Identifying word translations in non-parallel texts

Abstract

Common algorithms for sentence and word-alignment allow the automatic identification of word translations from parallel texts. This study suggests that the identification of word translations should also be possible with non-parallel and even unrelated texts. The method proposed is based on the assumption that there is a correlation between the patterns of word cooccurrences in texts of different languages.

The effect of bilingual term list size on dictionary-based cross-language information retrieval

“…The process can then be iterated to further improve coverage. Three variants on this basic idea have been tried by Rapp [16], Fung [5], and Picchi and Peters [14]. Every technique that has been tried relies on some way of estimating the importance of individual terms and then combining those estimates to evaluate term importance.…”

Automatic creation of bilingual dictionaries for Finno-Ugric languages

“…However, as foreseen by Rapp [7], "the availability of a large enough parallel corpus in a specific field and for a given pair of languages will always be the exception, not the rule", such corpora are still available only for the best-resourced language pairs. is is the reason of the increased interest in compiling comparable (non-parallel) corpora.…”

“…[7,8]), which consist of the following steps: building context vectors, translation of context vectors, and comparison of source and target vectors. ese methods need a seed lexicon which is then used to acquire additional translations of the context words.…”

A novel method for dictionary translation

“…The idea of lexicon translation with nonparallel corpora has been presented in Rapp (1995). The presented method is based on the observation that if the words a and b collocate often in one language then their translations should collocate in the repository in the target language too.…”