Information Technology Reference
In-Depth Information
Fig. 4.
Integration of our topic detection model into the existing architecture of the ECA Max
4.3
Integrating Topic Information
The complete system underlying the ECA Max is based on a multi-agent system com-
posed of several interacting agents. The conversational behavior, for example, is real-
ized via a dialog system in terms of an intelligent
dialog agent
. According to this, we
built up a
topic agent
implementing the presented processes and integrated this agent
into the existing system as shown in Figure 4.
The topic agent obtains every dialog contribution, that is the user's inputs as well
as the agent's outputs, and constantly provides up-to-date information about the cur-
rent topic situation of the ongoing dialog. It is directly connected to the dialog agent
due to interdependencies. More precisely, the interpreter of the dialog agent sends its
interpretation results to the topic agent which decides on the topical relevance of the
considered utterance on the basis of the identified modifier. That is, if an utterance is
specified as a greeting or farewell, the topic agent does not consider it as being topi-
cally relevant. Additionally, if one interlocutor proposes a dialog topic directly and the
interpreter specifies a rejection in response to this suggestion, the topic agent again as-
signs the proposed topic to irrelevant topics. To give an example, if Max says
“Let's
talk about music!”
and his human dialog partners answers with
“I don't want to talk
about music!”
, the topic agent does not identify “Music” to be the dialog topic even if
it was mentioned in two successive utterances.
The topic agent in turn sends the results of its topic detection process to the dialog
manager which has an impact on the conversational behavior of Max. For this purpose,
new dialog rules were defined allowing the agent to give information about the current
dialog topic, to wonder about sudden topic shifts (i.e. topic leaps) and to restrain the
search domain for the question answer component [16]. Moreover, the rules contained
in the knowledge base triggering or processing topic suggestions are topically arranged
to distinguish between their adequacies according to the given dialog setting. In the fol-
lowing, an example extract of the resulting rule library based on the agent architecture
JAM
[17] is given.
