FP7-funded project which will deploy and develop language technology to automatically detect and remove obstacles to reading comprehension for people with Autism Spectrum Disorders.

Project details:

Autistic Spectrum Disorders (ASD) are neurodevelopmental disorders characterised by qualitative impairment in communication and stereotyped repetitive behaviour. ASD are serious disabilities that affect about 60 people out of every 10000 in the EU. People with ASD usually have language deficits with a life-long impact on their psychosocial functioning. These deficits are in the comprehension of speech and writing, including misinterpretation of literal meanings and difficulty understanding complex instructions (Minshew and Goldstein, 1998). In many cases, people with ASD are unable to derive the gist or meaning of written documents (Nation et al., 2006; O'Connor and Klein, 2004; Frith and Snowing, 1983). Difficulties in reading comprehension that ASD cause represent a significant barrier to inclusion. People require access to written material for many purposes and in many contexts, from searching for employment opportunities and obtaining information to support their education to communicating by email or learning about local entertainment or news. Several studies have indicated a link between reading comprehension, (and more generally, literacy) and access to education, employment, culture, and communication (OECD, 2000; Brugha et al., 2007). In healthcare too, the empowerment of patients depends on their ability to understand information about potential medical treatments, and make informed decisions about interventions. Thus, there is considerable evidence that European citizens with ASD (of whom there are estimated to be more than three million) are at risk of exclusion from the opportunities available within the information society. The benefits of inclusion to people with mental health problems, including ASD, have been described in studies by Fryxell and Kennedy (1995) and McDaid (2008), among others.

Recent developments in Language Technology (LT) have enabled the processing of natural languages. These technologies can analyse language structures (Klein and Manning, 2003) and their meanings (Navigli, 2009), and also generate useful/meaningful text (Liben-Nowell, 2000; Max, 2005; Yanguas, 2009). These advances have brought many benefits to the EU, of which machine translation is an example. Nevertheless, very little effort has been made to apply these technologies for the benefit of people with ASD. We aim to improve this situation, and to develop and deploy LT to address, and thus alleviate the aforementioned language deficits of people with ASD. More specifically, LT will be used to reduce and/or remove structural complexity and ambiguity in meaning in texts. Additional content (such as illustrative pictures, concise summaries of documents, pre-reading questions on related concepts, and document navigation tools, which are shown to have a role to play in improving the comprehensibility of written documents (O'Connor and Klein, 2004; Clay, 1991) can also be generated. At present, LT are mature enough to provide a sound basis for the development of such components (to automatically detect and remove obstacles to reading comprehension and to generate additional content to facilitate reading comprehension). This, in turn, will improve the ability to access to written information by people with ASD, thereby reducing their risk of exclusion.

Wolverhampton team
Publications by the Wolverhampton team
(c) 2006 - 2014 Research Group in Computational Linguistics
Last modified: March 20 2017