Domain-speciﬁc query translation for multilingual access to digital libraries
Jones, Gareth J.F. and Fantino, Fabio and Fuller , Marguerite and Newman, Eamonn and Zhang, Ying (2009) Domain-speciﬁc query translation for multilingual access to digital libraries. In: 1st First Natural Language Processing for Digital Libraries (NLP4DL) Workshop, 15 June 2009, Viareggio, Italy.
Full text available as:
Accurate high-coverage translation is a vital component of reliable cross language information access (CLIR) systems. This is particularly true of access to archives such as Digital Libraries which are often speciﬁc to certain domains. While general machine translation (MT) has been shown to be effective for CLIR tasks in information retrieval evaluation workshops, it is not well suited to specialized tasks where domain speciﬁc translations are required. We demonstrate that effective query translation
in the domain of cultural heritage (CH) can be achieved by augmenting a standard MT system with domain-speciﬁc phrase dictionaries automatically mined from the online Wikipedia. Experiments using our hybrid translation system with sample query logs from users of CH websites demonstrate a large improvement in the accuracy of domain speciﬁc phrase detection and translation.
Archive Staff Only: edit this record