Non-verbal communication incorporating visual, audio, and contextual information are important to make sense of and navigate the social world. We propose a Computer-Based Training (CBT) regimen to train non-verbal communication skills for people with social and communication difficulties. We hypothesized that the Autism-Spectrum Quotient (AQ) is correlated with non-verbal communication ability, and use the AQ as a basis for our analysis. In previous work, we performed a factor analysis using AQ to determine the important non-verbal factors contributing to communication and social skills. As a result of the factor analysis, we defined non-verbal information on the axes of politeness/impoliteness (partner information) and intention & interest (intention information). Using this information, we developed an iPad interactive multimedia application called NOCOA to help users train their ability to recognize partner information and intention information. However, NOCOA uses only audio data, and this is potentially limiting because both audio and visual information are important in the real world. To help remedy this problem, in this work we recorded dialogue using the visual and audio modalities, and further proposed a training tool NOCOA+ that uses utterances in these modalities to help bring our training environment closer to the real world. We conducted two experimental evaluations with a total of 18 Japanese adults. The first experiment confirmed a significant relationship between AQ scores of the communication and socialization subcategories and non-verbal communication skills as measured by NOCOA+. The second experiment showed a score improvement for all subjects through systematic CBT.
Hiroki Tanaka, Nara Institute of Science and Technology, Japan
Sakriani Sakti, Nara Institute of Science and Technology, Japan
Graham Neubig, Nara Institute of Science and Technology, Japan
Tomoki Toda, Nara Institute of Science and Technology, Japan
Satoshi Nakamura, Nara Institute of Science and Technology, Japan
This paper is part of the ACE2013 Conference Proceedings (View)
View / Download the full paper in a new tab/window