Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).
Please note that all times are shown in the time zone of the conference. The current conference time is: 14th June 2025, 06:09:59pm WEST
Session Chair: Miguel Escobar Varela, National University of Singapore
Location:B210 (TB)
60 places
Presentations
Connecting Threads: Creating a Participatory and Globally Accessible Platform for the Study of Checked Indian Cotton Textiles
Deepthi Murali, Jason Heppler
George Mason University, United States of America
Connecting Threads explores connections between South Indian weavers and Caribbean consumers by linking small and large textile collections and archives to enhance access to global fashion histories. Featuring a PostgreSQL database and interactive visualizations, the paper details its technical development, collaborative methodology, and impact on equity and accessibility in DH.
Centering Civic Engagement with Open Scholarship: The Revolutionary City as a Model for Fostering Public Use of Digital Cultural Heritage
David Ragnar Nelson, Bayard L. Miller
American Philosophical Society, United States of America
The paper presents a two-pronged approach for fostering access to and use of digital archival holdings. This approach combines public use of HTR technologies and public involvement in producing interpretative digital scholarship. The framework presented seeks to encourage civic engagement and dialogue around the holdings.
Advancing OCR and Word Sense Disambiguation for the Jawi Script using LLMs and VLMs
Miguel Escobar Varela, Stephane Bressan, Faizah Zakaria, Ganesh Neelalkanta Iyer, Guo Quan Seng, Pratik Karmakar
National University of Singapore, Singapore
We introduce novel datasets and fine-tuned VLM and LLM models for OCR and word-sense disambiguation for Jawi (a writing system used historically for Malay). Our OCR system that outperforms previous solutions with a Character Error Rate (CER) of 8.66%, and a context-aware word sense disambiguation model that achieves 99.2% accuracy.
