Towards making automated multimedia communicate more effectively: A multi-disciplinary approach

For over a decade, research has been conducted into artificial intelligence systems that have the ability to automatically generate multimedia presentations. Instead of drawing on discourses commonly involved in the creation of multimedia presentations, such as graphic design, the responsibility for insuring the communicative and aesthetic effectiveness of presentations generated by these systems has been restricted to a limited set of linguistic and print-centric discourses. As a result, users (and developers) are often disappointed that automatically-generated presentations lack much of the communicative coherence and multimodal aesthetic qualities of real-world multimedia. In this paper we show that defining a presentation