Come and enjoy a taste of Digital Scholarship as an HKUST HUMA faculty member shares on his recent projects

Manchu (滿文) OCR & AI: From Fine-Tuned Models to Agentic Workflows

Manchu (滿文) is a critically endangered language: few people can read or understand it. Yet it is vital to help us understand early modern Chinese (Qing dynasty) and Eastern Eurasian history. Among the challenges of using Manchu sources the lack of effective OCR systems to handle real-world historical documents.

In this talk, Prof. Chung will share his recent project, which uses fine-tuned vision–language models to enable OCR for low-resource historical texts. He will also discuss how agentic AI can automate post-correction, validation, and iterative review.

Using this project as a case study, it shows that combining model development with scalable workflows creates new possibilities for cross-disciplinary collaboration and innovation in large-scale digitization and analysis of historical materials. It also demonstrates how humanities-driven challenges—and the methods of humanities scholars—can help advance more robust and adaptable AI systems across disciplines.

About the speaker:

Professor Michael Chung received his PhD in History from Emory University in 2025, and his BA and MPhil from the Chinese University of Hong Kong in 2012 and 2016, respectively. His research focuses on the early Qing dynasty, particularly the transfer of European artillery technology and the formation of the Hanjun Eight Banners. As a digital humanist, he works on applying AI methods to historical research, including the development of a Manchu OCR system for low-resource historical texts and a database of seventeenth- and eighteenth-century bannermen.