Information Technology Reference
In-Depth Information
It is also important to develop a better method to give feedback to a system
user. In prototype system, we give feedback by using visual information. This
method, however, is not effective when an ECA reflects the eye directions of the
user. The user cannot perform appropriate actions without feedback. On the other
hand, large feedback prevents the user from natural interaction. In addition, hu-
man behavior in cross-cultural communication is often expressed unconsciously.
We have to consider whether we had better to give feedback or not.
5.3 Target Interactions
The prototype CEBE which we are implementing will be able to detect head
direction, posture of the upper body and 3D angles of arms, and control a head
and arms of a robot or a virtual agent based on the detected data. By using
CEBE, we will be able to gather and analyze the information needed for real-
izing enculturated agent-human interaction. We can use the captured data for
unsupervised simultaneous learning of gestures, actions and their associations,
such as Embodied Interactive Control Architecture (EICA) [16].
6Conluon
6.1 Summary
In this paper, we proposed an environment in which a person can interact with
ECAs controlled by captured behavior of another person with cultural back-
ground, and we can analyze the data obtained in actual enculturated agent-
human interaction. In particular, we proposed a design and an implementation
of the system, CEBE.
In future works, we will focus on a negotiation situation in which people
have to express their opinions and speculate on the inner states of a commu-
nication partner, and we will conduct a preliminary experiment to confirm the
effectiveness of CEBE in an enculturated agent-human interaction. In the exper-
iment, we expect to reveal influences of cultural background in some situations,
such as detecting the degree of agreement or disagreement, commitment to the
communication, emotional representation, and how strong their assertion in the
communication.
References
1. Cassell, J., Sullivan, J., Prevost, S., Churchill, E. (eds.): Embodied Conversational
Agents. The MIT Press, Cambridge (2000)
2. Cassell, J., Tartaro, A., Oza, V., Rankin, Y., Tse, C.: Virtual Peers for Literacy
Learning. Educational Technology. Special Issue on Pedagogical Agents XLVII, pp.
39-43 (2007)
