Towards a Language Independent Encoding of Documents: A Novel Approach to Multilingual Question Answering

Given source text in several languages, can one answer queries in some other language, without translating any of the sources into the language of the questioner? In this paper we try to address this question as we report our work on a restricted domain, multilingual Question - Answering system, with current implementations for source text in English and questions posed in English and Hindi. The cross - language functionality has been achieved by converting the queries and the documents to an intermediate representation, an inter-lingua called UNL. Built under the premise of a homogeneous language- independent encoding, the UNL can be used as the predicate knowledge base on which inferences can be performed effectively.