Pivot, Box and Trilingual: Lexicon Extraction for Low-Resource Language Pairs with Extended Topic Models

Data-driven approaches to natural language processing have been shown to be greatly effective, and the case of bilingual lexicon extraction is no exception. While training data is readily available for many language pairs, many existing approaches fail for languages for which there simply does not exist parallel data. While there have been many studies on bilingual lexicon extraction, there has been little focus on the important problem of accommodating low-resource language pairs. We present a variety of solutions to this problem, demonstrating their application to a practical scenario, and compare their effectiveness to mainstream approaches. In this paper we develop pivot-based approaches for bilingual lexicon extraction using the framework of topic modelling [1]. Topic modelling has been a popular approach for bilingual lexicon extraction, however its use as a pivot model has yet to be explored.