Page 112 - Socially Intelligent Agents Creating Relationships with Computers and Robots

EgoChat Agent 95

that members in a general community have the same inclination, and that this
inclination represents the relationships of community members.
The EgoChat system consists of three storytelling processes and two storage
processes (Figure 11.1(b)). In the storytelling processes, community members
can share their knowledge by creating a knowledge stream with their VEs.
When a user wants to some community knowledge, she inputs her interest into
EgoChat by voice [storytelling process (a)]. The voice messages are recognized
by a commercial speech recognition system, and then VEs start talking about
related topics [storytelling process (b)]. VEs will continue talking with one
another unless the user interrupts the conversation to make a comment or change
the subject [storytelling process (c)].
As the community member talks with VEs, her personal memory is enriched
intheprocessesof storing comments. BeforeusingEgoChat, thepersonal mem-
ories are stored in VEs by using automated summarizing technology or summa-
rizing humanly from past exchanges on electronic media such as mailing-lists
[storage process (1)]. Besides text-based messages, VEs store new oral mes-
sages from a user on EgoChat [storage process (2)]. In this way, VEs share
community knowledge in past messages of community members and users add
new ideas to their virtualized-egos in a loop of community knowledge creation.

3. Storytelling by VEs

VEs generate messages in turn, and these messages are ordered as a story as
follows.
1. Start a turn. taking turns
2. Each VE selects a message associated with a topic from personal memory.
3. Only a VE whose message is presented in the most orderly and most reason-
able way in the context speaks the message.
4. The turn ends, and the next begins.
By repeating the above process of generating and selecting messages, VEs
can tell a story.
In the following three sections, we propose two sets of representations of
the personal memory that characterize a personality of a VE as an agent of
a community member; and lay out how an appropriate VE is selected during
process 3.

3.1 Topics-and-summaries representation

Each VE has a set of topics. Past posts from a community member related
to each topic are filed away into the personal memory of a VE. For example,
when a VE in a community of liquor fans has topics such as brandy, beer and
sake (Figure 11.2(a)), comments about brandy like “I used to drink V.S.O.P”