Linguist’s Ambiguity Tutor and Rehearsal System (LATARS)

BCL proposes developing a Linguist’s Ambiguity Tutor and Rehearsal System (LATARS) that will disambiguate idiomatic Arabic, Korean and Urdu text. The Phase II research will be built on BCL’s Phase I investigation into possible methods for developing the LATARS. In its Phase I research, BCL’s research indicated that, for languages like Arabic, the automated translation capabilities were severely limited. Similarly, developing the LATARS was hindered by the lack of an adequate parser, an electronically accessible lexicon and lack of formal syntactic or semantic rules. In its Phase II work, BCL, building on its Phase I work, will develop the LATARS using a COTS part of speech (POS) tagger. This LATARS will include a method for semi-automatically developing an electronic lexicon, and generic syntax and semantic rules. This approach will be tried primarily with Arabic but also applied to two languages that are associated with WMD developing countries, namely Korean (N. Korea) and Urdu (Pakistan). This will develop a methodology that can be applied by the DoD to develop translators for other languages in the future.