An Approach to a Knowledge Reconstruction

Engine for Supporting Event Planning

Shigeki AMITANI, Mikihiko MORI and Koichi HORI
{amitani, mori, hori}@ai.rcast.u-tokyo.ac.jp
Research Center for Advanced Science and Technology (RCAST),
The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo, 153-8904, JAPAN
Tel: +81-3-5452-5289 Fax: +81-3-5452-5312

Abstract. The main goal of our current research is to establish a methodology to articulate the gap
between the event designers' intention and the event visitors' mental impression, which is the first
step for "a knowledge reconstruction system''. At the actual events, the visitors' interactions with the
event objects were observed and the verbal reports (protocol data) were recorded. From analysis of
these data, a lot of gaps between the planners and the visitors are found as well as unexpected mental
processes. This analysis provides "chances'' to create new ideas with planners. In this paper, we are
going to propose "a supporting system for knowledge reconstruction'' as a tool for "Chance
Discovery'' by applying the framework of creativity support to the actual event designing work.

1. Introduction

Every year, event-planning companies hold various events. So far event planning is conducted
with implicit knowledge of the experienced planners and the visitors' impression of the events is
usually measured only by questionnaires. In the actual situation, it is said that the planners cannot
obtain adequate and proper knowledge for the future planning from the statistical data from
questionnaires because the data is information "without context where knowledge was produced".
That means the planners are unable to evaluate the event they designed. Planners need to know
what visitors to the event actually feel and how they behave when they are at the event booth to
construct the strategies for next event planning. This is because knowledge itself cannot be
applied to the real situation if it does not have the context where it was produced [1]. Qualitative
data and quantitative data should complement with each other.
In this paper, we are going to propose a methodology to articulate gaps between planners'
intention and visitors' impression. Our methodology was applied to the actual events, -"Tokyo
Motor Show 2001'' and "World PC Expo 2001'', in cooperation with Dentsu Inc. The visitors'
interactions with the event objects were observed and the verbal reports (protocol data) were
recorded. The results of this experiment and, furthermore, the image of "a knowledge
reconstruction system" is described. From analysis of the data, we obtained the prospect that our
microscopic and detailed approach is useful and effective toward event planning in the real world.

2. What is "Knowledge Reconstruction"?

Though the importance of knowledge has been claimed since middle of the 1980s, the main
concern of business theory is how to obtain and accumulate established knowledge. Little
research has been conducted on how innovative knowledge is created [2]. Nonaka [2] claims
that there are four modes of knowledge transition: Socialization, Externalization, Combination
and Internalization. Though a lot of companies have attempted to apply this theory to their actual
works, it does not seem to be successful. Nonaka theory is proposed as a theory and no method
has been indicated. In addition, this is the theory for "transition of knowledge mode", not the
theory for "manipulation of knowledge mode by the user". In our research we are going to apply
Nonaka's theory to the actual situations. We are going to propose a methodology and a system
to support to apply this theory to the actual situations (Fig 1). The italic phrases are what we are
going to provide.

Fig 1 A Cyclic Process of Knowledge Reconstruction

Our system aims to "use in the real world". Knowledge cannot be separated from actual
contexts to utilize the knowledge, i.e., knowledge management. That means that it is necessary to
preserve knowledge together with "the real context". Event planners need to know how the
knowledge was produced. It is necessary for planners to obtain and understand knowledge, to
integrate and create innovative knowledge, and to apply knowledge to actual event planning. This
is what we call "knowledge reconstruction". Knowledge must be dynamically integrated and
innovated. We are going to propose "a supporting system for knowledge reconstruction". At the
current state, our research obtained such perspective that our approach can be driving force of
this cyclic process. In the following sections, the experiments at actual events and an example of
findings are described. Moreover, the example that the planner hit upon an innovative idea with
providing the results of our investigation is also described. The system image is described in
section 4.

3. Experiment

The experiment of our methodology was conducted to extract the actual cognitive processes
at event sites of "World PC Expo 2001 (WPC: 19-22, Sept., 2001)" and "Tokyo Motor Show
2001 (Motor Show: 26, Oct. - 7, Nov., 2001)" held at Makuhari Messe in Japan, to see if our
methodology works. Three booths for WPC and one booth for the Motor Show were selected
for the experiment in cooperation with the event organizers.
To investigate what the designers' intentions are and how they implement the event objects to
express their intentions, we had interviews with the planners in advance.
To collect protocol data, two wearable computers were prepared1. Because of the sponsors'
intention, the normal digital video camera was adopted for collecting protocol data. 9 subjects
(one person for 3 sessions + one pair for 3 sessions) at WPC and 12 subjects (one person for
each session) at Motor Show were employed. Subjects are asked to look around the designated
booth(s). After visiting the booth, the procedure in the interview room is:

1
In cooperation with MIT Media Lab and Intelligent Cooperative Systems Laboratory at Research Center for
Advanced Science and Technology, the University of Tokyo. And a normal digital video camera was also prepared
1. Retrospective reports with visual aid the subject recorded: The subjects were asked to
report "what you look at'', "what you think about it" and "what you do" along with the VTR
as an memory aid.
2. Questions about the subjects' impression on the event objects: The questions were made
based on the interviews with the planners and the planning papers. This is to investigate how
the planners' intention and the visitors' impression match or mismatch with each other.
3. Keyword questionnaires: The keywords are also extracted from the interviews with the
planners and the planning papers. This is to investigate what keywords the planners
presented were impressive for visitors.

4. Results and Discussions

Though there are a lot of findings that surprised the planners, in this paper we are going to show
one example of a lot of unexpected findings that are beyond the planners' expectation at Motor
Show. These findings devote to create new knowledge, that is "knowledge reconstruction''. One
example of "the effect of the other visitors" at Motor Show is described.
The other visitors can provide a context that raises the degree of satisfaction of the visitor. The
following report was obtained:

A companion took a picture with a family. Both of the companion and the child
smiled. My (= the subject's) children also like cars. They would be delighted if I took
them here. That is a good idea.

This observed data was reported to a planner and he hit upon a new strategy:

By inviting families that are customers of the company, the other visitors will feel in a
way mentioned above. Moreover, the invited family will also feel better because they
feel "they are invited as special guests'' and this family can enjoy being a customer of
the company, which will be great benefit to the company, too.

This is a good example of "knowledge reconstruction''. We call it "reconstruction'' because

implicitly they might know the follows:
??A visitor is affected by another visitor at an event site.
??Customers are delighted if they are invited as special guests.
??If customers like the company, it is beneficial to the company.
But these pieces of information have not been connected. That means the knowledge obtained
through this analysis can support the event designing if it is properly reported to the planners. We
are going to propose a supporting system that provides "knowledge for strategies with real
context''.

5. Toward a Supporting System for Knowledge Reconstruction

As our investigation externalized a lot of knowledge successfully, we are going to propose the
supporting system for the transition among modes in Fig 1, i.e. knowledge reconstruction. Each
mode is supported in the following way:

Externalization: This mode is supported by the methodology we proposed in the former section.
Combination: From the data obtained with the analysis, this mode is supported by "exhaustive
search" and "presentation of contexts" Based on the conclusion of Hori and Yamamoto et al., the
spatial representation is adopted.
Internalization: It is expected that this system promote the user's understanding of the
knowledge.
Socialization: For example, "pursuasion of clients" is expected by using this system.
To accomplish this aim, our system should be implemented with the following features:

??Accumulating the knowledge: the system should accumulate the data of the analysis
??Browsing and reconstruct the knowledge: the system should present the knowledge with
the real context to promote and amplify the user's reflection [3][4]. And moreover, it
should foster the discovering and creating process of knowledge.

We adopt spatial representation to present data to the planner. Hori [ 5] conducted the
experiment for the effect of the spatial representation on the conceptual design. The system
named AA1 presents words, which represents the concept the user vaguely conceives, on the
two-dimensional space, the user can change the location of the words on the space. This action
helps the user clarify his/her concept gradually and such phenomenon was observed as the user
came to generate new concept by looking at the blank area on the space.

The image of the system is shown in Fig 2.

Fig 2 A Snap Shot of the System

The attribute set of "Object-VMT (Visitors' Mental Transition)-Action", i.e., the transition of
the visitors' cognitive process "what you percept" - "what you think" - "what you do" are
recorded together with the "real context" data. These attributes are composed of "unit object"
such as "family", "companion" and "unit cognitive process" such as "Remember", "FeelStarting"
are currently being defined to code the data [6][7]. The user can save the raw data with attaching
attributes tag onto it. The data can be refered by searching with the attribute tags or by full text
search. The user can add new attributes freely.
The expected interaction to reconstruct knowledge: The situation is that the planner would like
to browse the existing contexts to create an innovative one.

1. The planner selects attribute(s) from the lists. e.g.) "how could the visitors feel something is
starting?"
2. The system exhaustively searches contexts along with the selected attribute.
e.g.)"FeelStarting"
3. Then the system lists the contexts that contain the selected attribute tags, so that the planner
can refer to the real contexts. The related events are shown on the "Related Events" list and
the related contexts on the "Related Contexts" list (the third column of the left figure).
4. By clicking the button "Relate Context" (left most bottom of the left figure), the contexts are
extracted along with the units and all of the contexts (shown as ovals in the right figure) are
shown on the space. It shows the relations among contexts that the planner might not notice.
Moreover, it is expected that the planner hit upon innovative idea by looking at the blank or
combining the parts of existing contexts [8].
5. And it is also expected that the system automatically generate the new contexts by combining
the existing Objects, VMTs and Actions.

As for the automatic generation of the context, Ogata [9] developed a basic framework for
narrative conceptual structure generation. This is a system based on narrative techniques and
strategies. By changing the strategies, that is, tuning the strategies for a certain domain, it is
possible to generate a certain structures suitable for the domain.

6. Conclusion

In this paper, we proposed a real application of the methodology to investigate the cognitive
processes at event site toward "knowledge reconstruction system". The microscopic investigation
was conducted to understand the cognitive processes and their transitions. Through protocol
analysis, the following knowledge is found with "the real context", which has not been obtainable
from traditional statistical investigation.
??Effect of the other visitors (in our result, for example, "effect of a family")
??Effect and difference of understanding of congestion between designers and visitors
These knowledge with the real contexts are useful to establish new strategies, although the
statistic data lacks hence they are discarded. That is why we propose "knowledge reconstruction
system" that provides with the planners "knowledge with real context".
From now on, we are going to develop a knowledge reconstruction system and evaluate it at
the actual event-planning phase.

Acknowledgement
The authors gratefully acknowledge the generous assistance of Ms. Shoji, Mr. Shibata and Mr. Kanazaki
for experiments and discussions. With thanks to Ms. Ueoka in RCAST and Mr. Clarkson at MIT Media Lab
for wearable computers.

References
[1] Gerhard Fischer, Jonathan Ostwald: "Knowledge Management: Problems, Promises, Realities, and
Challenges", IEEE Intelligent Systems, Vol.16, No.1, pp. 60-72, January/February 2001
[2] Ikujiro Nonaka, Hirotaka Takeuchi: "The Knowledge-Creating Company : How Japanese Companies Create
the Dynamics of Innovation", Oxford University Press, 1995
[3] Schoen, D A: "The Reflective Practitioner: How Professionals Think in Action", Basic Books, NY, 1983
[4] Donald A. Norman: "Things That Make Us Smart: Defending Human Attributes in the Age of the
Machine", Addison Wesley Publishing Company, 1994
[5] Hori, K.: "A System for Aiding Creative Concept Formation", IEEE Trans. Systems, Man, and Cybernetics,
Vol.24, No.6, pp.882-894, 1994
[6] M. Suwa and T. Purcell and J. Gero: "Macroscopic analysis of design processes based on a scheme for
coding designers' cognitive actions", Design Studies, Vol.19, No.4, pp.455--483, 1998
[7] Shigeki Amitani, Koichi Hori: "Supporting Musical Composition by Externalizing the Composer's Mental
Space", Journal of Information Processing Society of Japan, Vol.42, No.10, pp.2369-2378, 2001
[8] Hori, K.: "A Model to Explain and Predict the Effect of Human-Computer Interaction in the Articulation
Process for Concept Formation", Information Modelling and Knowledge Bases, Vol.7, pp.36--43, IOS
press, 1996
[9] Takashi Ogata, Koichi Hori, Setsuo Ohsuga: "A Basic Framework for Narrative Conceptual Structure
Generation Based on Narrative Techniques and Strategies", J. of JSAI, Vol.11, No.1, pp.148-159, Jan.,
1996