Improving reading comprehension for hearing-impared students using Natural Language Processing

The main objective of this paper is developing a system, known as SimplexEduReading, capable of transforming educational texts in Spanish into very easy reading texts by using Natural Language Processing (NLP). This system is mainly oriented to support people with problems in reading comprehension, for instance, deaf people. The process of simplification and enrichment of texts consists of the automatically detection of linguistic features of input texts and: a) the reduction and removal of obstacles, but preserving in all cases the original meaning of the text, and b) the enrichment of texts using different open tools. To achieve our aim, at this point different NLP tools will be used: 1) OpeNER in order to detect name entities. Once detected, our system will provide extra information about them, for example, related images, information from Wikipedia or synonyms; 2) TERSEO in order to detect and resolve temporal expressions giving also a chronological timeline of the events; and 3) Wiktionary in order to clarify complex words. As further work Wordnet Domains will be used in order to detect the main domains of the text and therefore providing a context of the text to the reader.