Theoretical studies in Digital Humanities have paid ample attention to the concepts of "models" and "modelling". A relatively recent addition to this line of inquiry are so-called "generative" models, where algorithms are optimized to generate new, "synthetic" data, rather than describe or analyse pre-existing, "authentic" data. Many applications of this rapidly evolving technology (e.g. adversarial networks) are exciting and trigger the collective imagination, ranging from generating artifical bedroom images to generating cooking recipes and hiphop lyrics. The epistemologic status of these models, nevertheless remains problematic (i.e. we have a new tool, but we don't know yet what to use it for). This talk will introduce the notion of generative (or "predictive") models and further explore a series of applications in the Digital Humanities.
Mike Kestemont is assistant professor in Digital Text Analysis at the University of Antwerp. He specializes in computational text analysis through artificial intelligence for the humanities. Mike has a background in computational stylistics and published on the topic of authorship attribution in various fields, such as classics or medieval European literature. Mike actively engages in the debate surrounding the digital humanities and attempts to fuse modern methods from artificial intelligence with traditional scholarship in the humanities, including literary theory. His website (www.mike-kestemont.org) contains pointers to his recent scholarly activities, including an open access scientific documentary about stylometry and Hildegard of Bingen (vimeo.com/70881172).
The lecture is organized as part of the Summer school Cast 2017.
When: 21.9.2017, 18:00 Uhr
Where: Zentrum für Informationsmodellierung, Seminarraum, Elisabethstraße 59/III, 8010 Graz