Shifting Sentiments? What happens to BERT-based Sentiment Classification when derived text formats are used for fine-tuning
Du, Keli (1); Schöch, Christof (2)
1: Trier Center for Digital Humanities, Germany; 2: University of Trier, Germany
HTML XMLIn response to copyright law restricting Text and Data Mining, derived text formats (aka extracted features) were proposed. This work shows that selectively reducing information in DTFs has minor negative impact on DistilBERT-based sentiment classification. Therefore, it is useful to publish copyrighted texts in DTFs, e.g. for fine-tuning LLMs.
Developing the Bernburg 3D Model for the Digital Transformation of Euthanasia Education.
Østergaard, Camilla Vang (1); Rachow, Esther (2); Gebauer, Judith (3); Jänicke, Stefan (1)
1: University of Southern Denmark, Denmark; 2: The Hebrew University of Jerusalem, Israel; 3: Gedenkstätte Bernburg, Germany
HTML XML3D models are valuable for conveying and preserving cultural heritage and enhancing education. We developed a 3D model of the Bernburg Euthanasia Centre. This paper explores its potential use in teaching, illustrating its benefits and implications for future educational practices and historical understanding in school curricula.
A digital 4D information system at world scale: Research challenges, approaches and preliminary results
Muenster, Sander (1); Maiwald, Ferdinand (2); Bruschke, Jonas (1); Kröber, Cindy (1); Sun, Ying (1); Dworak, Daniel (1); Komorowicz, Dávid (1); Munir, Iqra (3); Beck, Clemens (1); Münster, Dora Luise (1)
1: Digital Humanities, FSU Jena, Germany; 2: Photogrammetry, TU Dresden, Germany; 3: Geospatial Science, School of Science, STEM College, RMIT University, Melbourne, VIC 3000, Australia
HTML XMLDigital 3D reconstructions of no more extant cityscapes are so far almost solely done for specific cases and via manual modelling techniques. Within the Jena4D research group we are investigating and developing methods and technologies for transferring historical media and their contextual information into 4D models.