The use of artificial intelligence in the assessment of live subtitling quality: the NER Buddy

Abstract

Translation quality assessment has been subject to high levels of subjectivity. However, in areas such as audiovisual translation it has become common practice to objectively evaluate the quality of the captions of live TV broadcasts. In intralingual live subtitling — an accessibility service for people with hearing loss where captions are in the same language as the original — the NER model was proposed by Romero-Fresco and Martínez (2015). However, it is complex and time-consuming. The purpose of this contribution is to present the results of our research on the development of an AI-based application for the (semi-)automatic assessment of live captions using the NER methodology. International TV broadcasters are testing this app.