Information Technology Reference
In-Depth Information
Talking Topically to Artificial Dialog Partners:
Emulating Humanlike Topic Awareness
in a Virtual Agent
Alexa Breuing and Ipke Wachsmuth
Artificial Intelligence Group, Bielefeld University, Bielefeld, Germany
{
abreuing,ipke
}
@techfak.uni-bielefeld.de
Abstract.
During dialog, humans are able to track ongoing topics, to detect topi-
cal shifts, to refer to topics via labels, and to decide on the appropriateness of po-
tential dialog topics. As a result, they interactionally produce coherent sequences
of spoken utterances assigning a thematic structure to the whole conversation.
Accordingly, an artificial agent that is intended to engage in natural and sophisti-
cated human-agent dialogs should be endowed with similar conversational abili-
ties. This paper presents how to enable topically coherent conversations between
humans and interactive systems by emulating humanlike topic awareness in the
virtual agent Max. Therefore, we firstly realized automatic topic detection and
tracking on the basis of contextual knowledge provided by Wikipedia and sec-
ondly adapted the agent's conversational behavior by means of the gained topic
information. As a result, we contribute to improve human-agent dialogs by en-
abling topical talk between human and artificial interlocutors. This paper is a
revised and extended version of [1].
Keywords:
Automatic topic awareness, Embodied conversational agents, Human-
agent interaction, Topic detection and tracking, Wikipedia.
1
Motivation
Topic awareness plays an important role in human conversations. Besides resolving lin-
guistic references and ambiguities which often arise in natural language talks, it enables
the interlocutors to interactionally produce coherent sequences of spoken utterances.
More precisely, every spoken contribution may raise new potential topics whose actual
realization depends on the co-participant's acceptance by picking up one of these topics
within his or her reply [2]. Hence, a topic can be described as a
joint project
[3] as it
is jointly established during ongoing conversations. Furthermore, being aware of top-
ics helps us to touch the right subject according to the social circumstances enclosing
the interactional situation. Assuming an everyday small talk conversation, for example,
so-called unsafe topics such as religion and death, should be avoided [4]. Altogether,
the competence to talk topically constitutes a basic requirement to carry on meaningful,
flexible, and appropriate conversations with other persons.
Embodied conversational agents (ECAs) are virtual characters possessing humanlike
conversational behaviors to establish an intuitive human-machine interface [5]. That is,
they are capable of holding face-to-face conversations with humans by understanding
