Supporting narrative understanding of children with autism: A story interface with autonomous autobiographic agents

Abstract

Recent studies in education for children with autism demonstrated that computer assisted learning can increase children's narrative understanding. However, software interfaces which include autonomous intelligent agents have yet to be explored in this context. In this paper we investigate how a narrative storytelling environment which is populated with virtual characters may help children with autism to understand narrative structure and to remember meaningful events within in a story. The behaviours of characters in the narrative storytelling environment are driven by a cognitive agent architecture and, in particular, a specially designed computational autobiographic memory (AM) which allows the virtual characters to remember and express emotional experiences. An application named virtual agent story interface (VASI) has been developed to visualise the agents' AMs and to help the children to reconstruct their own version of a story. In this paper we describe related research in the area, the overall design rationale of VASI, details of the VASI software design and the preliminary evaluation study.