The Computational Humanities (CH) group develops methods for analyzing cultural raw data such as images, video, and audio recordings with applications for the humanities. For this purpose, suitable machine-learning algorithms are developed and adapted in order to meet the special requirements of humanities research, which demands for multimodal, objective, robust, and interpretable methods. Applying these algorithms to large corpora allows for comparative analyses of different art genres. With this, the group is embedded into the Centre for Philology and Digitality (ZPD) and aims for establishing multilateral connections within JMU Würzburg and beyond.
A particular focus lies on the analysis of music audio recordings, contributing to research fields such as music information retrieval, audio signal processing, machine learning, and AI. This involves the creation of multimodal datasets and the development of suitable representations, which are robust, interpretable, and disentangled regarding the semantic dimensions of music. Applying the analysis algorithms to large music collections (corpus analysis) indicates the high potential of computational approaches for musicological research, thus serving as an example for exploiting cultural raw data for the digital humanities.