Search

The Collision between AI and Old Uighur Manuscripts: Opportunities and Challenges in the Dunhuang Academy Collection

Author(s)

Abudouriyimu Rousitaimujiang*

Affiliation(s)

Dunhuang Textual Research Institute, Dunhuang Academy, Lanzhou, China *Corresponding Author

Abstract

The Old Uyghur manuscripts collected by Dunhuang Academy are among China's most significant such relics, divided into two main collections: early collected manuscripts and those unearthed from the Northern Area of Mogao Grottoes. The materials under discussion here include Buddhist scriptures, literary works and socio-economic documents, and they are considered to be of the utmost importance for the study of the history and culture of ancient ethnic groups in Northwest China. The subject has been long constrained by fragmentation, with many fragments remaining unclassified and unpublished due to difficulties in deciphering and interpreting the extant material. This is compounded by the paucity of scholars specialising in the field, resulting in a lack of expertise and understanding. The integration of artificial intelligence with image recognition, big data analysis and intelligent collation offers a novel technical approach. The present paper methodically organizes the sources and research status of these manuscripts, with a focus on the opportunities and challenges of integrating AI with manuscript research. The utilisation of artificial intelligence has been demonstrated to yield significant benefits in the domains of fragment assembly, character recognition, digitisation, and comparison. The application of deep learning-based multi-modal feature extraction has been demonstrated to accelerate fragment collation, with convolutional neural networks achieving over 96% accuracy in the recognition of ancient Old Turkic script. The utilisation of intelligent databases and big data comparison has been demonstrated to enhance efficiency and reveal hitherto overlooked textual connections. Nevertheless, the process of deep integration continues to encounter significant challenges. The extinct language known as Old Uyghur is characterised by damaged documents, inconsistent norms and scarce training data. It is evident that artificial intelligence is not yet capable of comprehending religious terminology, cultural nuances, and academic logic with the same depth and breadth as human beings. The proposed methodology involves the construction of high-quality datasets, the optimisation of models, the establishment of an "AI preliminary recognition + expert verification" mechanism, the construction of semantic knowledge graphs, the development of linguistics-embedded models and the promotion of interdisciplinary collaboration. The advancement of the intelligent study of these manuscripts and the provision of a reference for the preservation of ancient script documents are the intended outcomes.

Keywords

Dunhuang Academy; Old Uighur Manuscripts; AI; Digital Protection

References

[1] Yang, Fuxue. Uighur Documents and Uighur Culture. Beijing: Ethnic Publishing House, 2002, pp. 26–27. [2] Tai, Huili. The Origin and Value of Dunhuang Manuscripts Collected by the Dunhuang Academy. China Calligraphy, 2019(17), pp. 50–57, 48–49. [3] Geng, Shimin. A Study on the Fragments of the Uighur Version of the Avatamsaka Sutra (Eighty Fascicles). Minority Languages of China, 1986(3), pp. 59–65. [4] Zhang, Tieshan; Zieme, Peter. A Study on the Uighur Fragments of the Commentary on the Yuanjue Jing Collected by the Dunhuang Academy. Dunhuang Research, 2015(2), pp. 92–101. [5] Zhang Tieshan, Peter Zieme, Two Old Uigur Fragments From Dunhuang Connected With The Pure Land Belief, Acta Orientalia Academiae Scientiarum Hung. Volume 71 (3), 2018, pp.253–261. [6] Zhang Tieshan, Peter Zieme, An Old Uigur Version Of The Kasibhāradvāja Sutta Extended By A Poem, Acta Orientalia Academiae Scientiarum Hung. Volume 72 (2), 2019, pp. 179–206. [7] Zhang Tieshan, Peter Zieme, Further fragments of the Commentary on the Yuanjue jing in Old Uigur from Dunhuang, Acta Orientalia Hung. 73 (2020) 3, pp. 439–450. [8] Zhang, Tieshan; Cui, Yan. A Study on Three Fragments of the Uighur Ekottaragama Sutra Formerly Collected by the Dunhuang Academy. Minority Languages of China, 2020(1), pp. 3–10. [9] Zhang, Tieshan. A Study on the Uighur Fragments of Abhidharmakośa-bhāṣya Formerly Collected by the Dunhuang Academy. Studies on the Inner Asian Languages 36, 2021.10, pp. 1–10. [10] Zhang, Tieshan; Zieme, Peter. A Study on the Uighur Fragments of the Siddham Formerly Collected by the Dunhuang Academy. Minority Languages of China, 2021(6), pp. 106–111. [11]Zhang, Tieshan. A Study on Two Uighur Fragments of the Āgama Sutras Formerly Collected by the Dunhuang Academy. Journal of Dunhuang Studies, 2021(1), pp. 6–11. [12]Tusunjan, Imin; Abdurehim, Rustamjan. A Study on the Uighur Scroll of Dasakarmapathāvadānamālā Collected by the Dunhuang Academy. Journal of Dunhuang Studies, 2022(2), pp. 121–136. [13]Tusunjan, Imin; Abdurehim, Rustamjan. A Study on the Uighur Fragments of Mahā­Amitāyur­vyākhyāna Sutra Formerly Collected by the Dunhuang Academy. Dunhuang and Turfan Studies, Vol. 22, 2023, pp. 353–365. [14]Abdurehim, Rustamjan; Tusunjan, Imin. A Study on the Conjunction of Uighur Poetic Fragments of the Great White Lotus Society Sutra Formerly Collected by the Dunhuang Academy. Dunhuang Research, 2025(1), pp. 97–106. [15]Abdurehim, Rustamjan. A Study on the Uighur Fragments of the Diamond Sutra Formerly Collected by the Dunhuang Academy. Collected Papers of Historical and Linguistic Studies of the Western Regions, 2025(1) (Vol. 23), pp. 163–175. [16]Zhang, Tieshan; Aydar, Mirkamal. A Study on a Fragment of the Chinese Miaofa Lianhua Jing Xuanzan in Uighur Script Formerly Collected by the Dunhuang Academy. Dunhuang and Turfan Studies, Vol. 22, 2023, pp. 345–351. [17]Aydar, Mirkamal. A Study on a Uighur Fragment of the Samyukta Āgama Formerly Collected by the Dunhuang Academy. Collected Papers of Historical and Linguistic Studies of the Western Regions, 2025(1) (Vol. 23), pp. 149–162. [18]Peng, Jinzhang; Wang, Jianjun; Dunhuang Academy (eds.). Northern Caves of Mogao Grottoes at Dunhuang (Vol. 1). Beijing: Cultural Relics Press, 2000. [19]Peng, Jinzhang; Wang, Jianjun; Dunhuang Academy (eds.). Northern Caves of Mogao Grottoes at Dunhuang (Vol. 2). Beijing: Cultural Relics Press, 2004. [20]Peng, Jinzhang; Wang, Jianjun; Dunhuang Academy (eds.). Northern Caves of Mogao Grottoes at Dunhuang (Vol. 3). Beijing: Cultural Relics Press, 2004. [21]Yakup,Abdurishid, On the newly unearthed Uighur Buddhist texts from thenorthern grottoes of Dunhuang.In:Sven Bretfeld and Jens Wilkens (eds.) Indienund Zentralasien:Sprach und Kulturkontakt.Vorträge des Göttinger Symposionsvom 7. Mai bis 10. Mai. (Veröffentlichungen der Societas Uralo-Altaica 61.)Wiesbaden, 2001, pp.259-276; [22]Abdurishit Yakup. An Overview of Uighur Buddhist Documents Unearthed from the Northern Caves of Dunhuang. In: Peng Jinzhang (ed.) Studies on the Northern Caves of Mogao Grottoes at Dunhuang (Vol. 2). Lanzhou: Reader Publishing House, 2011, pp. 485–502. [23]Abdurishid YAKUP. Uighurica from the Northern Grottoes of Dunhuang. In: Festschrift für Shoichiro Shōnaichi: Studies on the Eurasian Languages. Tokyo: The Society for Eurasian Languages, 2006, pp. 1–41. [24]Abdurishit Yakup. A Comprehensive Study on Uighur Documents Unearthed from the Northern Caves of Dunhuang. In: Peng Jinzhang (ed.) Studies on the Northern Caves of Mogao Grottoes at Dunhuang (Vol. 2). Lanzhou: Reader Publishing House, 2011, pp. 429–477. [25]Cai, Mengling. Current Situation and New Opportunities: An Analysis of the Application of Artificial Intelligence in the Preservation and Restoration of Paper Documents. Journal of Archives, 2026(1), pp. 28–37. [26]Abdurushid YAKUP, Altuigurische Aparimitāyus-Literatur und kleinere tantrische Texte, BerlinerTurfantexte XXXVI, Turnhout: Brepols Publishers n．v．, 2016. [27]Seyed Hossein Taheri at all, A Deep Learning Based Optical Character Recognition Model for Old Turkic, EAI Endorsed Transactions on AI and Robotics, Volume 4 , 2025. pp. 1-12. [28]Zhang, Tieshan. A Study on Three Precious Uighur Buddhist Fragments Unearthed from the Northern District of Mogao Grottoes. Dunhuang Research, 2004(1), pp. 78–82. [29]Niu, Ruji. A Restudy of the Syriac Christian-Uighur Buddhist Bilingual Manuscript Found in the Northern District of Mogao Grottoes. Dunhuang Research, 2002(2), pp. 56–63. [30]Peter Zieme, Zwei uigurische Gedichte aus Dunhuang – Ein Deutungsversuch. Türk Dilleri Araştırmalı, Cilt 11 (2001), pp. 25-136.