For many years translation software has been making extensive use of natural language processing techniques. These techniques are used to develop features which help translators carry out their work. In this presentation we will describe a novel approach to post-translation analysis: MFTA (Multi-faceted Translation Analysis). It relies on a combination of natural language processing techniques such as stemming, lemmatization, pos-tagging, and parsing. It also involves the proprietary mechanism of computing word similarities between languages, called ILVS (Inter-language Vector Space). The applications of MFTA are seen in the following areas: -immediate post-translation post-processing (e.g. automatic placement of inline elements in translation, automatic quality assurance) - terminology analysis (bilingual terminology extraction and glossary creation) - translation memory analysis (sentence aligner, sub-segment matcher). The key idea behind MFTA is the integration of multiple heterogeneous natural language processing techniques to enable a wide range of useful features.
