Focus
The Natural Language Processing and Accessibility group conducts research on several areas related to adaptable, relational and cognitive software environments.
The main focus of the group is in the area of representation and manipulation of linguistic information. The contributions are in two directions: on the one hand, computational linguistics and ontology based techniques are adopted for written and oral contents analysis, on the other hand, iconic and nonverbal languages adopted for complex pattern recognition are investigated.
Automatic summarisation, character extraction, text readability evaluation, semantic text indexing, advanced spell checker, information extraction, prediction, dialog analysis, and emotional and affective skills detection, represent the most important research interests of the group in this area.
The group also focuses on HW/SW devices supporting persons with disability. Multimodality and virtual reality are used to design innovative and adaptable applications for users with congenital or acquired impairments (either physical, sensorial, cognitive or relational). The group examines special applications for learning, social life and therapy, to study software properties in critical contexts.
Finally, the group is working on user profiling and service customisation; an extended version of the World Health Organization (WHO) International Classification of Functioning, Disability and Health (ICF) has been defined, permitting user interface customisation (e.g. simplified, screen reader friendly interface for blind students), content adaptation (e.g. simplified texts for students with dyslexia), content transcoding (e.g. the translation of text messages into voice by means of an ASR tool), personalised campus navigation (e.g. how to find the route to a given room that does not contain architectural barriers for students with motor impairments).
Most relevant research achievements
DIESIRAE, a semantic powered search engine has been developed, based on a proprietary model exploiting MaxEnt and HMMs.
KeaKI, a tool for automatic text summarisation has been produced; it is able to extract the principal actors of a given text, as well as to compute a summary and a conceptual map.
SPARTAA, an advanced text complexity estimator, exploiting both statistical and grammatical information, has been developed.
In collaboration with the AI/BCI group, a predictive speller for Brain Computer Interface (BCI) applications has been produced.
ICF*, an extended version of the WHO ICF, has been defined, and a web based survey is ongoing.
FLUtE, a C# Fuzzy Logic framework, developed for a master thesys, is available at
http://www.flutehomepage.com/ and is maintained by the author under LGPL.
