Leveraging the latest digital capabilities to take historical research to the next level is certainly the goal of the HistKI project. Multimodal resources such as photographs, illustrations, 3D models, but also texts and descriptions serve as research material for a variety of subject- and theory-related investigations in humanities. Today, photos are essential for the digital reconstruction of a building or help to identify the building aspects of a certain ear. However, linguistic sources described objects and buildings long before the invention of photography.
Critical Assessment of Sources
The starting point of many research scenarios from architecture, art history, and cultural studies is a critical assessment of the source supported by tools of the respective subjects, on which the further evaluations and analysis in the scientific context build. Related questions are: How do historians and other scholars find and evaluate multimodal sources? What are connected generic procedures and issues? How can steps and procedures be promoted with computational approaches? How do artificial intelligence techniques affect the humanities research process?
Artificial Intelligence and Machine Learning
Different questions will be investigated using selected scenarios in which images, texts and 3D models connected to an object will be used together for analysis. With the help of machine learning, object sources and text sources (e.g.: captions) will be linked in order to allow a detailed contextualization and location of photographs exceeding the mere of distant viewing. Recognizing and contextualizing entities in text and image sources provides new insights for multimodal approaches in computational linguistics, which so far have had little focus on more abstract descriptions (such as perspective).