\)\. Second, it is only partly visible: most China\-affiliated research lies outside the ACL Anthology, and hence outside surveys restricted to it\([Liu and Best, 2025](https://arxiv.org/html/2610.13398#bib.bib141)\)\. Third, its outputs are only partly reusable: 41 of the 140 publications \(29\.3%\) release data, code, or models at a public URL \(Appendix[F](https://arxiv.org/html/2610.13398#A6)\)\. Release has become more common, from none of the 31 publications of 2004–2018 to 27 of the 61 of 2024–2026 \(44\.3%\), but remains less frequent\.

### 5\.2Lessons for the Global South

Many low\-resource languages of the Global South face obstacles usually studied in isolation: dispersed speakers, dialectal diversity, scripts that complicate encoding and segmentation, and limited infrastructure and investment\. All co\-occur on the Qinghai\-Xizang Plateau, yet Tibetan research has advanced across the entire pipeline, suggesting four transferable lessons: \(1\) invest first in the base layers of script processing, segmentation, and annotated corpora, on which later work depends; \(2\) treat resource and benchmark construction as a continuing frontier, not a one\-off step \(§[4\.2](https://arxiv.org/html/2610.13398#S4.SS2)\); \(3\) exploit a high\-resource contact language, here Mandarin Chinese, for cross\-lingual transfer, while guarding against translated data and benchmarks, which can lose cultural content and misstate difficulty\([Zhuang and Sun, 2025](https://arxiv.org/html/2610.13398#bib.bib14);[Li et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib1)\); and \(4\) use dedicated venues to consolidate a research community \(§[4\.1](https://arxiv.org/html/2610.13398#S4.SS1)\)\. Transfer is conditional: Tibetan NLP has had sustained public funding \(§[4\.3](https://arxiv.org/html/2610.13398#S4.SS3)\), whereas African NLP has grown through community initiatives and targeted funding\([Alabi et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib143)\); communities with neither will rely more on international collaboration and open resources\.

### 5\.3The Transition to LLMs

Since 2022, the field has pivoted toward pre\-trained models and LLMs, from three separate LLaMA\-2 extensions with Tibetan vocabularies and 6–26 GB of text\([Lv et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib86);[Sha et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib80);[Zhuang et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib87)\)to dense and mixture\-of\-experts models trained on a curated 72 GB corpus\([Yang et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib2)\), vocabulary sizing\([Han et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib5)\), reinforcement learning that limits forgetting of Chinese\([Su et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib3)\), and vision–language models\([Xu et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib4)\)\. Two obstacles recur\. Tokenization is unresolved: a syllable vocabulary of 35K types still leaves 15% of tokens unknown\([Liu et al\., 2022](https://arxiv.org/html/2610.13398#bib.bib51)\), whereas multilingual tokenizers fragment Tibetan, so that a 27\-token English sentence takes 100 or more in Tibetan\([Zhang et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib13)\), and expanding byte\-level byte\-pair\-encoding vocabularies helps only beyond about 3,000 new tokens per language\([Han et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib5)\)\. Data are noisy and partly closed: 42% of sampled Tibetan CulturaX pages retain headers or footers\([Zhang et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib15)\), CINO’s pre\-training corpus is not public\([Zhuang and Sun, 2025](https://arxiv.org/html/2610.13398#bib.bib14)\), and 41 GB of the 72 GB corpus are private\([Yang et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib2)\)\.

Evaluation reveals a large gap\. GPT\-3\.5 reaches 3\.6% F1 on Tibetan WikiAnn named\-entity recognition, the lowest of ten low\-resource Sino\-Tibetan languages\([Liu and Best, 2025](https://arxiv.org/html/2610.13398#bib.bib141)\); on TLUE, most LLMs perform below a random baseline, and Qwen2\.5\-72B drops from 84\.7% on the Chinese CMMLU to 16\.5% on its Tibetan counterpart\([Gao et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib12)\); and general multimodal LLMs recognize 49–66% of characters on TibOCR\-Bench, against 92% for a 7B model tuned on 4\.5K Tibetan images\([Jie et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib71)\)\. Evaluation is itself a bottleneck\. Encoders are compared chiefly on one news\-classification corpus, TNCC\([Liu et al\., 2022](https://arxiv.org/html/2610.13398#bib.bib51);[Zhou et al\., 2024b](https://arxiv.org/html/2610.13398#bib.bib43);[Hao et al\., 2024a](https://arxiv.org/html/2610.13398#bib.bib41);[Cao et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib38)\), and task data are increasingly obtained by pivoting through other languages, as in offensive\-language and sentiment datasets translated from English or Chinese\([Ren et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib46);[Zhang et al\., 2026a](https://arxiv.org/html/2610.13398#bib.bib78)\)and summaries bootstrapped through Chinese with GPT\-3\.5\([Hao et al\., 2024b](https://arxiv.org/html/2610.13398#bib.bib47)\)\. Alignment data, too, are wholly or partly translated, from GPT\-4 answers to English prompts\([Zhuang et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib87)\)or Chinese preference data\([Chen et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib69)\)\. Many LLM benchmarks are likewise translated\([Gao et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib12);[Xu et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib4);[Yang et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib2)\), yet machine\-translated test items depress scores, most of all for Tibetan\([Zhang et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib13)\), and tasks built from native minority\-language knowledge are 11\.4–13\.9 points harder than translated ones\([Li et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib1)\)\. The move to LLMs thus risks outpacing its foundations, with dialects underserved and evaluation standards only emerging \(§[6](https://arxiv.org/html/2610.13398#S6)\)\.

### 5\.4Classical and Modern Tibetan

The corpus contains two research traditions, centred on modern Tibetan and its spoken varieties and on Old and Classical Tibetan, which differ in who conducts them and where they appear \(§[4\.3](https://arxiv.org/html/2610.13398#S4.SS3)\)\. China\-affiliated work on historical materials has focused on document\-image analysis\([Wang et al\., 2018a](https://arxiv.org/html/2610.13398#bib.bib55);[Ma et al\., 2020](https://arxiv.org/html/2610.13398#bib.bib120);[Li et al\., 2022c](https://arxiv.org/html/2610.13398#bib.bib106)\), the frontier of character recognition once OCR of modern print was in effect solved\([Long and Hill, 2021](https://arxiv.org/html/2610.13398#bib.bib142)\), and on interlinearization\([Long and An, 2023](https://arxiv.org/html/2610.13398#bib.bib19)\)\. Shared archives already link the traditions: Buddhist Digital Resource Center scans underlie both a China\-built character dataset\([Li et al\., 2022c](https://arxiv.org/html/2610.13398#bib.bib106)\)and US\-based OCR post\-correction\([Luo and Chuang, 2024](https://arxiv.org/html/2610.13398#bib.bib91)\)\. Transfer, however, works more readily within writing than for speech: Old Tibetan has been normalized to Classical orthography to reuse its tools\([Faggionato and Meelen, 2019](https://arxiv.org/html/2610.13398#bib.bib27)\), and a single pipeline annotates texts from the 11th to the 21st century because literary Tibetan has changed little\([Meelen and Roux, 2020](https://arxiv.org/html/2610.13398#bib.bib25)\), whereas pre\-training on Classical text did not help segment spoken Standard Tibetan\([Brown, 2024](https://arxiv.org/html/2610.13398#bib.bib17)\), and a segmenter with a classical and literary dictionary marks 87% of syllable boundaries in transcribed Lhasa speech as word boundaries, against 41% for an annotator\([Moore et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib75)\)\. Domain also matters within writing: general\-domain Tibetan–Chinese MT drops from about 50 BLEU to at most 5 on Buddhist text\([Gyatso et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib89)\)\. Classical treebanks and parallel corpora could support modern written Tibetan, and modern OCR and parsing the digitization of classical texts, while spoken varieties need their own resources\.

## 6Recommendations

#### Open science and cross\-community exchange\.

As fewer than a third of surveyed publications release resources \(§[5\.1](https://arxiv.org/html/2610.13398#S5.SS1)\), we encourage public release, not only on request, of datasets, models, and code under documented licences, and of pre\-training corpora where rights permit\([Trinley et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib115);[Zhang et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib15), as in\), following the practices in Appendix[F](https://arxiv.org/html/2610.13398#A6)\. To make work findable, authors should name the varieties studied in titles and abstracts, and venues could record them as metadata\([Liu and Best, 2025](https://arxiv.org/html/2610.13398#bib.bib141)\); joint workshops, special issues, and shared tasks could connect China\-affiliated research with research elsewhere, and the classical with the modern tradition \(§[5\.4](https://arxiv.org/html/2610.13398#S5.SS4)\)\.

#### Shared standards and native evaluation\.

To make results comparable across teams, the field needs shared conventions for encoding normalization, segmentation, and unit terminology; public test sets for core tasks; and reporting of splits, metrics, and the varieties and scripts covered\([Kumar et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib146)\)\. Shared segmentation tasks offer a basis, but their single gold segmentation covers only 22% of valid granularities\([Zhuoma and Zhijie, 2026](https://arxiv.org/html/2610.13398#bib.bib74)\)\. LLM benchmarks should increasingly be built from native material with expert involvement, as in CMiLBench’s minority\-knowledge tasks and DharmaBench\([Li et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib1);[Hashiloni et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib9)\), rather than by translation alone\.

#### Community participation\.

No surveyed publication assesses Tibetan speakers’ needs, yet needs assessments show that priorities vary across speaker groups, even across dialects of one language\([Al Kautsar et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib148)\); German dialect speakers, for instance, favour tools that take spoken dialect input over those producing dialect output\([Blaschke et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib147)\), a preference worth testing in diglossic Tibetan\. Resource creation and evaluation should involve speakers of all three major dialect groups, including diaspora communities, throughout research\([Tonja et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib145)\), as handwriting datasets with up to 235 writers from five regions\([Tashi et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib95);[Zhou et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib76)\)and a readability tool built on children’s speech\([Schmidt, 2020](https://arxiv.org/html/2610.13398#bib.bib118)\)begin to do\.

#### Responsible NLP\.

Beyond the gaps noted in §[4\.2](https://arxiv.org/html/2610.13398#S4.SS2), safety is evaluated only within recent benchmarks, which show that it does not carry over to Tibetan: most LLMs fall below chance on TLUE’s safety set\([Gao et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib12)\), and safety alignment transfers unreliably across minority languages\([Li et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib1)\)\. Bias is unmeasured, although zero\-shot LLMs tend to label Tibetan news as politics or religion\([Tang et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib81)\)\. Applications that mine user\-level social\-media data for public\-opinion monitoring or to locate user groups\([Ye et al\., 2014](https://arxiv.org/html/2610.13398#bib.bib61);[Qiu et al\., 2018b](https://arxiv.org/html/2610.13398#bib.bib134);[Xu and Gao, 2024](https://arxiv.org/html/2610.13398#bib.bib40)\), microblog corpora stored with user names\([Qiu et al\., 2017](https://arxiv.org/html/2610.13398#bib.bib138);[Qiu et al\., 2018a](https://arxiv.org/html/2610.13398#bib.bib135)\), speaker verification motivated partly by criminal investigation\([Gan et al\., 2022](https://arxiv.org/html/2610.13398#bib.bib107)\), and an OCR benchmark with identity cards and bank receipts but no stated anonymization\([Jie et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib71)\)raise privacy and dual\-use questions that the absence of ethics research leaves unexamined and that call for the consent and privacy safeguards urged for African languages\([Adebara and Abdul\-Mageed, 2022](https://arxiv.org/html/2610.13398#bib.bib144)\)\. Bias, safety, privacy, and human\-centred evaluation need study before Tibetan LLMs are deployed in education, healthcare, or public services\.

#### Dialects and related languages\.

Only eight publications name Lhasa, Amdo, or a multi\-dialect setting in their titles, all on speech recognition, synthesis, or translation\. Open speech data for all three groups total about 80–100 hours, against over 10,000 for Mandarin or English\([Liu et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib96)\), and Kham is their smallest, hardest part: the main multi\-dialect corpus, TIBMD@MUC, was released with 8\.7 hours of Kham speech against 30 of Lhasa speech\([Zhao et al\., 2020](https://arxiv.org/html/2610.13398#bib.bib123)\), and even the best recognizer trained on it errs on 23% of Kham characters, against 3–5% for the other groups\([Liu et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib96)\)\. Kham and Amdo resources should be prioritized, publicly released, and evaluated per variety\. Dzongkha and Ladakhi use Tibetan script, so OCR and segmentation tools may transfer directly, and Burmese could share low\-level resources\([Fang et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib44)\)\.

## 7Conclusion

We have surveyed 140 Tibetan NLP publications from 2004 to 2026\. Since 2018, the field has grown rapidly and expanded bottom\-up along the NLP pipeline while continuing to build resources, with 87% of the work first\-authored in China\. This full\-pipeline paradigm can serve as a reference model for other low\-resource languages of the Global South; its limits in dialect coverage, responsible NLP, native and comparable evaluation, resource release, and cross\-community visibility set the agenda for keeping plateau\-region languages from being left behind in the LLM era\.

## Limitations

This survey has several limitations\. First, its coverage is bounded by its sources: we include only publications indexed in the ACL Anthology or DBLP, so work in Chinese\-language venues, in venues indexed by neither source, and in preprints is not covered, and topics published mainly in such venues may be underrepresented\. Second, keyword\-based retrieval may miss multilingual studies that include Tibetan without naming it in the searched metadata fields; this risk is greater for DBLP, whose records do not include abstracts\. Third, assigning research areas required interpretive judgement, especially for work spanning several topics\. Fourth, the country\-level analysis relies on the first author’s primary affiliation, which may not reflect all contributors or the national context of the research\. Fifth, publication counts measure research output rather than impact or quality, and we did not perform a citation analysis; our audit of released resources \(Appendix[F](https://arxiv.org/html/2610.13398#A6)\) records whether resources are publicly available at a URL given by their authors, not whether they are complete, documented, or maintained, and may miss resources released elsewhere or whose links have lapsed\. Sixth, our qualitative synthesis, including the discussion of evaluation practice and of the LLM era, relies on results reported by the surveyed publications, which use different models, data, and metrics and which we did not reproduce\. Finally, the survey reflects the literature available up to 00:00 UTC, September 25, 2026, so 2026 is only partially covered\. We release the full bibliography, in raw and refined versions, to facilitate community\-driven correction and extension, and we will continue to track new work:[https://github\.com/metaphors/TibetanNLPSurvey](https://github.com/metaphors/TibetanNLPSurvey)\.

## Ethical Considerations

This survey analyses publicly available bibliographic metadata and peer\-reviewed publications; it involves no human subjects and no personal data beyond author names and affiliations as they appear in the public record\. Country\-level statistics describe institutional affiliations, not the nationality or ethnicity of researchers, and differences in resource release rates may reflect venue norms and legitimate constraints, such as consent and privacy requirements for speech data or missing permission to share collected data, rather than the practices of individual teams\. Where we discuss the data practices or stated applications of individual publications, we do so to motivate shared norms, not to single out particular teams\. We do not reproduce or redistribute copyrighted material beyond standard academic citation practices\. We have aimed to describe research trends accurately without advancing normative political positions, and we recognize that decisions about language technology for Tibetan should involve the communities who speak it\.

## Acknowledgments

This work is supported by the Science and Technology Strategic Consulting Project of the Chinese Academy of Engineering \(2025\-XZ\-16\-06\), the Project of the China Tibetology Literature Resources Data Center \(2025SJ003\), the National Social Science Foundation of China \(22&ZD035\), the Key Project of the Natural Science Foundation of Xizang Autonomous Region \(XZ202401ZR0040\), and the Major Project of the Science and Technology Program of Xizang Autonomous Region \(XZ202601ZD0010\)\.

## References

- I\. Adebara and M\. Abdul\-MageedTowards afrocentric NLP for African languages: where we are and where we can go\.InProceedings of the 60th Annual Meeting of the Association for Computational Linguistics \(Volume 1: Long Papers\),S\. Muresan, P\. Nakov, and A\. Villavicencio \(Eds\.\),Dublin, Ireland,pp\. 3814–3841\.External Links:[Link](https://aclanthology.org/2022.acl-long.265/),[Document](https://dx.doi.org/10.18653/v1/2022.acl-long.265)Cited by:[§2\.2](https://arxiv.org/html/2610.13398#S2.SS2.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Al Kautsaret al\.\(2025\)M\. D\. Al Kautsar, L\. Susanto, D\. T\. Wijaya, and F\. KotoWhat do Indonesians really need from language technology? A nationwide survey\.InProceedings of the 2025 Conference on Empirical Methods in Natural Language Processing,C\. Christodoulopoulos, T\. Chakraborty, C\. Rose, and V\. Peng \(Eds\.\),Suzhou, China,pp\. 7220–7245\.External Links:[Link](https://aclanthology.org/2025.emnlp-main.367/),[Document](https://dx.doi.org/10.18653/v1/2025.emnlp-main.367),ISBN 979\-8\-89176\-332\-6Cited by:[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px3.p1.1)\.
- Alabiet al\.\(2025\)J\. O\. Alabi, M\. A\. Hedderich, D\. I\. Adelani, and D\. KlakowCharting the landscape of African NLP: mapping progress and shaping the road ahead\.InProceedings of the 2025 Conference on Empirical Methods in Natural Language Processing,C\. Christodoulopoulos, T\. Chakraborty, C\. Rose, and V\. Peng \(Eds\.\),Suzhou, China,pp\. 27807–27841\.External Links:[Link](https://aclanthology.org/2025.emnlp-main.1414/),[Document](https://dx.doi.org/10.18653/v1/2025.emnlp-main.1414),ISBN 979\-8\-89176\-332\-6Cited by:[§2\.2](https://arxiv.org/html/2610.13398#S2.SS2.p1.1),[§5\.2](https://arxiv.org/html/2610.13398#S5.SS2.p1.1)\.
- An and Long \(2021\)B\. An and C\. LongNeural Dependency Parser for Tibetan Sentences\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.20\(2\),pp\. 30:1–30:16\.External Links:[Link](https://doi.org/10.1145/3429456),[Document](https://dx.doi.org/10.1145/3429456)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- An \(2023a\)B\. AnPrompt\-based for Low\-Resource Tibetan Text Classification\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.22\(8\),pp\. 207:1–207:13\.External Links:[Link](https://doi.org/10.1145/3603168),[Document](https://dx.doi.org/10.1145/3603168)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1)\.
- An \(2023b\)B\. AnTibetan dependency parsing with graph convolutional neural networks\.InProceedings of the Ancient Language Processing Workshop,A\. Anderson, S\. Gordin, B\. Li, Y\. Liu, and M\. C\. Passarotti \(Eds\.\),Varna, Bulgaria,pp\. 213–221\.External Links:[Link](https://aclanthology.org/2023.alp-1.24/)Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- Baoet al\.\(2024\)X\. Bao, W\. Wang, X\. Wang, G\. Zhao, H\. Li, and M\. LiuToward Better Generalization of Cross\-Domain Few\-Shot Classification in Tibetan Character With Contrastive Learning and Meta Fine\-Tuning\.IEEE Access12,pp\. 134439–134452\.External Links:[Link](https://doi.org/10.1109/ACCESS.2024.3459933),[Document](https://dx.doi.org/10.1109/ACCESS.2024.3459933)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1)\.
- Bilitskiet al\.\(2026\)G\. Bilitski, L\. Shechter, S\. Jamtsho, N\. Marciano, N\. Bajetta, R\. Sunden, O\. Drori, K\. G\. Hashiloni, O\. Zwebner, A\. Shina, O\. Almogi, D\. Wangchuk, and K\. BarAutomatic segmentation of Classical Tibetan texts into autochthonous and allochthonous regions\.InProceedings of the Fifteenth Language Resources and Evaluation Conference,S\. Piperidis, N\. Bel, H\. van den Heuvel, N\. Ide, S\. Krek, and A\. Toral \(Eds\.\),Palma de Mallorca, Spain,pp\. 1017–1030\.External Links:[Link](https://aclanthology.org/2026.lrec-1.79/),[Document](https://dx.doi.org/10.63317/2iyfjjv9boc6)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.6](https://arxiv.org/html/2610.13398#A3.SS6.p1.1),[§D\.21](https://arxiv.org/html/2610.13398#A4.SS21.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1)\.
- Blaschkeet al\.\(2024\)V\. Blaschke, C\. Purschke, H\. Schuetze, and B\. PlankWhat do dialect speakers want? A survey of attitudes towards language technology for German dialects\.InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics \(Volume 2: Short Papers\),L\. Ku, A\. Martins, and V\. Srikumar \(Eds\.\),Bangkok, Thailand,pp\. 823–841\.External Links:[Link](https://aclanthology.org/2024.acl-short.74/),[Document](https://dx.doi.org/10.18653/v1/2024.acl-short.74)Cited by:[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px3.p1.1)\.
- Brown \(2024\)C\. J\. BrownImproved neural word segmentation for standard Tibetan\.InProceedings of the 2nd Workshop on Resources and Technologies for Indigenous, Endangered and Lesser\-resourced Languages in Eurasia \(EURALI\) @ LREC\-COLING 2024,A\. Kr\. Ojha, S\. Ahmadi, S\. Cinková, T\. Fransen, C\. Liu, and J\. P\. McCrae \(Eds\.\),Torino, Italia,pp\. 12–17\.External Links:[Link](https://aclanthology.org/2024.eurali-1.2/)Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.8](https://arxiv.org/html/2610.13398#A3.SS8.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Cai and Wang \(2018\)Z\. Cai and W\. WangOnline Handwriting Tibetan Character Recognition Based on Two\-Dimensional Discriminant Locality Alignment\.InPattern Recognition and Computer Vision \- First Chinese Conference, PRCV 2018, Guangzhou, China, November 23\-26, 2018, Proceedings, Part III,J\. Lai, C\. Liu, X\. Chen, J\. Zhou, T\. Tan, N\. Zheng, and H\. Zha \(Eds\.\),Lecture Notes in Computer Science, Vol\.11258,pp\. 88–98\.External Links:[Link](https://doi.org/10.1007/978-3-030-03338-5/_8),[Document](https://dx.doi.org/10.1007/978-3-030-03338-5%5F8)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Caoet al\.\(2023\)X\. Cao, D\. Dawa, N\. Qun, and T\. NyimaPay attention to the robustness of Chinese minority language models\! Syllable\-level textual adversarial attack on Tibetan script\.InProceedings of the 3rd Workshop on Trustworthy Natural Language Processing \(TrustNLP 2023\),A\. Ovalle, K\. Chang, N\. Mehrabi, Y\. Pruksachatkun, A\. Galystan, J\. Dhamala, A\. Verma, T\. Cao, A\. Kumar, and R\. Gupta \(Eds\.\),Toronto, Canada,pp\. 35–46\.External Links:[Link](https://aclanthology.org/2023.trustnlp-1.4/),[Document](https://dx.doi.org/10.18653/v1/2023.trustnlp-1.4)Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.11](https://arxiv.org/html/2610.13398#A4.SS11.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px5.p1.1)\.
- Caoet al\.\(2025a\)X\. Cao, Q\. Gesang, Y\. Sun, N\. Qun, and T\. NyimaTSCheater: Generating High\-Quality Tibetan Adversarial Texts via Visual Similarity\.In2025 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2025, Hyderabad, India, April 6\-11, 2025,pp\. 1–5\.External Links:[Link](https://doi.org/10.1109/ICASSP49660.2025.10889732),[Document](https://dx.doi.org/10.1109/ICASSP49660.2025.10889732)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.11](https://arxiv.org/html/2610.13398#A4.SS11.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px5.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Caoet al\.\(2025b\)X\. Cao, Y\. Sun, J\. Li, Q\. Gesang, N\. Qun, and N\. TashiHuman\-in\-the\-loop generation of adversarial texts: a case study on Tibetan script\.InProceedings of The 14th International Joint Conference on Natural Language Processing and The 4th Conference of the Asia\-Pacific Chapter of the Association for Computational Linguistics: System Demonstrations,X\. Liu and A\. Purwarianti \(Eds\.\),Mumbai, India,pp\. 9–16\.External Links:[Link](https://aclanthology.org/2025.ijcnlp-demo.2/),[Document](https://dx.doi.org/10.18653/v1/2025.ijcnlp-demo.2),ISBN 979\-8\-89176\-301\-2Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.11](https://arxiv.org/html/2610.13398#A4.SS11.p1.1),[§D\.8](https://arxiv.org/html/2610.13398#A4.SS8.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px5.p1.1)\.
- Chenet al\.\(2026\)L\. Chen, G\. Duojie, and Y\. SunDataset Used for Reinforcement Learning Tasks of Tibetan Large Models\.Data Intell\.8\(2\),pp\. 20250044\.External Links:[Link](https://doi.org/10.3724/2096-7004.di.2025.0044),[Document](https://dx.doi.org/10.3724/2096-7004.DI.2025.0044)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Cohenet al\.\(2026\)S\. Cohen, J\. Yang, G\. Rabinovitz, S\. Choden, O\. Shtrosberg, N\. Bajetta, G\. B\. Horin, R\. Sundén, O\. Drori, S\. Jamtsho, D\. Wangchuk, K\. Bar, O\. Almogi, and S\. FineScaling sentence similarity for Classical Tibetan with automatic annotations\.InProceedings of the 6th International Conference on Natural Language Processing for the Digital Humanities,S\. Hamilton, E\. Öhman, R\. M\. M\. Hicke, Y\. Bizzoni, A\. Bax, J\. A\. Matthews, and M\. Hämäläinen \(Eds\.\),San Diego, USA,pp\. 150–166\.External Links:[Link](https://aclanthology.org/2026.nlp4dh-1.15/),[Document](https://dx.doi.org/10.18653/v1/2026.nlp4dh-1.15),ISBN 979\-8\-89176\-427\-9Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.6](https://arxiv.org/html/2610.13398#A3.SS6.p1.1),[§D\.21](https://arxiv.org/html/2610.13398#A4.SS21.p1.1),[§D\.25](https://arxiv.org/html/2610.13398#A4.SS25.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1)\.
- Dan and Sun \(2024\)Z\. Dan and Y\. SunTibetanQA2\.0: Dataset with Unanswerable Questions for Tibetan Machine Reading Comprehension\.Data Intell\.6\(4\),pp\. 1158–1167\.External Links:[Link](https://doi.org/10.3724/2096-7004.di.2024.0029),[Document](https://dx.doi.org/10.3724/2096-7004.DI.2024.0029)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.23](https://arxiv.org/html/2610.13398#A4.SS23.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Diet al\.\(2019\)Q\. Di, E\. Vylomova, and T\. BaldwinModelling Tibetan verbal morphology\.InProceedings of the 17th Annual Workshop of the Australasian Language Technology Association,M\. Mistica, M\. Piccardi, and A\. MacKinlay \(Eds\.\),Sydney, Australia,pp\. 35–40\.External Links:[Link](https://aclanthology.org/U19-1005/)Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.1](https://arxiv.org/html/2610.13398#A3.SS1.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1)\.
- Faggionatoet al\.\(2022\)C\. Faggionato, N\. Hill, and M\. MeelenNLP pipeline for annotating \(endangered\) Tibetan and Newar varieties\.InProceedings of the Workshop on Resources and Technologies for Indigenous, Endangered and Lesser\-resourced Languages in Eurasia within the 13th Language Resources and Evaluation Conference,A\. Kr\. Ojha, S\. Ahmadi, C\. Liu, and J\. P\. McCrae \(Eds\.\),Marseille, France,pp\. 1–6\.External Links:[Link](https://aclanthology.org/2022.eurali-1.1/)Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.7](https://arxiv.org/html/2610.13398#A3.SS7.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px1.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1)\.
- Faggionato and Meelen \(2019\)C\. Faggionato and M\. MeelenDeveloping the Old Tibetan treebank\.InProceedings of the International Conference on Recent Advances in Natural Language Processing \(RANLP 2019\),R\. Mitkov and G\. Angelova \(Eds\.\),Varna, Bulgaria,pp\. 304–312\.External Links:[Link](https://aclanthology.org/R19-1035/),[Document](https://dx.doi.org/10.26615/978-954-452-056-4%5F035)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.7](https://arxiv.org/html/2610.13398#A3.SS7.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Fanet al\.\(2026\)H\. Fan, Y\. Zhao, J\. Yu, Z\. Gong, X\. Xu, and H\. LiTibetan\-Chinese Speech\-to\-Speech Translation Based on Large Speech Models\.Data Intell\.8\(2\),pp\. 20250137\.External Links:[Link](https://doi.org/10.3724/2096-7004.di.2025.0137),[Document](https://dx.doi.org/10.3724/2096-7004.DI.2025.0137)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.16](https://arxiv.org/html/2610.13398#A4.SS16.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Fanget al\.\(2024\)J\. Fang, W\. Huang, A\. Yang, D\. Zhou, and N\. LinMLCL: A Framework for Reducing Language Imbalance in Sino\-Tibetan Languages through Adapter Structures\.InAsian Conference on Machine Learning, 5\-8 December 2024, Hanoi, Vietnam,V\. Nguyen and H\. Lin \(Eds\.\),Proceedings of Machine Learning Research, Vol\.260,pp\. 431–446\.External Links:[Link](https://proceedings.mlr.press/v260/fang25a.html)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px5.p1.1)\.
- Ganet al\.\(2022\)Z\. Gan, Y\. Yu, and M\. LuoA Tibetan\-dependent speaker recognition method based on deep learning\.Multim\. Tools Appl\.81\(21\),pp\. 30821–30840\.External Links:[Link](https://doi.org/10.1007/s11042-022-12540-9),[Document](https://dx.doi.org/10.1007/S11042-022-12540-9)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Gaoet al\.\(2025\)F\. Gao, C\. Huang, Y\. Liu, N\. Tashi, X\. Wang, T\. Tsering, B\. Ma\-bao, R\. Duojie, G\. Luosang, R\. Dongrub, D\. Tashi, X\. F\. Cd, Y\. Yu, and H\. WangTLUE: a Tibetan language understanding evaluation benchmark\.InProceedings of the 2025 Conference on Empirical Methods in Natural Language Processing,C\. Christodoulopoulos, T\. Chakraborty, C\. Rose, and V\. Peng \(Eds\.\),Suzhou, China,pp\. 35071–35097\.External Links:[Link](https://aclanthology.org/2025.emnlp-main.1777/),[Document](https://dx.doi.org/10.18653/v1/2025.emnlp-main.1777),ISBN 979\-8\-89176\-332\-6Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Gurgurovet al\.\(2024\)D\. Gurgurov, M\. Hartmann, and S\. OstermannAdapting multilingual LLMs to low\-resource languages with knowledge graphs via adapters\.InProceedings of the 1st Workshop on Knowledge Graphs and Large Language Models \(KaLLM 2024\),R\. Biswas, L\. Kaffee, O\. Agarwal, P\. Minervini, S\. Singh, and G\. de Melo \(Eds\.\),Bangkok, Thailand,pp\. 63–74\.External Links:[Link](https://aclanthology.org/2024.kallm-1.7/),[Document](https://dx.doi.org/10.18653/v1/2024.kallm-1.7)Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.4](https://arxiv.org/html/2610.13398#A3.SS4.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1)\.
- Gyatsoet al\.\(2025\)K\. Gyatso, S\. Tshering, T\. Norbu, N\. Tashi, T\. Xiao, J\. Zhu, G\. Tashi, and G\. LuosangAn Improved Byte Pair Encoding Method for Tibetan\.J\. Adv\. Comput\. Intell\. Intell\. Informatics29\(6\),pp\. 1273–1282\.External Links:[Link](https://doi.org/10.20965/jaciii.2025.p1273),[Document](https://dx.doi.org/10.20965/JACIII.2025.P1273)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.16](https://arxiv.org/html/2610.13398#A4.SS16.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Hanet al\.\(2026\)A\. Han, A\. Chen, Y\. Sun, and X\. ZhaoScaling laws or threshold effects: exploring the optimal vocabulary size for balancing performance and efficiency in low\-resource languages\.InFindings of the Association for Computational Linguistics: ACL 2026,M\. Liakata, V\. P\. Moreira, J\. Zhang, and D\. Jurgens \(Eds\.\),San Diego, California, United States,pp\. 31741–31758\.External Links:[Link](https://aclanthology.org/2026.findings-acl.1588/),[Document](https://dx.doi.org/10.18653/v1/2026.findings-acl.1588),ISBN 979\-8\-89176\-395\-1Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.5](https://arxiv.org/html/2610.13398#A4.SS5.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px6.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1)\.
- Hanet al\.\(2018\)Y\. Han, W\. Wang, Y\. Wang, and X\. WangResearch on the Method of Tibetan Recognition Based on Component Location Information\.InPattern Recognition and Computer Vision \- First Chinese Conference, PRCV 2018, Guangzhou, China, November 23\-26, 2018, Proceedings, Part III,J\. Lai, C\. Liu, X\. Chen, J\. Zhou, T\. Tan, N\. Zheng, and H\. Zha \(Eds\.\),Lecture Notes in Computer Science, Vol\.11258,pp\. 63–73\.External Links:[Link](https://doi.org/10.1007/978-3-030-03338-5/_6),[Document](https://dx.doi.org/10.1007/978-3-030-03338-5%5F6)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Haoet al\.\(2024a\)M\. Hao, X\. Yan, and X\. OuyangA Tibetan Text Classification Method Based on Hybrid Model and Channel Attention Mechanism\.In27th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2024, Tianjin, China, May 8\-10, 2024,W\. Shen, J\. A\. Barthès, J\. Luo, T\. Qiu, X\. Zhou, J\. Zhang, H\. Zhu, K\. Peng, T\. Xu, and N\. Chen \(Eds\.\),pp\. 1522–1527\.External Links:[Link](https://doi.org/10.1109/CSCWD61410.2024.10580150),[Document](https://dx.doi.org/10.1109/CSCWD61410.2024.10580150)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Haoet al\.\(2024b\)Y\. Hao, B\. Chen, and X\. ZhaoTiLTS:Tibetan Long Text Summarization Dataset\.InNatural Language Processing and Chinese Computing \- 13th National CCF Conference, NLPCC 2024, Hangzhou, China, November 1\-3, 2024, Proceedings, Part IV,D\. F\. Wong, Z\. Wei, and M\. Yang \(Eds\.\),Lecture Notes in Computer Science, Vol\.15362,pp\. 265–276\.External Links:[Link](https://doi.org/10.1007/978-981-97-9440-9/_21),[Document](https://dx.doi.org/10.1007/978-981-97-9440-9%5F21)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.28](https://arxiv.org/html/2610.13398#A4.SS28.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Hashiloniet al\.\(2025\)K\. G\. Hashiloni, S\. Cohen, A\. Shina, J\. Yang, O\. M\. Zwebner, N\. Bajetta, G\. Bilitski, R\. Sundén, G\. Maduel, R\. Conlon, A\. Barzilai, D\. Mass, S\. Jia, A\. Naaman, S\. Choden, S\. Jamtsho, Y\. Qu, H\. Isaacson, D\. Wangchuk, S\. Fine, O\. Almogi, and K\. BarDharmaBench: evaluating language models on Buddhist texts in Sanskrit and Tibetan\.InProceedings of the 14th International Joint Conference on Natural Language Processing and the 4th Conference of the Asia\-Pacific Chapter of the Association for Computational Linguistics,K\. Inui, S\. Sakti, H\. Wang, D\. F\. Wong, P\. Bhattacharyya, B\. Banerjee, A\. Ekbal, T\. Chakraborty, and D\. P\. Singh \(Eds\.\),Mumbai, India,pp\. 2088–2110\.External Links:[Link](https://aclanthology.org/2025.ijcnlp-long.114/),[Document](https://dx.doi.org/10.18653/v1/2025.ijcnlp-long.114),ISBN 979\-8\-89176\-298\-5Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.6](https://arxiv.org/html/2610.13398#A3.SS6.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px2.p1.1)\.
- Jianget al\.\(2005\)D\. Jiang, C\. Long, and J\. ZhangThe verbal entries and their description in a grammatical information\-dictionary of contemporary Tibetan\.InSecond International Joint Conference on Natural Language Processing: Full Papers,External Links:[Link](https://aclanthology.org/I05-1076/),[Document](https://dx.doi.org/10.1007/11562214%5F76)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- Jiang \(2006\)D\. JiangThe current status of sorting order of Tibetan dictionaries and standardization\.InProceedings of the 20th Pacific Asia Conference on Language, Information and Computation,Huazhong Normal University, Wuhan, China,pp\. 231–236\.External Links:[Link](https://aclanthology.org/Y06-1030/),[Document](https://dx.doi.org/http%3A//hdl.handle.net/2065/29059)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- Jieet al\.\(2026\)L\. Jie, M\. Cairang, K\. Khysru, H\. G\. C\. Rang, J\. Jiahui, Y\. Yingkai, and T\. QianTibOCR\-Bench: A Comprehensive Benchmark and Training Pipeline for Tibetan Multimodal OCR\.Data Intell\.8\(2\),pp\. 20250140\.External Links:[Link](https://doi.org/10.3724/2096-7004.di.2025.0140),[Document](https://dx.doi.org/10.3724/2096-7004.DI.2025.0140)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Joshiet al\.\(2020\)P\. Joshi, S\. Santy, A\. Budhiraja, K\. Bali, and M\. ChoudhuryThe state and fate of linguistic diversity and inclusion in the NLP world\.InProceedings of the 58th Annual Meeting of the Association for Computational Linguistics,D\. Jurafsky, J\. Chai, N\. Schluter, and J\. Tetreault \(Eds\.\),Online,pp\. 6282–6293\.External Links:[Link](https://aclanthology.org/2020.acl-main.560/),[Document](https://dx.doi.org/10.18653/v1/2020.acl-main.560)Cited by:[§1](https://arxiv.org/html/2610.13398#S1.p1.1),[§1](https://arxiv.org/html/2610.13398#S1.p2.1)\.
- Keet al\.\(2025\)W\. Ke, Y\. Liu, X\. Yang, J\. Wei, and Q\. HouAlign, enhance and read: Scene Tibetan text recognition with cross\-sequence reasoning\.Appl\. Soft Comput\.169,pp\. 112548\.External Links:[Link](https://doi.org/10.1016/j.asoc.2024.112548),[Document](https://dx.doi.org/10.1016/J.ASOC.2024.112548)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Khysruet al\.\(2019\)K\. Khysru, D\. Jin, and J\. DangMorphological Verb\-Aware Tibetan Language Model\.IEEE Access7,pp\. 72896–72904\.External Links:[Link](https://doi.org/10.1109/ACCESS.2019.2919328),[Document](https://dx.doi.org/10.1109/ACCESS.2019.2919328)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1)\.
- Khysruet al\.\(2021\)K\. Khysru, D\. Jin, Y\. Huang, H\. Feng, and J\. DangA Tibetan Language Model That Considers the Relationship Between Suffixes and Functional Words\.IEEE Signal Process\. Lett\.28,pp\. 459–463\.External Links:[Link](https://doi.org/10.1109/LSP.2021.3058896),[Document](https://dx.doi.org/10.1109/LSP.2021.3058896)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1)\.
- Khysruet al\.\(2025\)K\. Khysru, Yangzom, W\. Tang, and J\. WeiResearch on speech synthesis technology based on Tibetan rhythmic features\.Expert Syst\. Appl\.277,pp\. 127181\.External Links:[Link](https://doi.org/10.1016/j.eswa.2025.127181),[Document](https://dx.doi.org/10.1016/J.ESWA.2025.127181)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1)\.
- Konget al\.\(2024\)C\. Kong, X\. Lv, L\. Zhang, H\. Zhao, Z\. Cai, and Y\. ChenRPEPL: Tibetan Sentiment Analysis Based on Relative Position Encoding and Prompt Learning\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.23\(12\),pp\. 167:1–167:18\.External Links:[Link](https://doi.org/10.1145/3698575),[Document](https://dx.doi.org/10.1145/3698575)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.26](https://arxiv.org/html/2610.13398#A4.SS26.p1.1)\.
- Krishnaet al\.\(2021\)R\. Krishna, N\. Mu, and K\. KeutzerApplying Text Analytics to the Mind\-section Literature of the Tibetan Tradition of the Great Perfection\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.20\(2\),pp\. 21:1–21:32\.External Links:[Link](https://doi.org/10.1145/3392047),[Document](https://dx.doi.org/10.1145/3392047)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.8](https://arxiv.org/html/2610.13398#A3.SS8.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1),[§D\.21](https://arxiv.org/html/2610.13398#A4.SS21.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1)\.
- Kuanzhuoet al\.\(2020\)Z\. Kuanzhuo, L\. Lin, and Z\. WeinaSimpleNLG\-TI: adapting SimpleNLG to Tibetan\.InProceedings of the 13th International Conference on Natural Language Generation,B\. Davis, Y\. Graham, J\. Kelleher, and Y\. Sripada \(Eds\.\),Dublin, Ireland,pp\. 86–90\.External Links:[Link](https://aclanthology.org/2020.inlg-1.12/),[Document](https://dx.doi.org/10.18653/v1/2020.inlg-1.12)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.7](https://arxiv.org/html/2610.13398#A4.SS7.p1.1)\.
- Kumaret al\.\(2026\)R\. Kumar, D\. Raj, and S\. SahaBhashaSutra: a task\-centric unified survey of Indian NLP datasets, corpora, and resources\.InProceedings of the 64th Annual Meeting of the Association for Computational Linguistics \(Volume 1: Long Papers\),M\. Liakata, V\. P\. Moreira, J\. Zhang, and D\. Jurgens \(Eds\.\),San Diego, California, United States,pp\. 11998–12064\.External Links:[Link](https://aclanthology.org/2026.acl-long.551/),[Document](https://dx.doi.org/10.18653/v1/2026.acl-long.551),ISBN 979\-8\-89176\-390\-6Cited by:[§2\.2](https://arxiv.org/html/2610.13398#S2.SS2.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px2.p1.1)\.
- Laiet al\.\(2018\)W\. Lai, X\. Zhao, and W\. BaoTibetan\-Chinese neural machine translation based on syllable segmentation\.InProceedings of the AMTA 2018 Workshop on Technologies for MT of Low Resource Languages \(LoResMT 2018\),C\. Liu \(Ed\.\),Boston, MA,pp\. 21–29\.External Links:[Link](https://aclanthology.org/W18-2203/)Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.16](https://arxiv.org/html/2610.13398#A4.SS16.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Lhundrubet al\.\(2026\)G\. Lhundrub, Y\. Zhou, M\. Long, W\. Ke, G\. Chen, J\. Wang, and D\. ZhengTibSTR\-Former: Multi\-scale feature integration and attention mechanisms for Tibetan scene text recognition\.J\. Vis\. Commun\. Image Represent\.120,pp\. 104907\.External Links:[Link](https://doi.org/10.1016/j.jvcir.2026.104907),[Document](https://dx.doi.org/10.1016/J.JVCIR.2026.104907)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Liet al\.\(2024\)B\. Li, J\. Zhu, and D\. ZhaoTibetan Handwriting Recognition Method Based on Structural Re\-Parameterization ViT and Vertical Attention\.InPattern Recognition and Computer Vision \- 7th Chinese Conference, PRCV 2024, Urumqi, China, October 18\-20, 2024, Proceedings, Part VII,Z\. Lin, M\. Cheng, R\. He, K\. Ubul, W\. Silamu, H\. Zha, J\. Zhou, and C\. Liu \(Eds\.\),Lecture Notes in Computer Science, Vol\.15037,pp\. 156–169\.External Links:[Link](https://doi.org/10.1007/978-981-97-8511-7/_12),[Document](https://dx.doi.org/10.1007/978-981-97-8511-7%5F12)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Liet al\.\(2022a\)F\. Li, H\. Lv, L\. Duo, B\. Yong, and Q\. ZhouSentence Boundary Disambiguation for Tibetan Based on Attention Mechanism at the Syllable Level\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.21\(6\),pp\. 1–18\.External Links:[Link](https://doi.org/10.1145/3527663),[Document](https://dx.doi.org/10.1145/3527663)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1)\.
- Liet al\.\(2021\)Y\. Li, J\. Jiang, J\. Yangji, and N\. MaFinding Better Subwords for Tibetan Neural Machine Translation\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.20\(2\),pp\. 24:1–24:11\.External Links:[Link](https://doi.org/10.1145/3448216),[Document](https://dx.doi.org/10.1145/3448216)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.16](https://arxiv.org/html/2610.13398#A4.SS16.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Li and Yu \(2013\)Y\. Li and H\. YuStudy on Tibetan Word Segmentation as Syllable Tagging\.InNatural Language Processing and Chinese Computing \- Second CCF Conference, NLPCC 2013, Chongqing, China, November 15\-19, 2013, Proceedings,G\. Zhou, J\. Li, D\. Zhao, and Y\. Feng \(Eds\.\),Communications in Computer and Information Science, Vol\.400,pp\. 363–369\.External Links:[Link](https://doi.org/10.1007/978-3-642-41644-6/_34),[Document](https://dx.doi.org/10.1007/978-3-642-41644-6%5F34)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- Liet al\.\(2022b\)Y\. Li, X\. Li, Y\. Wang, H\. Lv, F\. Li, and L\. DuoCharacter\-based Joint Word Segmentation and Part\-of\-Speech Tagging for Tibetan Based on Deep Learning\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.21\(5\),pp\. 95:1–95:15\.External Links:[Link](https://doi.org/10.1145/3511600),[Document](https://dx.doi.org/10.1145/3511600)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- Liet al\.\(2026\)Y\. Li, X\. Cao, Y\. Sun, Q\. Minggad, A\. Ablikim, and J\. Q\. C\. WangDiversity in unity, theory in practice: hierarchical multitask benchmarks for Chinese minority languages\.InProceedings of the 64th Annual Meeting of the Association for Computational Linguistics \(Volume 1: Long Papers\),M\. Liakata, V\. P\. Moreira, J\. Zhang, and D\. Jurgens \(Eds\.\),San Diego, California, United States,pp\. 36353–36373\.External Links:[Link](https://aclanthology.org/2026.acl-long.1684/),[Document](https://dx.doi.org/10.18653/v1/2026.acl-long.1684),ISBN 979\-8\-89176\-390\-6Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§5\.2](https://arxiv.org/html/2610.13398#S5.SS2.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px2.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Liet al\.\(2022c\)Z\. Li, W\. Wang, Y\. Wang, and Q\. ZhangCharacter recognition of Tibetan Historical document in Uchen font: Dataset and bench mark\.J\. Comput\. Methods Sci\. Eng\.22\(5\),pp\. 1779–1794\.External Links:[Link](https://doi.org/10.3233/JCM-226167),[Document](https://dx.doi.org/10.3233/JCM-226167)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Liang and Zerong \(2026\)S\. Liang and Z\. ZerongThe tonogenesis continuum in Tibetan: a computational investigation\.InThe Proceedings for the 6th International Workshop on Computational Approaches to Language Change \(LChange’26\),N\. Tahmasebi, P\. Cassotti, S\. Montariol, A\. Kutuzov, N\. Huebscher, E\. Spaziani, and N\. Baes \(Eds\.\),Rabat, Morocco,pp\. 27–34\.External Links:[Link](https://aclanthology.org/2026.lchange-1.3/),[Document](https://dx.doi.org/10.18653/v1/2026.lchange-1.3),ISBN 979\-8\-89176\-362\-3Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.8](https://arxiv.org/html/2610.13398#A3.SS8.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1)\.
- Lianget al\.\(2024\)Y\. Liang, H\. Lv, Y\. Li, L\. Duo, C\. Liu, and Q\. ZhouTibetan\-BERT\-wwm: A Tibetan Pretrained Model With Whole Word Masking for Text Classification\.IEEE Trans\. Comput\. Soc\. Syst\.11\(5\),pp\. 6268–6277\.External Links:[Link](https://doi.org/10.1109/TCSS.2024.3374633),[Document](https://dx.doi.org/10.1109/TCSS.2024.3374633)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1)\.
- Liuet al\.\(2011\)H\. Liu, M\. Nuo, L\. Ma, J\. Wu, and Y\. HeTibetan word segmentation as syllable tagging using conditional random field\.InProceedings of the 25th Pacific Asia Conference on Language, Information and Computation,H\. H\. Gao and M\. Dong \(Eds\.\),Singapore,pp\. 168–177\.External Links:[Link](https://aclanthology.org/Y11-1018/)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- Liuet al\.\(2012\)H\. Liu, M\. Nuo, J\. Wu, and Y\. HeBuilding large scale text corpus for Tibetan natural language processing by extracting text from web pages\.InProceedings of the 10th Workshop on Asian Language Resources,R\. Weerasinghe, S\. Hussain, V\. Sornlertlamvanich, and R\. E\. O\. Roxas \(Eds\.\),Mumbai, India,pp\. 11–20\.External Links:[Link](https://aclanthology.org/W12-5202/)Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- Liuet al\.\(2014\)H\. Liu, M\. Nuo, and J\. WuZipf’s law and statistical data on Modern Tibetan\.InProceedings of COLING 2014, the 25th International Conference on Computational Linguistics: Technical Papers,J\. Tsujii and J\. Hajic \(Eds\.\),Dublin, Ireland,pp\. 322–333\.External Links:[Link](https://aclanthology.org/C14-1032/)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.14](https://arxiv.org/html/2610.13398#A4.SS14.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1)\.
- Liuet al\.\(2010\)H\. Liu, W\. Zhao, M\. Nuo, L\. Jiang, J\. Wu, and Y\. HeTibetan number identification based on classification of number components in Tibetan word segmentation\.InColing 2010: Posters,C\. Huang and D\. Jurafsky \(Eds\.\),Beijing, China,pp\. 719–724\.External Links:[Link](https://aclanthology.org/C10-2082/)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1)\.
- Liu and Best \(2025\)S\. Liu and M\. BestA survey of NLP progress in Sino\-Tibetan low\-resource languages\.InProceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies \(Volume 1: Long Papers\),L\. Chiruzzo, A\. Ritter, and L\. Wang \(Eds\.\),Albuquerque, New Mexico,pp\. 7804–7825\.External Links:[Link](https://aclanthology.org/2025.naacl-long.396/),[Document](https://dx.doi.org/10.18653/v1/2025.naacl-long.396),ISBN 979\-8\-89176\-189\-6Cited by:[§1](https://arxiv.org/html/2610.13398#S1.p3.1),[Table 1](https://arxiv.org/html/2610.13398#S2.T1.p1.1.2.1.2.1),[§5\.1](https://arxiv.org/html/2610.13398#S5.SS1.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px1.p1.1)\.
- Liuet al\.\(2022\)S\. Liu, J\. Deng, Y\. Sun, and X\. ZhaoTiBERT: Tibetan Pre\-trained Language Model\.InIEEE International Conference on Systems, Man, and Cybernetics, SMC 2022, Prague, Czech Republic, October 9\-12, 2022,pp\. 2956–2961\.External Links:[Link](https://doi.org/10.1109/SMC53654.2022.9945074),[Document](https://dx.doi.org/10.1109/SMC53654.2022.9945074)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[Table 2](https://arxiv.org/html/2610.13398#S4.T2),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Liuet al\.\(2024\)Y\. Liu, Y\. Zhao, X\. Xu, L\. Xu, X\. Zhang, and Q\. JiExploring task\-diverse meta\-learning on Tibetan multi\-dialect speech recognition\.EURASIP J\. Audio Speech Music\. Process\.2024\(1\),pp\. 37\.External Links:[Link](https://doi.org/10.1186/s13636-024-00361-7),[Document](https://dx.doi.org/10.1186/S13636-024-00361-7)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px5.p1.1)\.
- Liuet al\.\(2023\)Y\. Liu, Y\. Zhao, X\. Xu, L\. Xu, and X\. ZhangMulti\-task Pre\-training for Lhasa\-Tibetan Speech Recognition\.InArtificial Neural Networks and Machine Learning \- ICANN 2023 \- 32nd International Conference on Artificial Neural Networks, Heraklion, Crete, Greece, September 26\-29, 2023, Proceedings, Part IX,L\. Iliadis, A\. Papaleonidas, P\. Angelov, and C\. Jayne \(Eds\.\),Lecture Notes in Computer Science, Vol\.14262,pp\. 78–90\.External Links:[Link](https://doi.org/10.1007/978-3-031-44201-8/_7),[Document](https://dx.doi.org/10.1007/978-3-031-44201-8%5F7)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Long and An \(2023\)C\. Long and B\. AnOn the development of interlinearized ancient literature of ethnic minorities: a case study of the interlinearization of ancient written Tibetan literature\.InProceedings of the Ancient Language Processing Workshop,A\. Anderson, S\. Gordin, B\. Li, Y\. Liu, and M\. C\. Passarotti \(Eds\.\),Varna, Bulgaria,pp\. 222–231\.External Links:[Link](https://aclanthology.org/2023.alp-1.25/)Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.21](https://arxiv.org/html/2610.13398#A4.SS21.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Long and Hill \(2021\)C\. Long and N\. W\. HillRecent Developments in Tibetan NLP\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.20\(2\),pp\. 19:1–19:3\.External Links:[Link](https://doi.org/10.1145/3453692),[Document](https://dx.doi.org/10.1145/3453692)Cited by:[§1](https://arxiv.org/html/2610.13398#S1.p3.1),[Table 1](https://arxiv.org/html/2610.13398#S2.T1.p2.1.2.1.2.1),[§4\.1](https://arxiv.org/html/2610.13398#S4.SS1.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p3.1),[§5\.1](https://arxiv.org/html/2610.13398#S5.SS1.p1.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Longet al\.\(2021\)C\. Long, X\. Zhou, and M\. ZhouRecognition of Tibetan Maximal\-length Noun Phrases Based on Syntax Tree\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.20\(2\),pp\. 25:1–25:13\.External Links:[Link](https://doi.org/10.1145/3423324),[Document](https://dx.doi.org/10.1145/3423324)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1)\.
- Luo and Chuang \(2024\)Q\. Luo and Y\. ChuangCleansing Jewel: A Neural Spelling Correction Model Built On Google OCR\-ed Tibetan Manuscripts\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.23\(5\),pp\. 73\.External Links:[Link](https://doi.org/10.1145/3654811),[Document](https://dx.doi.org/10.1145/3654811)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.8](https://arxiv.org/html/2610.13398#A3.SS8.p1.1),[§D\.21](https://arxiv.org/html/2610.13398#A4.SS21.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Lvet al\.\(2024\)H\. Lv, H\. Lv, L\. Yang, J\. Shen, L\. Duo, Y\. Li, Q\. Zhou, and B\. YongImproved Tibetan Word Vectors Models Based on Position Information Fusion\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.23\(11\),pp\. 153:1–153:21\.External Links:[Link](https://doi.org/10.1145/3681787),[Document](https://dx.doi.org/10.1145/3681787)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§D\.25](https://arxiv.org/html/2610.13398#A4.SS25.p1.1)\.
- Lvet al\.\(2025\)H\. Lv, C\. Pu, L\. Duo, Y\. Li, Q\. Zhou, and J\. ShenT\-LLaMA: a Tibetan large language model based on LLaMA2\.Complex Intell\. Syst\.11\(1\)\.External Links:[Link](https://doi.org/10.1007/s40747-024-01641-7),[Document](https://dx.doi.org/10.1007/S40747-024-01641-7)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1)\.
- Maet al\.\(2011\)L\. Ma, H\. Liu, and J\. WuMRG\-OHTC Database for Online Handwritten Tibetan Character Recognition\.In2011 International Conference on Document Analysis and Recognition, ICDAR 2011, Beijing, China, September 18\-21, 2011,pp\. 207–211\.External Links:[Link](https://doi.org/10.1109/ICDAR.2011.50),[Document](https://dx.doi.org/10.1109/ICDAR.2011.50)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Maet al\.\(2020\)L\. Ma, C\. Long, L\. Duan, X\. Zhang, Y\. Li, and Q\. ZhaoSegmentation and Recognition for Historical Tibetan Document Images\.IEEE Access8,pp\. 52641–52651\.External Links:[Link](https://doi.org/10.1109/ACCESS.2020.2975023),[Document](https://dx.doi.org/10.1109/ACCESS.2020.2975023)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.21](https://arxiv.org/html/2610.13398#A4.SS21.p1.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Ma and Wu \(2013\)L\. Ma and J\. WuSemi\-automatic Tibetan Component Annotation from Online Handwritten Tibetan Character Database by Optimizing Segmentation Hypotheses\.In12th International Conference on Document Analysis and Recognition, ICDAR 2013, Washington, DC, USA, August 25\-28, 2013,pp\. 1340–1344\.External Links:[Link](https://doi.org/10.1109/ICDAR.2013.271),[Document](https://dx.doi.org/10.1109/ICDAR.2013.271)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Ma and Wu \(2015\)L\. Ma and J\. WuOnline handwritten Tibetan syllable recognition based on component segmentation method\.In13th International Conference on Document Analysis and Recognition, ICDAR 2015, Nancy, France, August 23\-26, 2015,pp\. 46–50\.External Links:[Link](https://doi.org/10.1109/ICDAR.2015.7333723),[Document](https://dx.doi.org/10.1109/ICDAR.2015.7333723)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Meelenet al\.\(2021\)M\. Meelen, É\. Roux, and N\. HillOptimisation of the Largest Annotated Tibetan Corpus Combining Rule\-based, Memory\-based, and Deep\-learning Methods\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.20\(1\),pp\. 7:1–7:11\.External Links:[Link](https://doi.org/10.1145/3409488),[Document](https://dx.doi.org/10.1145/3409488)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.7](https://arxiv.org/html/2610.13398#A3.SS7.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px1.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1)\.
- Meelen and Roux \(2020\)M\. Meelen and É\. RouxMeta\-dating the PArsed corpus of Tibetan \(PACTib\)\.InProceedings of the 19th International Workshop on Treebanks and Linguistic Theories,K\. Evang, L\. Kallmeyer, R\. Ehren, S\. Petitjean, E\. Seyffarth, and D\. Seddah \(Eds\.\),Düsseldorf, Germany,pp\. 31–42\.External Links:[Link](https://aclanthology.org/2020.tlt-1.3/),[Document](https://dx.doi.org/10.18653/v1/2020.tlt-1.3)Cited by:[§A\.2](https://arxiv.org/html/2610.13398#A1.SS2.p1.1),[§C\.7](https://arxiv.org/html/2610.13398#A3.SS7.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px1.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Mooreet al\.\(2026\)J\. Moore, S\. Li, and P\. LaurenEvaluating Tibetan ASR With Segmented Word Error Rate: Beyond Character\-Level Metrics\.IEEE Access14,pp\. 101790–101805\.External Links:[Link](https://doi.org/10.1109/ACCESS.2026.3709206),[Document](https://dx.doi.org/10.1109/ACCESS.2026.3709206)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.8](https://arxiv.org/html/2610.13398#A3.SS8.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Nehrdich \(2022\)S\. NehrdichSansTib, a Sanskrit \- Tibetan parallel corpus and bilingual sentence embedding model\.InProceedings of the Thirteenth Language Resources and Evaluation Conference,N\. Calzolari, F\. Béchet, P\. Blache, K\. Choukri, C\. Cieri, T\. Declerck, S\. Goggi, H\. Isahara, B\. Maegaard, J\. Mariani, H\. Mazo, J\. Odijk, and S\. Piperidis \(Eds\.\),Marseille, France,pp\. 6728–6734\.External Links:[Link](https://aclanthology.org/2022.lrec-1.724/)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.4](https://arxiv.org/html/2610.13398#A3.SS4.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1)\.
- Nuoet al\.\(2012\)M\. H\. Nuo, H\. D\. Liu, W\. N\. Zhao, L\. L\. Ma, J\. Wu, and Z\. M\. DingTibetan base noun phrase identification framework based on Chinese\-Tibetan sentence aligned corpus\.InProceedings of COLING 2012,M\. Kay and C\. Boitet \(Eds\.\),Mumbai, India,pp\. 2141–2158\.External Links:[Link](https://aclanthology.org/C12-1131/)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1)\.
- Nuoet al\.\(2015\)M\. Nuo, H\. Liu, C\. Long, and J\. WuTibetan unknown word identification from news corpora for supporting lexicon\-based Tibetan word segmentation\.InProceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing \(Volume 2: Short Papers\),C\. Zong and M\. Strube \(Eds\.\),Beijing, China,pp\. 451–457\.External Links:[Link](https://aclanthology.org/P15-2074/),[Document](https://dx.doi.org/10.3115/v1/P15-2074)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- Nuoet al\.\(2016\)M\. Nuo, C\. Lun, and H\. LiuTibetan Multi\-word Expressions Identification Framework Based on News Corpora\.InNatural Language Understanding and Intelligent Applications \- 5th CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2016, and 24th International Conference on Computer Processing of Oriental Languages, ICCPOL 2016, Kunming, China, December 2\-6, 2016, Proceedings,C\. Lin, N\. Xue, D\. Zhao, X\. Huang, and Y\. Feng \(Eds\.\),Lecture Notes in Computer Science, Vol\.10102,pp\. 16–26\.External Links:[Link](https://doi.org/10.1007/978-3-319-50496-4/_2),[Document](https://dx.doi.org/10.1007/978-3-319-50496-4%5F2)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.25](https://arxiv.org/html/2610.13398#A4.SS25.p1.1)\.
- Ouyanget al\.\(2024\)X\. Ouyang, X\. Yan, and M\. HaoA Pre\-Trained Language Model Based on LED for Tibetan Long Text Summarization\.In27th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2024, Tianjin, China, May 8\-10, 2024,W\. Shen, J\. A\. Barthès, J\. Luo, T\. Qiu, X\. Zhou, J\. Zhang, H\. Zhu, K\. Peng, T\. Xu, and N\. Chen \(Eds\.\),pp\. 992–997\.External Links:[Link](https://doi.org/10.1109/CSCWD61410.2024.10580451),[Document](https://dx.doi.org/10.1109/CSCWD61410.2024.10580451)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.28](https://arxiv.org/html/2610.13398#A4.SS28.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Qinet al\.\(2022a\)S\. Qin, L\. Wang, S\. Li, J\. Dang, and L\. PanImproving low\-resource Tibetan end\-to\-end ASR by multilingual and multilevel unit modeling\.EURASIP J\. Audio Speech Music\. Process\.2022\(1\),pp\. 2\.External Links:[Link](https://doi.org/10.1186/s13636-021-00233-4),[Document](https://dx.doi.org/10.1186/S13636-021-00233-4)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Qinet al\.\(2022b\)S\. Qin, L\. Wang, S\. Li, Y\. Lin, and J\. DangFiner\-grained Modeling units\-based Meta\-Learning for Low\-resource Tibetan Speech Recognition\.In23rd Annual Conference of the International Speech Communication Association, Interspeech 2022, Incheon, Korea, September 18\-22, 2022,H\. Ko and J\. H\. L\. Hansen \(Eds\.\),pp\. 2133–2137\.External Links:[Link](https://doi.org/10.21437/Interspeech.2022-10015),[Document](https://dx.doi.org/10.21437/INTERSPEECH.2022-10015)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Qiuet al\.\(2018a\)L\. Qiu, Q\. Lei, and Z\. ZhangAdvanced Sentiment Classification of Tibetan Microblogs on Smart Campuses Based on Multi\-Feature Fusion\.IEEE Access6,pp\. 17896–17904\.External Links:[Link](https://doi.org/10.1109/ACCESS.2018.2820163),[Document](https://dx.doi.org/10.1109/ACCESS.2018.2820163)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1),[§D\.26](https://arxiv.org/html/2610.13398#A4.SS26.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Qiuet al\.\(2018b\)L\. Qiu, J\. Yu, J\. Li, and H\. JiaTibetan Weibo User Group Division Based on Semantic Information in the Era of Big Data\.Complex\.2018,pp\. 5809043:1–5809043:11\.External Links:[Link](https://doi.org/10.1155/2018/5809043),[Document](https://dx.doi.org/10.1155/2018/5809043)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Qiuet al\.\(2017\)L\. Qiu, H\. Zhang, Z\. Zhang, and Q\. PuTibetan Microblog Emotional Analysis Based on Sequential Model in Online Social Platforms\.Complex\.2017,pp\. 5342601:1–5342601:8\.External Links:[Link](https://doi.org/10.1155/2017/5342601),[Document](https://dx.doi.org/10.1155/2017/5342601)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1),[§D\.26](https://arxiv.org/html/2610.13398#A4.SS26.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Renet al\.\(2024\)H\. Ren, J\. Yang, S\. Ni, Q\. Yang, J\. Zhang, and N\. QunA Method for Tibetan Offensive Language Detection Based on Prompt Learning and Information Theory\.InInternational Joint Conference on Neural Networks, IJCNN 2024, Yokohama, Japan, June 30 \- July 5, 2024,pp\. 1–8\.External Links:[Link](https://doi.org/10.1109/IJCNN60899.2024.10651369),[Document](https://dx.doi.org/10.1109/IJCNN60899.2024.10651369)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Renet al\.\(2026\)Z\. Ren, L\. Zhu, X\. Yang, and X\. ZhaoResearch on the Quality Evaluation of Amdo Tibetan\-Chinese Speech Translation Dataset\.Data Intell\.8\(2\),pp\. 20250172\.External Links:[Link](https://doi.org/10.3724/2096-7004.di.2025.0172),[Document](https://dx.doi.org/10.3724/2096-7004.DI.2025.0172)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.16](https://arxiv.org/html/2610.13398#A4.SS16.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Rohatgiet al\.\(2023\)S\. Rohatgi, Y\. Qin, B\. Aw, N\. Unnithan, and M\. KanThe ACL OCL corpus: advancing open science in computational linguistics\.InProceedings of the 2023 Conference on Empirical Methods in Natural Language Processing,H\. Bouamor, J\. Pino, and K\. Bali \(Eds\.\),Singapore,pp\. 10348–10361\.External Links:[Link](https://aclanthology.org/2023.emnlp-main.640/),[Document](https://dx.doi.org/10.18653/v1/2023.emnlp-main.640)Cited by:[Table 1](https://arxiv.org/html/2610.13398#S2.T1.2.3.3.1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px6.p1.1)\.
- Schmidt \(2020\)D\. SchmidtGrading Tibetan Children’s Literature: A Test Case Using the NLP Readability Tool “Dakje”\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.19\(6\),pp\. 75:1–75:19\.External Links:[Link](https://doi.org/10.1145/3392046),[Document](https://dx.doi.org/10.1145/3392046)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.5](https://arxiv.org/html/2610.13398#A3.SS5.p1.1),[§D\.21](https://arxiv.org/html/2610.13398#A4.SS21.p1.1),[§D\.8](https://arxiv.org/html/2610.13398#A4.SS8.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px5.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px3.p1.1)\.
- Shaet al\.\(2025\)J\. Sha, M\. Zhu, C\. Feng, and J\. CiTibetan\-LLaMA 2: Large Language Model for Tibetan\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.24\(12\),pp\. 141:1–141:19\.External Links:[Link](https://doi.org/10.1145/3776748),[Document](https://dx.doi.org/10.1145/3776748)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1)\.
- Shenet al\.\(2017\)T\. Shen, L\. Wang, X\. Chen, K\. Khysru, and J\. DangExploiting the Tibetan Radicals in Recurrent Neural Network for Low\-Resource Language Models\.InNeural Information Processing \- 24th International Conference, ICONIP 2017, Guangzhou, China, November 14\-18, 2017, Proceedings, Part II,D\. Liu, S\. Xie, Y\. Li, D\. Zhao, and E\. M\. El\-Alfy \(Eds\.\),Lecture Notes in Computer Science, Vol\.10635,pp\. 266–275\.External Links:[Link](https://doi.org/10.1007/978-3-319-70096-0/_28),[Document](https://dx.doi.org/10.1007/978-3-319-70096-0%5F28)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1)\.
- Shiet al\.\(2021\)S\. Shi, D\. Luo, X\. Wu, C\. Long, and H\. HuangMulti\-level Chunk\-based Constituent\-to\-Dependency Treebank Transformation for Tibetan Dependency Parsing\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.20\(2\),pp\. 26:1–26:12\.External Links:[Link](https://doi.org/10.1145/3424247),[Document](https://dx.doi.org/10.1145/3424247)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1)\.
- Suet al\.\(2013\)B\. Su, X\. Ding, L\. Peng, and C\. LiuCross\-Language Sensitive Words Distribution Map: A Novel Recognition\-Based Document Understanding Method for Uighur and Tibetan\.In12th International Conference on Document Analysis and Recognition, ICDAR 2013, Washington, DC, USA, August 25\-28, 2013,pp\. 255–259\.External Links:[Link](https://doi.org/10.1109/ICDAR.2013.58),[Document](https://dx.doi.org/10.1109/ICDAR.2013.58)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1)\.
- Suet al\.\(2026\)Z\. Su, Z\. Zhang, Z\. Liu, X\. Song, Z\. Xu, L\. Zheng, X\. Zhang, R\. Fu, G\. Xu, and W\. ZhangReinforcement learning with semantic rewards enables low\-resource language expansion without alignment tax\.InFindings of the Association for Computational Linguistics: ACL 2026,M\. Liakata, V\. P\. Moreira, J\. Zhang, and D\. Jurgens \(Eds\.\),San Diego, California, United States,pp\. 17772–17786\.External Links:[Link](https://aclanthology.org/2026.findings-acl.880/),[Document](https://dx.doi.org/10.18653/v1/2026.findings-acl.880),ISBN 979\-8\-89176\-395\-1Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1)\.
- Sunet al\.\(2021\)Y\. Sun, A\. Chen, C\. Chen, T\. Xia, and X\. ZhaoA Joint Model for Representation Learning of Tibetan Knowledge Graph Based on Encyclopedia\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.20\(2\),pp\. 27:1–27:17\.External Links:[Link](https://doi.org/10.1145/3447248),[Document](https://dx.doi.org/10.1145/3447248)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.20](https://arxiv.org/html/2610.13398#A4.SS20.p1.1),[§D\.9](https://arxiv.org/html/2610.13398#A4.SS9.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1)\.
- Sunet al\.\(2019a\)Y\. Sun, C\. Chen, T\. Xia, and X\. ZhaoQuGAN: Quasi Generative Adversarial Network for Tibetan Question Answering Corpus Generation\.IEEE Access7,pp\. 116247–116255\.External Links:[Link](https://doi.org/10.1109/ACCESS.2019.2934581),[Document](https://dx.doi.org/10.1109/ACCESS.2019.2934581)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.23](https://arxiv.org/html/2610.13398#A4.SS23.p1.1),[§D\.7](https://arxiv.org/html/2610.13398#A4.SS7.p1.1)\.
- Sunet al\.\(2022\)Y\. Sun, S\. Liu, Z\. Dan, and X\. ZhaoQuestion generation based on grammar knowledge and fine\-grained classification\.InProceedings of the 29th International Conference on Computational Linguistics,N\. Calzolari, C\. Huang, H\. Kim, J\. Pustejovsky, L\. Wanner, K\. Choi, P\. Ryu, H\. Chen, L\. Donatelli, H\. Ji, S\. Kurohashi, P\. Paggio, N\. Xue, S\. Kim, Y\. Hahm, Z\. He, T\. K\. Lee, E\. Santus, F\. Bond, and S\. Na \(Eds\.\),Gyeongju, Republic of Korea,pp\. 6457–6467\.External Links:[Link](https://aclanthology.org/2022.coling-1.562/)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.23](https://arxiv.org/html/2610.13398#A4.SS23.p1.1),[§D\.7](https://arxiv.org/html/2610.13398#A4.SS7.p1.1)\.
- Sunet al\.\(2019b\)Y\. Sun, L\. Wang, C\. Chen, T\. Xia, and X\. ZhaoImproved Distant Supervised Model in Tibetan Relation Extraction Using ELMo and Attention\.IEEE Access7,pp\. 173054–173062\.External Links:[Link](https://doi.org/10.1109/ACCESS.2019.2955977),[Document](https://dx.doi.org/10.1109/ACCESS.2019.2955977)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.9](https://arxiv.org/html/2610.13398#A4.SS9.p1.1)\.
- Sun and Xia \(2019\)Y\. Sun and T\. XiaA Hybrid Network Model for Tibetan Question Answering\.IEEE Access7,pp\. 52769–52777\.External Links:[Link](https://doi.org/10.1109/ACCESS.2019.2911320),[Document](https://dx.doi.org/10.1109/ACCESS.2019.2911320)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.23](https://arxiv.org/html/2610.13398#A4.SS23.p1.1)\.
- Tanget al\.\(2025\)C\. Tang, Z\. Tan, and X\. ZhaoTibetan Text Classification based on Prompt Learning and Ensemble Learning\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.24\(2\),pp\. 19:1–19:17\.External Links:[Link](https://doi.org/10.1145/3711827),[Document](https://dx.doi.org/10.1145/3711827)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Tashiet al\.\(2026\)D\. Tashi, B\. Chen, T\. Sheng, Y\. Yu, X\. Wang, J\. Zhang, L\. Yeshi, R\. Dongrub, T\. Tsering, and N\. TashiTibetan Data Augmentation via GAN\-Based Handwritten Text Generation\.CAAI Trans\. Intell\. Technol\.11\(1\),pp\. 55–65\.External Links:[Link](https://doi.org/10.1049/cit2.70078),[Document](https://dx.doi.org/10.1049/CIT2.70078)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1)\.
- Tashiet al\.\(2024\)D\. Tashi, T\. Sheng, B\. Chen, R\. Duojie, R\. Dongrub, Y\. Yu, X\. Wang, and N\. TashiConstruction of a Tibetan Handwriting Khyug\-yig Dataset\.Data Intell\.6\(3\),pp\. 870–887\.External Links:[Link](https://doi.org/10.3724/2096-7004.di.2024.0048),[Document](https://dx.doi.org/10.3724/2096-7004.DI.2024.0048)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px3.p1.1)\.
- Tonjaet al\.\(2024\)A\. Tonja, F\. Balouchzahi, S\. Butt, O\. Kolesnikova, H\. Ceballos, A\. Gelbukh, and T\. SolorioNLP progress in indigenous Latin American languages\.InProceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies \(Volume 1: Long Papers\),K\. Duh, H\. Gomez, and S\. Bethard \(Eds\.\),Mexico City, Mexico,pp\. 6972–6987\.External Links:[Link](https://aclanthology.org/2024.naacl-long.385/),[Document](https://dx.doi.org/10.18653/v1/2024.naacl-long.385)Cited by:[§2\.2](https://arxiv.org/html/2610.13398#S2.SS2.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px3.p1.1)\.
- Trinleyet al\.\(2021\)N\. Trinley, Tenzin, D\. Schmidt, H\. Hildt, and T\. KaldanTaming the Wild Etext: Managing, Annotating, and Sharing Tibetan Corpora in Open Spaces\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.20\(2\),pp\. 22:1–22:23\.External Links:[Link](https://doi.org/10.1145/3418060),[Document](https://dx.doi.org/10.1145/3418060)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.3](https://arxiv.org/html/2610.13398#A3.SS3.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.3](https://arxiv.org/html/2610.13398#S4.SS3.p2.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px1.p1.1)\.
- Wagner and Zeisler \(2004\)A\. Wagner and B\. ZeislerA syntactically annotated corpus of Tibetan\.InProceedings of the Fourth International Conference on Language Resources and Evaluation \(LREC’04\),M\. T\. Lino, M\. F\. Xavier, F\. Ferreira, R\. Costa, and R\. Silva \(Eds\.\),Lisbon, Portugal\.External Links:[Link](https://aclanthology.org/L04-1156/)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.4](https://arxiv.org/html/2610.13398#A3.SS4.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.29](https://arxiv.org/html/2610.13398#A4.SS29.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- Wanget al\.\(2020a\)H\. Wang, F\. Gao, Y\. Zhao, L\. Yang, J\. Yue, and H\. MaMultitask Learning with Local Attention for Tibetan Speech Recognition\.Complex\.2020,pp\. 8894566:1–8894566:10\.External Links:[Link](https://doi.org/10.1155/2020/8894566),[Document](https://dx.doi.org/10.1155/2020/8894566)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1)\.
- Wanget al\.\(2026\)Q\. Wang, Y\. Wang, M\. Amini, M\. Xian, and Z\. FuQuality assessment of Tibetan\-Chinese poetry translation: Integrating automated metrics and qualitative insights through a cross\-system comparison of dedicated NMT engines and a prompted LLM\.Lang\. Resour\. Evaluation60\(2\),pp\. 42\.External Links:[Link](https://doi.org/10.1007/s10579-026-09918-8),[Document](https://dx.doi.org/10.1007/S10579-026-09918-8)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.16](https://arxiv.org/html/2610.13398#A4.SS16.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Wanget al\.\(2019\)W\. Wang, Z\. Li, Z\. Cai, X\. Lv, C\. Zhaxi, and Y\. HanOnline Tibetan Handwriting Recognition for Large Character Set on New Databases\.Int\. J\. Pattern Recognit\. Artif\. Intell\.33\(10\),pp\. 1953003:1–1953003:21\.External Links:[Link](https://doi.org/10.1142/S0218001419530033),[Document](https://dx.doi.org/10.1142/S0218001419530033)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Wanget al\.\(2018a\)X\. Wang, W\. Wang, Z\. Li, Y\. Wang, Y\. Han, and Z\. HaoA Recognition Method of the Similarity Character for Uchen Script Tibetan Historical Document Based on DNN\.InPattern Recognition and Computer Vision \- First Chinese Conference, PRCV 2018, Guangzhou, China, November 23\-26, 2018, Proceedings, Part III,J\. Lai, C\. Liu, X\. Chen, J\. Zhou, T\. Tan, N\. Zheng, and H\. Zha \(Eds\.\),Lecture Notes in Computer Science, Vol\.11258,pp\. 52–62\.External Links:[Link](https://doi.org/10.1007/978-3-030-03338-5/_5),[Document](https://dx.doi.org/10.1007/978-3-030-03338-5%5F5)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.21](https://arxiv.org/html/2610.13398#A4.SS21.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§5\.4](https://arxiv.org/html/2610.13398#S5.SS4.p1.1)\.
- Wanget al\.\(2018b\)Y\. Wang, W\. Wang, Z\. Li, Y\. Han, and X\. WangResearch on Text Line Segmentation of Historical Tibetan Documents Based on the Connected Component Analysis\.InPattern Recognition and Computer Vision \- First Chinese Conference, PRCV 2018, Guangzhou, China, November 23\-26, 2018, Proceedings, Part III,J\. Lai, C\. Liu, X\. Chen, J\. Zhou, T\. Tan, N\. Zheng, and H\. Zha \(Eds\.\),Lecture Notes in Computer Science, Vol\.11258,pp\. 74–87\.External Links:[Link](https://doi.org/10.1007/978-3-030-03338-5/_7),[Document](https://dx.doi.org/10.1007/978-3-030-03338-5%5F7)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.21](https://arxiv.org/html/2610.13398#A4.SS21.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Wanget al\.\(2025\)Y\. Wang, Y\. Hao, B\. Chen, Y\. Xu, and X\. ZhaoDiSG: A Discourse Structure\-Aware Multi\-stage Approach for Long Tibetan Text Summarization\.InNatural Language Processing and Chinese Computing \- 14th National CCF Conference, NLPCC 2025, Urumqi, China, August 7\-9, 2025, Proceedings, Part IV,X\. Mao, Z\. Ren, and M\. Yang \(Eds\.\),Lecture Notes in Computer Science, Vol\.16105,pp\. 159–171\.External Links:[Link](https://doi.org/10.1007/978-981-95-3352-7/_13),[Document](https://dx.doi.org/10.1007/978-981-95-3352-7%5F13)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.28](https://arxiv.org/html/2610.13398#A4.SS28.p1.1),[§D\.4](https://arxiv.org/html/2610.13398#A4.SS4.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Wanget al\.\(2020b\)Z\. Wang, W\. Fang, X\. Zhao, W\. Song, Y\. Feng, and Y\. ChangTransliteration recognition of Tibetan person name based on Tibetan cultural knowledge\.Int\. J\. Comput\. Sci\. Eng\.22\(2/3\),pp\. 305–312\.External Links:[Link](https://doi.org/10.1504/IJCSE.2020.107352),[Document](https://dx.doi.org/10.1504/IJCSE.2020.107352)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.9](https://arxiv.org/html/2610.13398#A4.SS9.p1.1)\.
- Wuet al\.\(2026\)S\. Wu, Y\. Zhao, J\. Yu, X\. Xu, and H\. LiEfficient and Natural Tibetan Speech Synthesis via Gaussian Noise\-Improved Monotonic Alignment Search and Lightweight iSTFT Multiband Decoder\.Data Intell\.8\(2\),pp\. 20250128\.External Links:[Link](https://doi.org/10.3724/2096-7004.di.2025.0128),[Document](https://dx.doi.org/10.3724/2096-7004.DI.2025.0128)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1)\.
- Xu and Gao \(2024\)G\. Xu and W\. GaoTibetan microblogging user data analysis and topic identification\.In23rd IEEE International Conference on Trust, Security and Privacy in Computing and Communications, TrustCom 2024, Sanya, China, December 17\-21, 2024,pp\. 1077–1082\.External Links:[Link](https://doi.org/10.1109/TrustCom63139.2024.00154),[Document](https://dx.doi.org/10.1109/TRUSTCOM63139.2024.00154)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.10](https://arxiv.org/html/2610.13398#A4.SS10.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Xuet al\.\(2026\)G\. Xu, Y\. Liang, Z\. Su, X\. Song, Z\. Zhang, Y\. Dong, T\. Zhang, and X\. HanFTibSuite: a comprehensive resource suite for Tibetan vision–language modeling\.InFindings of the Association for Computational Linguistics: ACL 2026,M\. Liakata, V\. P\. Moreira, J\. Zhang, and D\. Jurgens \(Eds\.\),San Diego, California, United States,pp\. 18143–18159\.External Links:[Link](https://aclanthology.org/2026.findings-acl.903/),[Document](https://dx.doi.org/10.18653/v1/2026.findings-acl.903),ISBN 979\-8\-89176\-395\-1Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Xuet al\.\(2025\)G\. Xu, Z\. Su, Z\. Zhang, J\. Liu, X\. Han, T\. Zhang, and Y\. DongCMHG: a dataset and benchmark for headline generation of minority languages in China\.InProceedings of the 2025 Conference on Empirical Methods in Natural Language Processing,C\. Christodoulopoulos, T\. Chakraborty, C\. Rose, and V\. Peng \(Eds\.\),Suzhou, China,pp\. 12339–12346\.External Links:[Link](https://aclanthology.org/2025.emnlp-main.622/),[Document](https://dx.doi.org/10.18653/v1/2025.emnlp-main.622),ISBN 979\-8\-89176\-332\-6Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.7](https://arxiv.org/html/2610.13398#A4.SS7.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Xuet al\.\(2017\)G\. Xu, C\. Wang, L\. Wang, Y\. Zhou, W\. Li, H\. Xu, and Q\. HuangSemantic classification method for network Tibetan corpus\.Clust\. Comput\.20\(1\),pp\. 155–165\.External Links:[Link](https://doi.org/10.1007/s10586-017-0742-6),[Document](https://dx.doi.org/10.1007/S10586-017-0742-6)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.10](https://arxiv.org/html/2610.13398#A4.SS10.p1.1),[§D\.25](https://arxiv.org/html/2610.13398#A4.SS25.p1.1)\.
- Xuet al\.\(2019a\)G\. Xu, C\. Wang, H\. Yao, and Q\. QiResearch on Tibetan hot words, sensitive words tracking and public opinion classification\.Clust\. Comput\.22\(Supplement\),pp\. 9977–9990\.External Links:[Link](https://doi.org/10.1007/s10586-017-1026-x),[Document](https://dx.doi.org/10.1007/S10586-017-1026-X)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.10](https://arxiv.org/html/2610.13398#A4.SS10.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1)\.
- Xuet al\.\(2019b\)G\. Xu, H\. Yao, D\. Wu, Y\. Li, D\. Ouyang, and G\. ChenPublic opinion classification and text alignment based on Chinese and Tibetan corpus\.Clust\. Comput\.22\(Supplement\),pp\. 10263–10274\.External Links:[Link](https://doi.org/10.1007/s10586-017-1267-8),[Document](https://dx.doi.org/10.1007/S10586-017-1267-8)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1)\.
- Xuet al\.\(2018\)G\. Xu, Z\. Yu, and Q\. QiEfficient Sensitive Information Classification and Topic Tracking Based on Tibetan Web Pages\.IEEE Access6,pp\. 55643–55652\.External Links:[Link](https://doi.org/10.1109/ACCESS.2018.2870122),[Document](https://dx.doi.org/10.1109/ACCESS.2018.2870122)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.10](https://arxiv.org/html/2610.13398#A4.SS10.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Xuet al\.\(2021\)X\. Xu, L\. Yang, Y\. Zhao, and H\. WangEnd\-to\-End Speech Synthesis for Tibetan Multidialect\.Complex\.2021,pp\. 6682871:1–6682871:8\.External Links:[Link](https://doi.org/10.1155/2021/6682871),[Document](https://dx.doi.org/10.1155/2021/6682871)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Yanget al\.\(2015\)H\. Yang, K\. Oura, H\. Wang, Z\. Gan, and K\. TokudaUsing speaker adaptive training to realize Mandarin\-Tibetan cross\-lingual speech synthesis\.Multim\. Tools Appl\.74\(22\),pp\. 9927–9942\.External Links:[Link](https://doi.org/10.1007/s11042-014-2117-9),[Document](https://dx.doi.org/10.1007/S11042-014-2117-9)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1)\.
- Yanget al\.\(2026\)L\. Yang, L\. Pan, B\. Xiong, R\. Jin, S\. Zhang, Y\. Chen, L\. Shi, J\. Zhou, J\. Wu, Z\. Wang, J\. Peng, J\. Xiao, T\. Dong, Z\. Han, Z\. Chen, Y\. Ren, and D\. XiongFrom curated data to scalable models: continual pre\-training of dense and MoE large language models for Tibetan\.InProceedings of the 64th Annual Meeting of the Association for Computational Linguistics \(Volume 1: Long Papers\),M\. Liakata, V\. P\. Moreira, J\. Zhang, and D\. Jurgens \(Eds\.\),San Diego, California, United States,pp\. 40167–40182\.External Links:[Link](https://aclanthology.org/2026.acl-long.1866/),[Document](https://dx.doi.org/10.18653/v1/2026.acl-long.1866),ISBN 979\-8\-89176\-390\-6Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Yeet al\.\(2015\)C\. Ye, W\. Wen, and C\. WangChinese\-Tibetan bilingual clustering based on random walk\.Neurocomputing158,pp\. 32–41\.External Links:[Link](https://doi.org/10.1016/j.neucom.2015.02.003),[Document](https://dx.doi.org/10.1016/J.NEUCOM.2015.02.003)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.10](https://arxiv.org/html/2610.13398#A4.SS10.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1)\.
- Yeet al\.\(2014\)C\. Ye, W\. Wen, and P\. YangTM\-ToT: An Effective Model for Topic Mining from the Tibetan Messages\.InNatural Language Processing and Chinese Computing \- Third CCF Conference, NLPCC 2014, Shenzhen, China, December 5\-9, 2014\. Proceedings,C\. Zong, J\. Nie, D\. Zhao, and Y\. Feng \(Eds\.\),Communications in Computer and Information Science, Vol\.496,pp\. 438–448\.External Links:[Link](https://doi.org/10.1007/978-3-662-45924-9/_40),[Document](https://dx.doi.org/10.1007/978-3-662-45924-9%5F40)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.10](https://arxiv.org/html/2610.13398#A4.SS10.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px4.p1.1)\.
- Yuet al\.\(2024\)L\. Yu, F\. Li, L\. Yu, W\. Li, Z\. Dong, D\. Cai, and Z\. WangA novel Chinese\-Tibetan mixed\-language rumor detector with multi\-extractor representations\.Comput\. Speech Lang\.87,pp\. 101625\.External Links:[Link](https://doi.org/10.1016/j.csl.2024.101625),[Document](https://dx.doi.org/10.1016/J.CSL.2024.101625)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.2](https://arxiv.org/html/2610.13398#A4.SS2.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Zhanget al\.\(2024\)C\. Zhang, M\. Tao, Q\. Huang, J\. Lin, Z\. Chen, and Y\. FengMC2\{\}^\{2\}: towards transparent and culturally\-aware NLP for minority languages in China\.InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics \(Volume 1: Long Papers\),L\. Ku, A\. Martins, and V\. Srikumar \(Eds\.\),Bangkok, Thailand,pp\. 8832–8850\.External Links:[Link](https://aclanthology.org/2024.acl-long.479/),[Document](https://dx.doi.org/10.18653/v1/2024.acl-long.479)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px1.p1.1)\.
- Zhanget al\.\(2025a\)C\. Zhang, M\. Tao, Z\. Liao, and Y\. FengMiLiC\-eval: benchmarking multilingual LLMs for China’s minority languages\.InFindings of the Association for Computational Linguistics: ACL 2025,W\. Che, J\. Nabende, E\. Shutova, and M\. T\. Pilehvar \(Eds\.\),Vienna, Austria,pp\. 11086–11102\.External Links:[Link](https://aclanthology.org/2025.findings-acl.578/),[Document](https://dx.doi.org/10.18653/v1/2025.findings-acl.578),ISBN 979\-8\-89176\-256\-5Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Zhanget al\.\(2023\)G\. Zhang, W\. Wang, C\. Zhang, P\. Zhao, and M\. ZhangHUTNet: An Efficient Convolutional Neural Network for Handwritten Uchen Tibetan Character Recognition\.Big Data11\(5\),pp\. 387–398\.External Links:[Link](https://doi.org/10.1089/big.2021.0333),[Document](https://dx.doi.org/10.1089/BIG.2021.0333)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.5](https://arxiv.org/html/2610.13398#A4.SS5.p1.1)\.
- Zhanget al\.\(2025b\)J\. Zhang, Z\. Zhang, L\. Yeshi, D\. Tashi, X\. Wang, Y\. Cai, Y\. Yu, X\. Wang, N\. Tashi, and G\. LuosangTibetan Medical Named Entity Recognition Based on Syllable\-Word\-Sentence Embedding Transformer\.CAAI Trans\. Intell\. Technol\.10\(4\),pp\. 1148–1158\.External Links:[Link](https://doi.org/10.1049/cit2.70029),[Document](https://dx.doi.org/10.1049/CIT2.70029)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.1](https://arxiv.org/html/2610.13398#A4.SS1.p1.1),[§D\.9](https://arxiv.org/html/2610.13398#A4.SS9.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px5.p1.1)\.
- Zhanget al\.\(2026a\)R\. Zhang, H\. Liu, G\. Li, Z\. Wang, Z\. Ye, Z\. Liu, and C\. JiaTibetan Sentiment Analysis Based on Cross\-Language Sentence Vector Contrastive Distillation Learning\.KSII Trans\. Internet Inf\. Syst\.20\(6\),pp\. 2851–2873\.External Links:[Link](https://doi.org/10.3837/tiis.2026.06.001),[Document](https://dx.doi.org/10.3837/TIIS.2026.06.001)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.26](https://arxiv.org/html/2610.13398#A4.SS26.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Zhanget al\.\(2026b\)W\. Zhang, M\. Wang, J\. Li, and H\. YangCro\-MTVITS: An end\-to\-end cross\-lingual speech synthesis model for Mandarin and multi\-dialect Tibetan based on VITS\.Comput\. Speech Lang\.100,pp\. 101956\.External Links:[Link](https://doi.org/10.1016/j.csl.2026.101956),[Document](https://dx.doi.org/10.1016/J.CSL.2026.101956)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Zhanget al\.\(2019\)W\. Zhang, H\. Yang, X\. Bu, and L\. WangDeep Learning for Mandarin\-Tibetan Cross\-Lingual Speech Synthesis\.IEEE Access7,pp\. 167884–167894\.External Links:[Link](https://doi.org/10.1109/ACCESS.2019.2954342),[Document](https://dx.doi.org/10.1109/ACCESS.2019.2954342)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1)\.
- Zhang and Yang \(2023\)W\. Zhang and H\. YangImproving Sequence\-to\-sequence Tibetan Speech Synthesis with Prosodic Information\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.22\(9\),pp\. 225:1–225:13\.External Links:[Link](https://doi.org/10.1145/3616012),[Document](https://dx.doi.org/10.1145/3616012)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1)\.
- Zhanget al\.\(2025c\)Z\. Zhang, Y\. Yu, X\. Wang, X\. Feng, Y\. Li, J\. Shen, D\. Tashi, J\. Zhang, L\. Yeshi, L\. Li, N\. Tashi, and J\. CaiTibetan Few\-Shot Learning Model With Deep Contextualised Two\-Level Word Embeddings\.CAAI Trans\. Intell\. Technol\.10\(5\),pp\. 1394–1410\.External Links:[Link](https://doi.org/10.1049/cit2.70047),[Document](https://dx.doi.org/10.1049/CIT2.70047)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1)\.
- Zhaoet al\.\(2024\)G\. Zhao, W\. Wang, X\. Wang, X\. Bao, H\. Li, and M\. LiuIncremental Recognition of Multi\-Style Tibetan Character Based on Transfer Learning\.IEEE Access12,pp\. 44190–44206\.External Links:[Link](https://doi.org/10.1109/ACCESS.2024.3381039),[Document](https://dx.doi.org/10.1109/ACCESS.2024.3381039)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Zhaoet al\.\(2018\)Q\. Zhao, L\. Ma, and L\. DuanA Touching Character Database from Tibetan Historical Documents to Evaluate the Segmentation Algorithm\.InPattern Recognition and Computer Vision \- First Chinese Conference, PRCV 2018, Guangzhou, China, November 23\-26, 2018, Proceedings, Part IV,J\. Lai, C\. Liu, X\. Chen, J\. Zhou, T\. Tan, N\. Zheng, and H\. Zha \(Eds\.\),Lecture Notes in Computer Science, Vol\.11259,pp\. 309–321\.External Links:[Link](https://doi.org/10.1007/978-3-030-03341-5/_26),[Document](https://dx.doi.org/10.1007/978-3-030-03341-5%5F26)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Zhaoet al\.\(2019a\)Y\. Zhao, P\. Hu, X\. Xu, L\. Wu, and X\. LiLhasa\-Tibetan Speech Synthesis Using End\-to\-End Model\.IEEE Access7,pp\. 140305–140311\.External Links:[Link](https://doi.org/10.1109/ACCESS.2019.2940125),[Document](https://dx.doi.org/10.1109/ACCESS.2019.2940125)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Zhaoet al\.\(2012\)Y\. Zhao, X\. Xu, and G\. YangUnsupervised Tibetan speech features Learning based on Dynamic Bayesian Networks\.InProceedings of the 21st International Conference on Pattern Recognition, ICPR 2012, Tsukuba, Japan, November 11\-15, 2012,pp\. 2319–2322\.External Links:[Link](https://ieeexplore.ieee.org/document/6460629/)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.15](https://arxiv.org/html/2610.13398#A4.SS15.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Zhaoet al\.\(2020\)Y\. Zhao, X\. Xu, J\. Yue, W\. Song, X\. Li, L\. Wu, and Q\. JiAn open speech resource for Tibetan multi\-dialect and multitask recognition\.Int\. J\. Comput\. Sci\. Eng\.22\(2/3\),pp\. 297–304\.External Links:[Link](https://doi.org/10.1504/IJCSE.2020.107351),[Document](https://dx.doi.org/10.1504/IJCSE.2020.107351)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px1.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px5.p1.1)\.
- Zhaoet al\.\(2019b\)Y\. Zhao, J\. Yue, X\. Xu, L\. Wu, and X\. LiEnd\-to\-End\-Based Tibetan Multitask Speech Recognition\.IEEE Access7,pp\. 162519–162529\.External Links:[Link](https://doi.org/10.1109/ACCESS.2019.2952406),[Document](https://dx.doi.org/10.1109/ACCESS.2019.2952406)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px1.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Zhouet al\.\(2026\)M\. Zhou, X\. Qi, W\. Bao, and X\. ZhaoTibHCR: A Large\-Scale Dataset for Offline Handwritten Tibetan Character Recognition\.IEICE Trans\. Inf\. Syst\.109\(1\),pp\. 193–205\.External Links:[Link](https://doi.org/10.1587/transinf.2025edp7098),[Document](https://dx.doi.org/10.1587/TRANSINF.2025EDP7098)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.18](https://arxiv.org/html/2610.13398#A4.SS18.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px6.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px3.p1.1)\.
- Zhouet al\.\(2024a\)M\. Zhou, Q\. Gesang, N\. Qun, T\. Nyima, and R\. DongrubTibetan\-Chinese Machine Translation Enhanced on Cross\-Lingual Pre\-Trained Model\.InIEEE International Conference on Systems, Man, and Cybernetics, SMC 2024, Kuching, Malaysia, October 6\-10, 2024,pp\. 1618–1623\.External Links:[Link](https://doi.org/10.1109/SMC54092.2024.11169736),[Document](https://dx.doi.org/10.1109/SMC54092.2024.11169736)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.16](https://arxiv.org/html/2610.13398#A4.SS16.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px4.p1.1)\.
- Zhouet al\.\(2024b\)M\. Zhou, D\. Zhuoma, N\. Qun, and T\. NyimaEfficient Fine\-Tuning for Low\-Resource Tibetan Pre\-trained Language Models\.InArtificial Neural Networks and Machine Learning \- ICANN 2024 \- 33rd International Conference on Artificial Neural Networks, Lugano, Switzerland, September 17\-20, 2024, Proceedings, Part VII,M\. Wand, K\. Malinovská, J\. Schmidhuber, and I\. V\. Tetko \(Eds\.\),Lecture Notes in Computer Science, Vol\.15022,pp\. 410–422\.External Links:[Link](https://doi.org/10.1007/978-3-031-72350-6/_28),[Document](https://dx.doi.org/10.1007/978-3-031-72350-6%5F28)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.5](https://arxiv.org/html/2610.13398#A4.SS5.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Zhuet al\.\(2019\)J\. Zhu, S\. Wang, Y\. Wu, M\. Guan, and Y\. FengA New Algorithm for Component Decomposition and Type Recognition of Tibetan Syllable\.InNatural Language Processing and Chinese Computing \- 8th CCF International Conference, NLPCC 2019, Dunhuang, China, October 9\-14, 2019, Proceedings, Part II,J\. Tang, M\. Kan, D\. Zhao, S\. Li, and H\. Zan \(Eds\.\),Lecture Notes in Computer Science, Vol\.11839,pp\. 381–392\.External Links:[Link](https://doi.org/10.1007/978-3-030-32236-6/_34),[Document](https://dx.doi.org/10.1007/978-3-030-32236-6%5F34)Cited by:[§B\.1](https://arxiv.org/html/2610.13398#A2.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§2\.1](https://arxiv.org/html/2610.13398#S2.SS1.p2.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1)\.
- Zhu and Huang \(2020\)X\. Zhu and H\. HuangEnd\-to\-End Amdo\-Tibetan Speech Recognition Based on Knowledge Transfer\.IEEE Access8,pp\. 170991–171000\.External Links:[Link](https://doi.org/10.1109/ACCESS.2020.3023783),[Document](https://dx.doi.org/10.1109/ACCESS.2020.3023783)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.27](https://arxiv.org/html/2610.13398#A4.SS27.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px2.p1.1)\.
- Zhuanget al\.\(2025a\)W\. Zhuang, D\. Cairen, and Y\. SunTIFD: Tibetan Instruction\-Following Dataset for Large Language Models Supervised Fine\-Tuning\.Data Intell\.7\(3\),pp\. 776–785\.External Links:[Link](https://doi.org/10.3724/2096-7004.di.2024.0010),[Document](https://dx.doi.org/10.3724/2096-7004.DI.2024.0010)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.12](https://arxiv.org/html/2610.13398#A4.SS12.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p2.1)\.
- Zhuang and Sun \(2025\)W\. Zhuang and Y\. SunCUTE: a multilingual dataset for enhancing cross\-lingual knowledge transfer in low\-resource languages\.InProceedings of the 31st International Conference on Computational Linguistics,O\. Rambow, L\. Wanner, M\. Apidianaki, H\. Al\-Khalifa, B\. D\. Eugenio, and S\. Schockaert \(Eds\.\),Abu Dhabi, UAE,pp\. 10037–10046\.External Links:[Link](https://aclanthology.org/2025.coling-main.670/)Cited by:[§A\.1](https://arxiv.org/html/2610.13398#A1.SS1.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.17](https://arxiv.org/html/2610.13398#A4.SS17.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[Appendix F](https://arxiv.org/html/2610.13398#A6.SS0.SSS0.Px3.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§5\.2](https://arxiv.org/html/2610.13398#S5.SS2.p1.1),[§5\.3](https://arxiv.org/html/2610.13398#S5.SS3.p1.1)\.
- Zhuanget al\.\(2025b\)Y\. Zhuang, Y\. Sun, Y\. Li, S\. Liu, and X\. ZhaoTibetan Question Generation Based on Key Sentence and Knowledge Graph\.ACM Trans\. Asian Low Resour\. Lang\. Inf\. Process\.24\(5\),pp\. 42:1–42:16\.External Links:[Link](https://doi.org/10.1145/3725531),[Document](https://dx.doi.org/10.1145/3725531)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.23](https://arxiv.org/html/2610.13398#A4.SS23.p1.1),[§D\.7](https://arxiv.org/html/2610.13398#A4.SS7.p1.1)\.
- Zhuoma and Zhijie \(2026\)C\. Zhuoma and C\. ZhijieA multi\-granularity word segment evaluation set construction method for Tibetan language\.Eng\. Appl\. Artif\. Intell\.181,pp\. 115301\.External Links:[Link](https://doi.org/10.1016/j.engappai.2026.115301),[Document](https://dx.doi.org/10.1016/J.ENGAPPAI.2026.115301)Cited by:[§B\.2](https://arxiv.org/html/2610.13398#A2.SS2.p1.1),[§C\.2](https://arxiv.org/html/2610.13398#A3.SS2.p1.1),[§D\.22](https://arxiv.org/html/2610.13398#A4.SS22.p1.1),[§D\.24](https://arxiv.org/html/2610.13398#A4.SS24.p1.1),[§4\.2](https://arxiv.org/html/2610.13398#S4.SS2.SSS0.Px1.p1.1),[§6](https://arxiv.org/html/2610.13398#S6.SS0.SSS0.Px2.p1.1)\.

## Appendix ATibetan NLP Conference/Workshop Papers Found from ACL Anthology

### A\.1Conference Papers

### A\.2Workshop Papers

### A\.3Conference List

⋅\\cdotAACL: Asia\-Pacific Chapter of the Association for Computational Linguistics

⋅\\cdotACL: Annual Meeting of the Association for Computational Linguistics

⋅\\cdotCOLING: International Conference on Computational Linguistics

⋅\\cdotEMNLP: Conference on Empirical Methods in Natural Language Processing

⋅\\cdotIJCNLP: International Joint Conference on Natural Language Processing

⋅\\cdotINLG: International Natural Language Generation Conference

⋅\\cdotLREC: International Conference on Language Resources and Evaluation

⋅\\cdotPACLIC: Pacific Asia Conference on Language, Information and Computation

⋅\\cdotRANLP: International Conference on Recent Advances in Natural Language Processing

### A\.4Workshop List

⋅\\cdotALP: Workshop on Ancient Language Processing

⋅\\cdotALR: Workshop on Asian Language Resources

⋅\\cdotALTA: Australasian Language Technology Association Workshop

⋅\\cdotEURALI: Workshop on Resources and Technologies for Indigenous, Endangered and Lesser\-resourced Languages in Eurasia

⋅\\cdotKaLLM: Workshop on Knowledge Graphs and Large Language Models

⋅\\cdotLChange: International Workshop on Computational Approaches to Historical Language Change

⋅\\cdotLoResMT: Workshop on Technologies for MT of Low Resource Languages

⋅\\cdotNLP4DH: International Workshop on Natural Language Processing for Digital Humanities

⋅\\cdotTLT: International Workshop on Treebanks and Linguistic Theories

⋅\\cdotTrustNLP: Workshop on Trustworthy Natural Language Processing

## Appendix BTibetan NLP Conference Papers / Journal Articles Found from DBLP

### B\.1Conference Papers

### B\.2Journal Articles

### B\.3Conference List

⋅\\cdotACML: Asian Conference on Machine Learning

⋅\\cdotCSCWD: International Conference on Computer Supported Cooperative Work in Design

⋅\\cdotICANN: International Conference on Artificial Neural Networks

⋅\\cdotICASSP: IEEE International Conference on Acoustics, Speech, and Signal Processing

⋅\\cdotICDAR: International Conference on Document Analysis and Recognition

⋅\\cdotICONIP: International Conference on Neural Information Processing

⋅\\cdotICPR: International Conference on Pattern Recognition

⋅\\cdotIJCNN: International Joint Conference on Neural Networks

⋅\\cdotInterspeech: Annual Conference of the International Speech Communication Association

⋅\\cdotNLPCC: CCF International Conference on Natural Language Processing and Chinese Computing

⋅\\cdotPRCV: Chinese Conference on Pattern Recognition and Computer Vision

⋅\\cdotSMC: IEEE International Conference on Systems, Man, and Cybernetics

⋅\\cdotTrustCom: IEEE International Conference on Trust, Security and Privacy in Computing and Communications

### B\.4Journal List

⋅\\cdotACM Transactions on Asian and Low\-Resource Language Information Processing

⋅\\cdotApplied Soft Computing

⋅\\cdotBig Data

⋅\\cdotCAAI Transactions on Intelligence Technology

⋅\\cdotCluster Computing

⋅\\cdotComplex & Intelligent Systems

⋅\\cdotComplexity

⋅\\cdotComputer Speech & Language

⋅\\cdotData Intelligence

⋅\\cdotEngineering Applications of Artificial Intelligence

⋅\\cdotEURASIP Journal on Audio, Speech, and Music Processing

⋅\\cdotExpert Systems with Applications

⋅\\cdotIEEE Access

⋅\\cdotIEEE Signal Processing Letters

⋅\\cdotIEEE Transactions on Computational Social Systems

⋅\\cdotIEICE Transactions on Information and Systems

⋅\\cdotInternational Journal of Computational Science and Engineering

⋅\\cdotInternational Journal of Pattern Recognition and Artificial Intelligence

⋅\\cdotJournal of Advanced Computational Intelligence and Intelligent Informatics

⋅\\cdotJournal of Computational Methods in Sciences and Engineering

⋅\\cdotJournal of Visual Communication and Image Representation

⋅\\cdotKSII Transactions on Internet and Information Systems

⋅\\cdotLanguage Resources and Evaluation

⋅\\cdotMultimedia Tools and Applications

⋅\\cdotNeurocomputing

## Appendix CPublications Classified by Country of First Author’s Affiliation

### C\.1Australia

### C\.2China

### C\.3France

### C\.4Germany

### C\.5India

### C\.6Israel

### C\.7UK

### C\.8USA

## Appendix DPublications Classified by Research Area in ARR

### D\.1Clinical and Biomedical Applications

### D\.2Computational Social Science and Cultural Analytics

### D\.3Dialogue and Interactive Systems

None\.

### D\.4Discourse and Pragmatics

### D\.5Efficient Methods for NLP

### D\.6Ethics, Bias, and Fairness

None\.

### D\.7Generation

### D\.8Human\-Centered NLP and Human\-AI Interaction

### D\.9Information Extraction

### D\.10Information Retrieval and Text Mining

### D\.11Interpretability and Analysis of Models for NLP

### D\.12Language Modeling

### D\.13LLM agents

None\.

### D\.14Linguistic Theories, Cognitive Modeling, and Psycholinguistics

### D\.15Machine Learning for NLP

### D\.16Machine Translation

### D\.17Multilingualism and Cross\-Lingual NLP

### D\.18Multimodality and Language Grounding to Vision, Robotics and Beyond

### D\.19NLP and Code Models

None\.

### D\.20NLP and Symbolic Reasoning

### D\.21NLP Applications

### D\.22Phonology, Morphology, and Word Segmentation

### D\.23Question Answering

### D\.24Resources and Evaluation

### D\.25Semantics: Lexical and Sentence\-Level

### D\.26Sentiment Analysis, Stylistic Analysis, and Argument Mining

### D\.27Speech Recognition, Text\-to\-Speech and Spoken Language Understanding

### D\.28Summarization

### D\.29Syntax: Tagging, Chunking and Parsing

## Appendix EFull\-Scale Versions of Figure[3](https://arxiv.org/html/2610.13398#S4.F3),[4](https://arxiv.org/html/2610.13398#S4.F4),[5](https://arxiv.org/html/2610.13398#S4.F5), and[6](https://arxiv.org/html/2610.13398#S4.F6)

Figure[7](https://arxiv.org/html/2610.13398#A7.F7),[8](https://arxiv.org/html/2610.13398#A7.F8),[9](https://arxiv.org/html/2610.13398#A7.F9), and[10](https://arxiv.org/html/2610.13398#A7.F10)are the full\-scale versions of Figure[3](https://arxiv.org/html/2610.13398#S4.F3),[4](https://arxiv.org/html/2610.13398#S4.F4),[5](https://arxiv.org/html/2610.13398#S4.F5), and[6](https://arxiv.org/html/2610.13398#S4.F6), respectively\.

## Appendix FOpen\-Source Resources Released by the Surveyed Publications

Bibliographic metadata cannot reveal whether the datasets, models, and code described in a publication are publicly released\. To complement the bibliometric analysis, we manually examined each of the 140 publications for resources released by its authors; Tablelists the results\.

#### Compilation\.

A resource is included when the authors of a publication release code, data, or models at a publicly accessible URL; resources available only on request and third\-party re\-uploads are excluded\. For each entry, the table gives the name of the resource, the publications that introduce it, its URLs, and a one\-sentence summary of its content and, where reported, its scale and licence\. A resource described in several publications is listed once, under the earliest of them: this applies to the TIBMD@MUC speech corpus\([Zhao et al\., 2019b](https://arxiv.org/html/2610.13398#bib.bib132);[Zhao et al\., 2020](https://arxiv.org/html/2610.13398#bib.bib123)\)and to the PACTib/ACTib corpora and annotation pipelines\([Meelen and Roux, 2020](https://arxiv.org/html/2610.13398#bib.bib25);[Meelen et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib112);[Faggionato et al\., 2022](https://arxiv.org/html/2610.13398#bib.bib23)\)\. Entries are ordered by year of publication and then alphabetically by first author\.

#### Coverage and trend\.

The table contains 38 entries from 41 of the 140 publications \(29\.3%\)\. Public release is a recent practice in Tibetan NLP\. None of the 31 publications from 2004–2018 releases a resource that meets our criteria, compared with 14 of the 48 publications from 2019–2023 \(29\.2%\) and 27 of the 61 from 2024–2026 \(44\.3%\), including 12 of the 18 publications of 2025\. This rise coincides with the transition to pre\-trained and large language models, a period in which releasing corpora, models, and benchmarks has become common practice across NLP\.

#### What is released\.

The released resources span every layer of the pipeline described in Section[4\.2](https://arxiv.org/html/2610.13398#S4.SS2)\. At the foundational layer, they include a lexicon of verbal paradigms\([Di et al\., 2019](https://arxiv.org/html/2610.13398#bib.bib26)\), the PACTib/ACTib treebank and annotation pipelines\([Meelen and Roux, 2020](https://arxiv.org/html/2610.13398#bib.bib25);[Meelen et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib112);[Faggionato et al\., 2022](https://arxiv.org/html/2610.13398#bib.bib23)\), a knowledge graph\([Sun et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib114)\), a word segmenter\([Gyatso et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib89)\), and expert word\-boundary annotations\([Moore et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib75)\)\. Script processing is represented by character datasets of historical documents, handwriting, and cursive script\([Li et al\., 2022c](https://arxiv.org/html/2610.13398#bib.bib106);[Tashi et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib95);[Zhou et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib76)\), code for few\-shot character recognition\([Bao et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib97)\), a multimodal OCR benchmark\([Jie et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib71)\), and a vision–language model\([Xu et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib4)\)\. Speech is represented by the TIBMD@MUC multi\-dialect corpus\([Zhao et al\., 2019b](https://arxiv.org/html/2610.13398#bib.bib132);[Zhao et al\., 2020](https://arxiv.org/html/2610.13398#bib.bib123)\)and a package for measuring recognition error at the word level\([Moore et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib75)\)\. Since 2022, releases have increasingly consisted of PLMs and LLMs\([Liu et al\., 2022](https://arxiv.org/html/2610.13398#bib.bib51);[Liang et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib99);[Lv et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib86);[Sha et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib80);[Zhuang and Sun, 2025](https://arxiv.org/html/2610.13398#bib.bib14)\), the data used to train and align them\([Zhang et al\., 2024](https://arxiv.org/html/2610.13398#bib.bib15);[Zhuang et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib87)\), and benchmarks for evaluating them; six benchmarks date from 2025–2026 alone\([Gao et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib12);[Xu et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib11);[Hashiloni et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib9);[Zhang et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib13);[Li et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib1);[Jie et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib71)\)\. Three further releases, all from one research team, support the evaluation of adversarial robustness\([Cao et al\., 2023](https://arxiv.org/html/2610.13398#bib.bib20);[Cao et al\., 2025a](https://arxiv.org/html/2610.13398#bib.bib38);[Cao et al\., 2025b](https://arxiv.org/html/2610.13398#bib.bib10)\)\. Finally, Classical Tibetan and Buddhist textual heritage accounts for a distinct cluster of releases: a Sanskrit–Tibetan parallel corpus\([Nehrdich, 2022](https://arxiv.org/html/2610.13398#bib.bib22)\), tools and text analytics for Tibetan Buddhist literature\([Trinley et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib115);[Krishna et al\., 2021](https://arxiv.org/html/2610.13398#bib.bib109)\), and recent resources for benchmarking, distinguishing native from translated passages, and measuring sentence similarity\([Hashiloni et al\., 2025](https://arxiv.org/html/2610.13398#bib.bib9);[Bilitski et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib7);[Cohen et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib6)\)\.

#### Release rates by research area\.

The share of publications that release resources varies widely across ARR areas \(Appendix D\)\. It is highest in Resources and Evaluation \(22 of 40; 55\.0%\), Language Modeling \(7 of 16; 43\.8%\), and Multilingualism and Cross\-Lingual NLP \(16 of 41; 39\.0%\)\. Among the larger areas, it is lowest in Speech Recognition, Text\-to\-Speech and Spoken Language Understanding \(3 of 21; 14\.3%\) and in Phonology, Morphology, and Word Segmentation \(2 of 17; 11\.8%\), and none of the publications in Information Retrieval and Text Mining \(6\), Sentiment Analysis, Stylistic Analysis, and Argument Mining \(4\), or Summarization \(3\) releases a resource\. Speech and word segmentation are the areas most directly shaped by two defining properties of Tibetan, dialectal diversity and unmarked word boundaries, yet they have the fewest shared resources\. Consent and privacy considerations can legitimately restrict the release of speech data, but the result is that TIBMD@MUC is the only entry covering the three main dialect groups, and no text resource in the table targets a specific dialect through genres such as proverbs and folk song lyrics\.

#### Release rates by country and venue\.

Release rates also differ by the country of the first author’s affiliation \(Appendix C\)\. Of the 18 publications from outside China, 13 \(72\.2%\) release resources, compared with 28 of the 122 China\-affiliated publications \(23\.0%\)\. Publication venue accounts for part of this difference\. Publications in the ACL Anthology release resources more often than those retrieved from DBLP \(18 of 37, 48\.6%, vs\. 23 of 103, 22\.3%\), and among China\-affiliated publications, the rate is 40\.0% \(10 of 25\) in the ACL Anthology but 18\.6% \(18 of 97\) elsewhere\. This pattern is consistent with the artifact\-release expectations of ACL venues, such as the Responsible NLP Research Checklist\. In absolute terms, China\-affiliated teams remain the largest contributors, accounting for 28 of the 41 releasing publications, all of the released LLMs, and five of the six recent benchmarks\. All of these rates are lower bounds, because resources shared on request or through channels that the publications do not cite are not captured\.

#### Accessibility and sustainability\.

GitHub666[https://github\.com/](https://github.com/)is the most common host \(28 entries\), followed by Hugging Face777[https://huggingface\.co/](https://huggingface.co/)\(13 entries\), and 11 entries distribute their resources across more than one platform, for example Hugging Face with GitHub\([Zhou et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib76)\)or ModelScope888[https://www\.modelscope\.cn/](https://www.modelscope.cn/)with GitHub\([Han et al\., 2026](https://arxiv.org/html/2610.13398#bib.bib5)\)\. Only four entries are deposited in research\-data repositories that assign persistent identifiers \(ScienceDB999[https://www\.scidb\.cn/](https://www.scidb.cn/)and Zenodo101010[https://zenodo\.org/](https://zenodo.org/)\); the others depend on platforms on which repositories can be renamed, moved, or deleted\. Two datasets, TIBMD@MUC and THCU, are distributed only through Baidu Netdisk111111[https://pan\.baidu\.com/](https://pan.baidu.com/), which can be difficult to use from outside China, while some international platforms are, conversely, not reliably accessible from mainland China\. Building on the open\-science recommendations in Section[6](https://arxiv.org/html/2610.13398#S6), we encourage authors of Tibetan NLP resources to \(1\) deposit released data and models in archives that issue persistent identifiers, \(2\) state an explicit licence, \(3\) document each resource with a datasheet or model card, and \(4\) mirror resources on platforms accessible to researchers both inside and outside China\.

#### Caveats\.

Inclusion in Tableindicates that a resource has been publicly released, not that it is complete, documented, or maintained; we did not evaluate the released resources themselves\. As with the rest of the survey, 2026 is only partially covered\. We ensure that all links in Tableare accessible at 00:00 UTC, September 25, 2026, but we cannot guarantee their accessibility in the future\.

## Appendix GNLP Publications on Other Bodish Languages

Complementing the recommendation on related languages in Section[6](https://arxiv.org/html/2610.13398#S6), Tableextends the retrieval of Section[3](https://arxiv.org/html/2610.13398#S3)from Tibetan to all 82 languages of Bodish, the subgroup of Tibeto\-Burman to which Ethnologue121212[https://www\.ethnologue\.com/subgroup/4085/](https://www.ethnologue.com/subgroup/4085/)assigns Tibetan\. For each language, we repeated the retrieval and screening with the keyword in the third column in place ofTibetan\. Entries that differ only in a qualifier, such as the three Tibetan or the three Tamang entries, share a keyword, and Sikkimese was also searched under its alternative nameBhutia\.

The keywordTibetanretrieves the survey corpus itself \(Appendices[A](https://arxiv.org/html/2610.13398#A1)and[B](https://arxiv.org/html/2610.13398#A2)\)\. For the other 79 languages, the search yielded only 12 publications, against 140 for Tibetan: six on Dzongkha, three on Tamang, and one each on Balti, Ladakhi, and Sikkimese\. They address text processing for Dzongkha and Sikkimese \(e\.g\., word segmentation and tagged corpora\), machine translation for Tamang, and speech processing for Balti and Ladakhi\. The Tamang publications, whose titles and abstracts name no variety, are listed under Eastern Tamang\. The keywords of the remaining 72 languages, including all those outside Central and West Bodish, yielded none\. These 12 publications are outside the survey corpus and the analyses of Section[4](https://arxiv.org/html/2610.13398#S4)\. As retrieval covers only the ACL Anthology and DBLP and depends on the keywords, an empty cell means that no publication was retrieved, not that no research on the language exists\.

Figure 7:Full\-scale version of Figure[3](https://arxiv.org/html/2610.13398#S4.F3): distribution of the 140 Tibetan NLP publications by year and venue type, 2004–2026\.Left:stacked bars give the annual number of conference papers \(blue\), workshop papers \(green\), and journal articles \(tan\); numbers inside the segments are per\-type counts, and numbers above the bars are annual totals\. No publications were identified for 2007–2009\. Output rises from three papers in 2017 to nine in 2018 and peaks at 22 in 2024; 2026 is covered only up to 00:00 UTC, September 25, 2026\.Right:overall share of each venue type: journal articles 52\.9% \(74\), conference papers 38\.6% \(54\), and workshop papers 8\.6% \(12\)\.
Figure 8:Full\-scale version of Figure[4](https://arxiv.org/html/2610.13398#S4.F4): number of Tibetan NLP publications per venue; within each panel, venues are sorted by count in descending order, with ties listed alphabetically\.Top left:22 conferences, led by ACL \(8, including Findings of ACL\), NLPCC \(6\), PRCV \(6\), and COLING \(5\)\. Papers at joint conferences \(ACL\-IJCNLP 2015 and IJCNLP\-AACL 2025\) are counted under each constituent venue, so the bars sum to 57 attributions for 54 conference papers\.Top right:10 workshops with 12 papers, led by ALP and EURALI \(2 each\)\.Bottom:25 journals with 74 articles, led byACM Transactions on Asian and Low\-Resource Language Information Processing\(19\),IEEE Access\(14\), andData Intelligence\(8\); journal names follow DBLP abbreviations\. Full venue names are listed in Appendices A and B\.
Figure 9:Full\-scale version of Figure[5](https://arxiv.org/html/2610.13398#S4.F5): number of Tibetan NLP publications by ACL Rolling Review \(ARR\) research area and by country of the first author’s primary affiliation\.Left:the 29 ARR research areas, sorted in descending order\. Each publication was assigned up to three areas \(in practice, 115 received two and 25 received one\), so the counts sum to 255 rather than 140\. Multilingualism and Cross\-Lingual NLP \(41\) and Resources and Evaluation \(40\) lead, whereas Dialogue and Interactive Systems; Ethics, Bias, and Fairness; LLM Agents; and NLP and Code Models have no publications\.Right:countries, sorted in ascending order: China accounts for 122 \(87\.1%\), followed by the USA \(5\), the UK \(4\), Germany \(3\), Israel \(3\), and Australia, France, and India \(one each\)\. Per\-country and per\-area publication lists are given in Appendices C and D, respectively\.
![Refer to caption](https://arxiv.org/html/2610.13398v1/fig_heatmap.png)Figure 10:Full\-scale version of Figure[6](https://arxiv.org/html/2610.13398#S4.F6): number of Tibetan NLP publications in each of the 29 ARR research areas \(rows\) and each year \(columns\), 2004–2026\. Rows follow the order of Figure[9](https://arxiv.org/html/2610.13398#A7.F9)\(descending total count\)\. Cell shading darkens with the number of publications \(scale 0–10\); each non\-empty cell shows its count, and blank cells indicate no publications\. Because a publication can belong to more than one area, the cells sum to 255\. The largest values are Resources and Evaluation in 2026 \(10\) and Multilingualism and Cross\-Lingual NLP in 2024 \(9\); the columns for 2007–2009 are empty, and 2026 is covered only up to 00:00 UTC, September 25, 2026\.
Open\-source ResourcePaperRepository URLDescription