Information Technology Reference
In-Depth Information
Figure 2.
Prototype of the Kyoto digital city tour.
Ta b l e 1 .
Decision rule for prototype tour guide agent
Valence of Conversation Contents
Quantity of Talk
Negative
Positive
Low
medium length
long length
High
short length
medium length
ple decision rule (see Table 1). We plan to adjust the agent's decision-making
mechanism after examining its initial performance with tour-takers.
To make sure the tour stops long enough (but not too long) at each tour lo-
cation, the agent asks visitors (through dialog delivered by the Microsoft Agent)
to provide an explicit verbal cue that they are ready to go on. The agent moves
on to the next tour stop when it gets feedback from the majority of tour-takers
that they want to move forward. We believe this explicit request for feedback
serves two purposes: it allows the tour-takers to modulate the stop length far
more subtly than the agent could, and it gives them a feeling that the tour
(and guide) are adjusting to them - that they have an influence on events, and
needn't feel hurried or bored.
There is also a post-tour questionnaire that asks visitors who have taken the
tour to give feedback about the stories and the experience itself. As mentioned
