Thoughts on FML: Behavior Generation in the Virtual Human Communication Architecture
Jina Lee, David DeVault, Stacy Marsella, David R. Traum · 2008
We discuss our current architecture for the generation of natural language and non-verbal behavior in ICT virtual humans. We draw on our experience developing this architecture to present our current perspective on several issues related to the standardization of FML and to the SAIBA framework more generally. In particular, we discuss our current use, and non-use, of FML-inspired representations in generating natural language, eye gaze, and emotional displays. We also comment on some of the shortcomings of our design as currently implemented. 1. OVERVIEW In this paper, we discuss our experience developing multimodal generation capabilities within the ICT virtual human architecture. This paper is intended to contribute to an