Date: 23 January 2014
In this seminar I will explain the work done so far for my doctoral thesis, which focuses on the development of tools to support people with dyslexia mixing Natural Language Processing (NLP) and Human Computer Interaction (HCI).
First, I will motivate the problem presenting some results that estimate the presence of dyslexia in the Web and in social media.
Then, I will detail how we used HCI evaluation techniques — eye-tracking, questionnaires and interviews — to find out which text presentations and NLP strategies are beneficial for people with dyslexia. These experiments explore different factors regarding the readability and understandability of a text. These factors are: (1) the text presentation, (2) its lexical quality, (3) the inclusion of graphical schemes in addition to the text, (4) the application of lexical simplification strategies, (5) the use of verbal paraphrases, (6) the presence of different numerical representations and (7) the use of highlighted keywords in the text.
Finally, I will show the applications developed so far based on the results of the previous experiments such as an eBook reader, a web service, or a game for children with dyslexia. *(the later is not part of my PhD but profits from these results)* My thesis is being co-supervised by Ricardo Baeza-Yates and Horacio Saggion and includes several collaborations with other people.