Oliver Lemon, Alexander Gruenstein, Randolph Gullett, Alexis Battle, Laura Hiatt, and Stanley Peters
We explain generation methods in a dialogue system allowing humans and complex devices or applications to collaborate in real-time dialogue about ongoing activities in dynamic environments. The generation component must be able to handle contexts where there are multiple topics being co-ordinated by the conversation and where world and system-states can vary independently. We describe turn-management, truthchecking, "relevance"-checking, and the incremental message selection, aggregation, and generation methods employed in this context. We demonstrate that these techniques are viable in a demonstration dialogue system for multi-modal conversations with semiautonomous mobile robots.