Risorse e applicazioni computazionali per l’accesso ai beni culturali: il Corpus CHerIDesCo

Abstract

The paper presents CHerIDesCo - Cultural Heritage - Italian Description Corpus, a domain-specific linguistic resource designed for the training and testing of novel NLP tools in the Cultural Heritage field. The corpus has been developed by the UNIOR NLP Research group as a part of the SMACH project, a three-year project funded by the National Operative Program to pursue the Smart Specialization Strategy defined by the EU. The project aims at improving language-based human-computer interaction in the Cultural Heritage domain through the development of innovative applications for multilingual access to the contents based on semantic language technologies. In particular, the paper describes the design of the CHerIDesCo corpus, the annotation procedures, and the platforms where the resource has been uploaded. As pointed out in the conclusion, this linguistic resource can be exploited in several NLP tasks (e.g., NER - Named-Entity Recognition, NEL - Named-Entity Linking, and Topic Modeling).