Digital humanities is a newly-emerging interdisciplinary field bringing together experts in humanities, social science, and computer science. Several efforts have been recently carried out in both fundamental and applied research, providing valid algorithms to deal with multimodal data gathered from cultural heritage scenarios. In this respect, the large amount of data and the peculiar challenges arising constitute a perfect ground for developing tailored AI-based solutions. Notably, the notions presented in the course can be easily applied to other fields: indeed, as cultural heritage represents a complex domain, it ends up being a valuable and compelling benchmark to assess general learning paradigms.
To sum up, this course will provide an introduction to both task-specific algorithms (e.g. style recognition) and general-purpose learning paradigms (e.g. self-supervised contrastive learning) that can be applied to tackle the challenges posed by the varied, multimodal data from an interdisciplinary field such as cultural heritage.
The course evaluation will consist of a short paper presentation delivered by small groups of students.