General agent-based architecture for collaborative dialogue management
Tomáš Nestorovič · 2010
In this paper, we focus on our agent-based approach to task-oriented dialogue management. We present our deliberation process based on optimizing the length of dialogue. As an uncommon feature, the manager accommodates a two-layered structure for representing the dialogue context, i.e., user's intentions detection and beliefs management. At the end of the paper, we suggest future extensions to the architecture.