The Collision between AI and Old Uighur Manuscripts: Opportunities and Challenges in the Dunhuang Academy Collection
Academic Conferences Series · 2026-02 · 19 (1)
DOI: 10.62381/acs.hsms2026.03 ↗Kurz gefasst
Fachartikel in der Zeitschrift Academic Conferences Series, 2026-02. Verfasst von Abudouriyimu Rousitaimujiang.
Abstract (Originalsprache)
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 The subject has been long constrained by fragmentation, with many fragments remaining unclassified and unpublished due to difficulties in deciphering and interpreting the extant material.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.
Zitieren
APA
Rousitaimujiang, A. (2026). The Collision between AI and Old Uighur Manuscripts: Opportunities and Challenges in the Dunhuang Academy Collection. Academic Conferences Series, 19(1). https://doi.org/10.62381/acs.hsms2026.03
Vancouver
Rousitaimujiang A. The Collision between AI and Old Uighur Manuscripts: Opportunities and Challenges in the Dunhuang Academy Collection. Academic Conferences Series. 2026;19(1). doi:10.62381/acs.hsms2026.03
BibTeX
@article{Rousitaimujiang2026,
author = {Rousitaimujiang, Abudouriyimu},
title = {The Collision between AI and Old Uighur Manuscripts: Opportunities and Challenges in the Dunhuang Academy Collection},
journal = {Academic Conferences Series},
year = {2026},
volume = {19},
number = {1},
doi = {10.62381/acs.hsms2026.03}
}
EUVIMED