Knowledge Portability with Semantic Expansion of Ontology Labels

Abstract

Our research focuses on the multilingual enhancement of ontologies that, often represented only in English, need to be translated in different languages to enable knowledge access across languages. Ontology translation is a rather different task then the classic document translation, because ontologies contain highly specific vocabulary and they lack contextual information. For these reasons, to improve automatic ontology translations, we first focus on identifying relevant unambiguous and domain-specific sentences from a large set of generic parallel corpora. Then, we leverage Linked Open Data resources, such as DBPedia, to isolate ontologyspecific bilingual lexical knowledge. In both cases, we take advantage of the semantic information of the labels to select relevant bilingual data with the aim of building an ontology-specific statistical machine translation system. We evaluate our approach on the translation of a medical ontology, translating from English into German. Our experiment shows a significant improvement of around 3 BLEU points compared to a generic as well as a domain-specific translation approach.

Polylingual Wordnet

“…Furthermore, we engage the idea of identifying relevant contextual information to support an SMT system translating short expressions, which showed better performance compared to approaches without a context. [55] built small domain-specific translation models for ontology translation from relevant sentence pairs that were identified in a parallel corpus based on the ontology labels to be translated. With this approach they improve the translation quality over the usage of large generic translation models.…”

Leveraging bilingual terminology to improve machine translation in a CAT environment

“…As for the integration of domain-specific parallel data such as dictionaries or bilingual terminology into an SMT system, three main strands of research have been explored in the past: Incorporating existing terminology within word alignment training (Okita and Way 2010), retraining additional in-domain parallel resources (Arcan et al 2012; Haddow and Koehn 2012) or adding new entries to the phrase table (Ren et al 2009). These approaches all allow the integration of domain-specific terms, but they require either switching-off the SMT system, which is unsuitable for our scenario or accessing prior knowledge to translate specific expressions.…”

Translating Ontologies in Real-World Settings

“…Due to the specificity of the ontology labels, just an n-gram overlap approach is not sufficient to select all the useful sentences. For this reason, we follow the idea of [11], where the authors extend the semantic information of ontology labels using Word2Vec 6 for computing distributed representations of words. The technique is based on a neural network that analyses the textual data provided as input and outputs a list of semantically related words [12].…”