| Session | |
SP-08
| |
| Presentations | |
Interpretable Computer Vision: Multiple Instance Learning for Colonial Korean Print Leiden University This study demonstrates how Multiple Instance Learning enables both accurate and interpretable analysis of visual features in colonial Korean printshops. While achieving 92% accuracy, our model reveals that reliable identification depends on examining common rather than distinctive elements, making computational analysis meaningful for humanities research. Digitising Fels Cave, Lelepa Island, Vanuatu 1Australian National University; 2Stepwise Heritage and Tourism Pty. Ltd; 3Lelema World Heritage Committee; 4Vanuatu Cultural Centre This paper reports on a project in which a multidisciplinary team, the Lelepa community, and Vanuatu cultural heritage staff digitised Fels Cave, a UNESCO World Heritage site on the island of Lelepa in Vanuatu. The site, with engraved and painted rock art walls, is of considerable cultural and spiritual significance. Revisiting Dalgado: Tracing the Heritage of the Portuguese Language in South Asia 1CNR-ILC, Italy; 2CLUNL, NOVA University Lisbon, Portugal; 3Lancaster University, UK; 4UPV/EHU University of the Basque Country, Spain; 5University of Colombo, Sri Lanka; 6Insight Centre for Data Analytics, NUI Galway, Ireland The current submission describes the latest developments within the project Cultural HeritAge and Multilingual Understanding through lexiCal Archives (CHAMUÇA). The latter initiative seeks to create a (linked data) knowledge graph that analyses the impact of Portuguese on the vocabulary of numerous Asian languages. Speculating on the Future of Digital Humanities Research with Copyrighted Materials 1Texas A&M University, United States of America; 2Temple University, United States of America The steep barriers that Digital Humanists face when assembling datasets are made insurmountable by perceived copyright restrictions. This paper will introduce the Data Speculations project, which combines a speculative approach with fair use interpretation to imagine cultural heritage workers and researchers stewarding - rather than licensing - corpora of copyrighted cultural data. | |