By demonstrating the great potential of Artificial Intelligence (AI) in processing, translating, and transcribing ancient manuscripts in traditional Chinese characters, the Universitas Indonesia (UI) Library Research Team won 1st place at the 15th Indonesian Digital Library Conference (KPDI) in 2024. The conference was organized by the National Library of the Republic of Indonesia together with the Indonesian Digital Library Forum and Universitas Lampung on 6-9 August 2024, at the Grand Mercure Hotel, Bandar Lampung.
From the research conducted by Mariyah, S.Sos., M.Hum., and Sony Pawoko, S.Sos., M.T.I., it is known that the process of cataloging ancient manuscripts with the help of AI technology has become easier and can be applied in the system at the UI Library.
Currently, UI has 10,000 titles of ancient manuscript collections in traditional Chinese script. These manuscripts are not only historically important for the Indonesian nation, but also provide insight into the life, culture, law, government, philosophy, values, and knowledge developed by Chinese civilization over the centuries.
Google Lens’ deep learning-based Optical Character Recognition (OCR) technology successfully recognizes traditional Chinese characters with high precision. The results of the OCR can be perfectly copied, making it easier to translate texts contained in ancient manuscripts.
“With the increasing capabilities of AI technology, users can also find and understand the contents of traditional Chinese ancient manuscripts even though they have never studied traditional Chinese writing and language,” said Mariyah, who is also the Head of the UI Library UPT.
According to Sony, many traditional Chinese ancient manuscripts are related to primary sources of historical literature in Indonesia. The Kudadu Inscription and the Kakawin Negara Kertagama which discuss the founding of Majapahit, need to be supported by contemporary literature from China.
Contemporary literary sources were written during the Yuan dynasty, and also, inscriptions were made by generals sent by Kublai Khan to conquer the archipelago. The mission was also to seek revenge for the treatment of King Kertanegara of Singhasari who refused to submit to the Mongol empire and even disfigured the face and cut off the ear of his envoy named Men Shi.
“From contemporary sources from China, we can complete the 5W + 1H of incomplete historical writing. In the Pasir Cina inscription on Serutu Island which is written in traditional Chinese script, we know that the attack on Java by the Yuan Dynasty used 500 ships, if one ship can carry 50 people then according to the records of the generals, namely Shi-bi, Ike Mese, and Gaoxing who said that the number of troops was between 20,000-30,000 war soldiers. From their records we also know the time of arrival and attack on Daha,” said Sony.
He further added that this finding is certainly very proud because it can be the beginning of processing the collection of Chinese ancient manuscripts that have not been managed and served to library users for decades.
The 15th KPDI in 2024 was attended by more than 350 participants and aimed to be a means of sharing experiences, knowledge, suggestions, and formulations regarding the implementation and utilization of AI in digital libraries. The series of events at the conference included paper presentations, workshops, panel discussions, exhibitions, and an introduction to Lampung arts and culture. A total of 110 call-for-paper papers were submitted to attend this conference, but only 16 papers were selected and passed to be presented at the 15th KPDI event.
