No abstract

No abstract

This paper proposes the comprehensive architecture of an integrated narrative generation system to discuss main two methods which are the foundations for integrating a variety of modules or elements and show a preliminary version utilizing our existing programs directly for seeing the picture of the integrated system. After we propose the macro structure of the proposed integrated system, we describe two fundamental methods to generate narrative structures and controlling the execution. And we show a pilot or preliminary version for the system developed by a kind of "bricolage" of existing programs corresponding to each module in the narrative generation process. Last, we discuss some issues for building the next version of the integrated narrative generation system. Moreover, this architecture is also differentiated from other many systems and approaches in the interdisciplinary concept of AI and narratology.

For a mechanism of the surface expression phase in the authors' narrative generation system architecture, this paper proposes an approach to the circulative generation by the mutual transformation between music and narrative. In the concept for musical mechanism, music is not just the accompaniment for narrative and the authors extremely aim at using musical knowledge & techniques and conceptual knowledge & techniques in fusion by the medium of music. Depending on this concept, in this paper, the authosr introduce the newest prototyping system composed of eight routes of transformation modules between narrative and music including automatic composition and musical variation mechanisms. In addition, through the simple preliminary evaluations, the authors consider several insufficient points in the quality of music and the correspondence of music to narrative.

No abstract