Kannada Kannada Paraphrasing

Abstract

Natural language processing (NLP) is one of the fields in computer science that utilizes newlinemachine learning and artificial intelligence techniques to process various languages. It can be newlineused to translate sentences from one language to another. It is an emerging area of computation newlinewhich facilitates easier and refined communication in different languages. newlineKannada language belongs to the Dravidian family of languages, characterized by a rich system newlineof morphology resulting in long and complex word forms. newlineSemantic similarity evaluation is pivotal issue in Natural Language Understanding, wherein main newlineobjective is to identify whether input data units are semantically similar. The main challenges newlinefaced in semantic similarity evaluation are ambiguity and rich variability of natural language newlinetext. The two text units, which are lexically and structurally very different, may convey same newlinemeaning. For given sentence semantically similar sentence can be generated by the following newlineapproaches: newline1. Re-ordering of words/phrases (lexical transformations) newline2. Re-arranging the grammatical structure (syntactic transformations) newline3. Replacing the words with their synonyms (Semantic transformations) newlineParaphrasing is a restating the same information using different words. It is not just a dictionary newlineexercise; however it should attempt to alter the information order and also grammatical structure newlineThis research work focused on certain type of paraphrases, structural and lexical paraphrases, newlinewhich are carried out by referring to the linguistic knowledge but not to the communicative newlinesituation.