In this talk I will focus on the use of NLP to automatically link information from plain text to a knowledge base, a task which can benefit digital history projects aimed at reconstructing paths, events and trajectories from the past starting from archival documents. I will present some examples, where we showed that linking persons’ mentions to DBPedia can be effectively used not only to enrich the original text with structured knowledge, but also to infer information about the mentions that do not have a DBPedia entry. Trying to do the same with event mentions, however, poses a number of additional questions. Indeed, from a technological point of view, linking tools can recognize historical events with a similar accuracy of other named entities. However a general definition of what an event is does not seem to hold in different contexts and an application-based operationalisation may be more appropriate for digital history projects.
This online lecture will be held on Zoom. It is part of the Linked Pasts 6 conference. Separate booking is required only if you have NOT booked to attend the whole conference.
A booking for the Linked Pasts 6 conference can be made here.