Towards semantic alignment of heterogeneous structures and its application to digital humanities

Abstract

Different variants of the notion of ‘alignment’ have been adopted in a range of areas, focusing on homogeneous structures (e.g., text alignment [8], database alignment [1] or ontology alignment [4]) or heterogeneous structures (e.g., annotation of text with on- tologies [3], alignment of dictionaries and ontologies [2], alignments between relational databases and ontologies [9]). These alignment approaches, however, take little account of the alignment of multiple structures. This type of approach is becoming increasingly necessary to manage the growing volume of unstructured information sources available on the Web (encyclopedias such as Wikipedia, social media data, etc.) and LOD knowl- edge bases. In addition, the approaches are mostly developed for the English language. These needs have to be addressed through a global vision of alignment that takes into account a multiplicity of structures in which knowledge can be expressed. This paper seeks a holistic approach to semantic computing and alignment, when considering het- erogeneous structures in which knowledge is represented.