It-Sr-NER: CLARIN compatible NER and geoparsing web services for parallel texts: case study Italian and Serbian
Progetto The main goal of the proposed project is development of the CLARIN compatible NER web service for parallel text with case study on Italian and Serbian, dubbed It-Sr-NER. Service could be used for recognizing and
classifying named entities in bilingual natural language texts. Input would be parallel texts expected to be TMX (Translation Memory eXchange) file, e.g. Sr-It. It-Sr-NER would recognize six NER classes: demonyms (DEMO), works of art (WORK), person names (PERS), places (LOC), events (EVENT) and organisations (ORG). Although primarily developed for aligned, parallel texts in TMX, the use of the service for monolingual text NER annotation for available spaCy NER models will be possible. It-Sr-NER uses a powerful Convolutional Neural Network architecture within the spaCy tool.