This paper presents a system that supports human communication via computers by providing knowledge to users according to individual's background knowledge and the context of the conversation. The system takes account of the fact that a word has several meanings which depends on the context or user's specialty, so that it helps the listener to present some annotations adding to the speaker's speech information by presuming their background knowledge in the case of such different fields of communication. The system manages the background knowledge, users' profiles and conceptual spaces as users' personal information internally. The presented knowledge is individually tailored for the needs of each user using this personal information. Additionally, to help the listener understand the speaker, the system displays the words that are closely related to the topic during the conversation as they can realize the topic area. This paper describes the method and its application to a conversation between speakers from different fields.
[1]
Terry Winograd,et al.
Understanding computers and cognition
,
1986
.
[2]
Hitoshi Iida,et al.
A Japanese-to-English speech translation system: ATR-MATRIX
,
1998,
ICSLP.
[3]
Toyoaki Nishida.
Dynamic knowledge interaction
,
2000
.
[4]
Chris Schmandt,et al.
Nomadic radio: scaleable and contextual notification for wearable audio messaging
,
1999,
CHI '99.
[5]
Yasuyuki Sumi,et al.
Computer-Aided Thinking by Mapping Text-Objects Into Metric Spaces
,
1997,
Artif. Intell..
[6]
Daniel G. Bobrow,et al.
Beyond the chalkboard: computer support for collaboration and problem solving in meetings
,
1988,
CACM.