Transcript Information and Communication Activities

Shifting ICT to ICA
- Towards Information and Communication Activity Navigation -
Hideaki Takeda
National Institute of Informatics
&
The Graduate University for Advanced Studies
[email protected]
http://www-kasm.nii.ac.jp/~takeda/
Hideaki Takeda / National Institute of Informatics
Table of Contents
Introduction: Shifting ICT to ICA
Our Research Topics
Kmedia: Finding relationship via WWW bookmarks
NMM:Re-configuration of personal human network
Community Navigator: Collaborative Scheduling Support System for Conferences
Social Scheduler: Collaborative Scheduler for Personal Resources
TelMeA: Expressive Media for Online Communities
TelMeA for e-Kyoshitsu: Application to Distance Learning
Summary
Hideaki Takeda / National Institute of Informatics
Shifting ICT to ICA
- Towards Information and Communication Activity Navigation -
H. Takeda
From “old computing” to “new computing”
“The old computing was about what computers could do; the
new computing is about what users can do. Successful
technologies are those that are in harmony with users’
needs. They must support relationships and activities that
enrich the users’ experiences.”
Ben Shneiderman, Leonardo's Laptop: Human Needs and the New Computing
Technologies, MIT Press, 2002

Paradigm shift is needed
 Technology-centered approach
 Human-centered approach
Hideaki Takeda / National Institute of Informatics
ART(Activity-Relationship-Table)
COLLECT
RELATE
CREATE
DONATE
(Information)
(Communication)
(Innovation)
(Dissemination)
Self
Family and friends
(2-50 intimates)
Colleagues and neighbors
(50-5,000 regular encounters)
Citizens and markets
(5,000+)
Ben Shneiderman, Leonardo's Laptop: Human Needs and the
New Computing Technologies, MIT Press, 2002
Hideaki Takeda / National Institute of Informatics
Information and Communication Activities

Two layers for our activities
Information layer only concerns explicitly represented and processed
information.
 Communication layer concerns potential information. Potential information
can be revealed through communication among people.
Communication Layer

Collaborate
Relate
Present
Information Layer
Collect
Create
Donate
Hideaki Takeda / National Institute of Informatics
Information Activities

A cycle of information exploitation
 Collect
 Find and retrieve information
 Create
 Process (classify, extract, combine, mix, …) information
 Generate new information
 Donate
 Publish and distribute information
Information Layer
Collect
Create
Donate
Hideaki Takeda / National Institute of Informatics
Communication Activities

A cycle of human relationship exploitation
 Relate
 Find and contact people
 Collaborate
 Work with other people (organized work, teamwork, cooperation, …)
 Present
 Identify and contribute ourselves to communities
Communication Layer
Collaborate
Relate
Present
Hideaki Takeda / National Institute of Informatics
Information and Communication Activity Navigation
(ICAN)

Support by computers for six categories of ICA
Community Navigator: Collaborative Scheduling Support System for Conferences
Social Scheduler: Collaborative Scheduler for Personal Resources
Collaborate
Relate
NMM:Re-configuration of
personal human network
Kmedia: Finding relationship
via WWW bookmarks
Collect
Present
TelMeA: Expressive Media for
Online Communities
TelMeA for e-kyoshitsu
Create
Donate
Hideaki Takeda / National Institute of Informatics
Discovery of Shared Topics Networks
among People
A Simple Approach to Find Community Knowledge from WWW Bookmarks
H. Takeda, M. Hamasaki, T. Matsuzuka, Y. Taniguchi
Purpose


Generation of human network guiding individual information
activities
 An example
 I want to watch sports programs on TV. What your
recommendation?
 Who and What
Shared Topics Network among Users (STN)
Hideaki Takeda / National Institute of Informatics
Bookmarks as Knowledge

= A topic interested by the user
User
Folder
(topic)

2
1
a
g
URLs in a bookmark folder
= Examples of the topic
x
WWW page
(information
on the topic)
A bookmark folder
b
Shared Topics Network
c
f
5
z
4
e
d
3
y
Bookmark (person’s interest)
Hideaki Takeda / National Institute of Informatics
Discovery of topic relations
Common relations
 (search, IR), (academia, research-related)
 similar but words themselves are different
 Un-common relations
computer-related
…(Unix, academia)
research-related
 Speciality of the
community
search

B
A
Information
retrieval
academia
C
Hideaki Takeda / National Institute of Informatics
Discovery of relationship among people


What are common topics with others?
Who is good at this topic?
computer-related
research-related
B
search
A
Information
retrieval
academia
C
Hideaki Takeda / National Institute of Informatics
Category Resemblance (1)
Categorization Is Human Relation?

Human relation can be measured by resemblance of folder structure
Nfij × Rfij Cij : Category resemblance
Cij =
Nfij : No. of recommended folders
Npij
Rfij : Folder relevance
Npij : No. of recommended pages
Folder structure is similar
Not similar
Hideaki Takeda / National Institute of Informatics
Effects of Category Resemblance (4)
for Page Recommendation
Better page recommendation results for new group made from
category resemblance (CR)

Score
2.8
2.7
Page Recommendation
New groups
2.6
2.5
2.4
2.3
2.2
All
In-community
Cross-community
2.1
2
Hideaki Takeda / National Institute of Informatics
Summary




Proposal of shared topic network to enhance user’s
communication
Proposal of algorithm of discovery of shared topic networks with
WWW bookmark files
Validity of our approach by an experiment
Proposal of category resemblance as measurement for community
effects
Hideaki Takeda / National Institute of Informatics
Re-configuration of personal networks
by the neighborhood matchmaker method
M. Hamasaki, H. Takeda
Purpose




Personal network is usually “ad hoc”
 We may miss better friends nearby
We need better network
One Solution:
 Collect data for all people, then generate the “best” network
 Disadvantage:
 Scalability
 Privacy
Our approach:
 Neighborhood Matchmaker Method (NMM)
Hideaki Takeda / National Institute of Informatics
Neighborhood Matchmaker Method (NMM)




A iterative approach to optimize the network
Every node works as a matchmaker for neighborhood nodes to
improve the network
The basic idea
 In our real life, introducing new friends by the current friends is a
practical way to optimize personal networks
 We can know persons who you have not known before
 Your friend can filter people for you
Advantages
 No need for central servers
 Applicable to any size of community
 Less computational cost
Hideaki Takeda / National Institute of Informatics
Algorithm



1. A node calculates connection values between its neighbor nodes
 We call that node “matchmaker”
2. If the matchmaker finds a pair of nodes which has a good enough connection
value, it selects this pair for recommendation. The matchmaker introduces both
nodes of recommended pair to each other
3. The node that receives recommendation decides whether it accepts or not. If
it accepts, it adds a path to the recommended node
!?
Good !!
matchmaker
matchmaker
Calculating connection values
matchmaker
OK
OK
Introducing each nodes
Adding a new path
Hideaki Takeda / National Institute of Informatics
Results: Cover-Rate w.r.t. Nodes


The path size is fixed as three times as the node size
All cases were converged
The average of cover-rate and the turn of convergence vary with the node size
1
20node
0.9
40node
0.8
0.7
60node
0.6
80node
0.5
Cover-Rate

100node
0.4
0.3
20node
40node
60node
80node
100node
0.2
0.1
0
0
50
100
150
200
Turn
250 300 350 400 450 500
Hideaki Takeda / National Institute of Informatics
Results: Average of Convergence Turn

The number of convergence turn is linearly increased with the
node size
Computational cost
 NMM: O(N)
 Central Server Model: O(N2)
600
5path/node
Convergence Turn

500
4path/node
400
3path/node
300
2path/node
1p/node
2p/node
3p/node
4p/node
5p/node
200
1path/node
100
0
10
20
30
40
50
60
70
80
90
100
The node size
Hideaki Takeda / National Institute of Informatics
Conclusion


Proposal of optimization of “ad hoc” network
Good news for the Internet communities
 No need for central servers
 Applicable to any size of community
 Anytime Algorithm
Hideaki Takeda / National Institute of Informatics
Community Navigator:
Collaborative Scheduling Support System for Conferences
H. Takeda, M. Hamasaki
In cooperation with Yutaka Matsuo and Takuichi Nishimura
Purpose




System Aim: Support people to find their friends in a specific group
Research Theme: Investigate different human networks in the same group
Three human networks
 Human network in the activity: I worked with him
 Human network by communication: I know him
 Human network by behavior: I meet him
Scheduling on conferences
 Plan and Action
Planning
“I worked with him”
action
“I know him”
“I meet him”
Hideaki Takeda / National Institute of Informatics
System Functions




Easy-to-use scheduling system for the conference
 Just add presentations what you want to watch
Can refer schedules of other people
 Manually collaborative scheduling
 Can only see schedules of who know you
Can recommend schedules
 collaborative scheduling
On-site support of schedules
 Small communication device with sensors
Cobit
Takuichi Nishimura, Hideo Itoh, Yoshinobu Yamamoto and Hideyuki
Nakashima. ``A compact battery-less information terminal (CoBIT) for
location-based support systems," In Proceeding of SPIE, number 4863B12, 2002.
Hideaki Takeda / National Institute of Informatics
Outlook
Hideaki Takeda / National Institute of Informatics
Recommendation


Recommendation by pattern similarity
Naïve collaborative filtering
 Paper recommendation
 Person recommendation
Recommendation by personal network
Reply on selection by friends
 Paper recommendation
 Person recommendation
Paper
1
Paper
2
Paper
3
Paper
4
Paper
5
Person A
1
0
1
1
1
Person B
0
1
0
1
0
Person C
1
1
1
1
0
Person D
0
0
1
1
1
Hideaki Takeda / National Institute of Informatics
Recommendation
Hideaki Takeda / National Institute of Informatics
Social Scheduler:
Collaborative Personal Task Scheduler
Ikki Ohmukai, H. Takeda
Social Scheduler:
Collaborative Personal Task Scheduler


Mobile Task Scheduler for Daily Life

Everyone belongs several groups and communities

Some groups were built emergently and have ambiguous boundaries

No one cannot manage “my” schedule
Collaborative Model of Personal Scheduling

Based on information sharing with one’s friends

Access control and filtering according to each group

Cell-phone application
Hideaki Takeda / National Institute of Informatics
Social Scheduler
Collaborative Personal Task Scheduler

Create friends network by authorization


Task conditions become viewable and updatable by
each other
Server merges all friends network into alliance network

The system considers partial complete graphs as
emerging groups (unit of information sharing/filtering)
Hideaki Takeda / National Institute of Informatics
Social Scheduler:
Collaborative Personal Task Scheduler

Experiment
12 subjects from various
communities
3 weeks in use
41.9 tasks/person
76 collaborative tasks

Social Network
The system can find multiple
communities around the
user
Distance of each friend can be
measured by the number of
groups they share
Hideaki Takeda / National Institute of Informatics
TelMeA
Show Me What You Mean Expressive Media for Online Communities
Toru Takahashi, Yasuhiro Katagiri, H. Takeda
TelMeA2002
Hideaki Takeda / National Institute of Informatics
Conversation Process in TelMeA2002
Hello
!
1. Edit a message
in terms of a
script language
Editor
Participant of a
TelMeA Community
Personified
Media
2. Submit the message
to the community
server
Conversation Log of the
TelMeA Community
6. The message is
asynchronously
enacted
Other Participants
of the Community
Hello!
Screen Shot of
TelMeA2003
4. Request for seeing the
message from others
5. The required message
is downloaded
3. The massage is accumulated
in the conversation log
Hideaki Takeda / National Institute of Informatics
Our Goal

Find pragmatic rules of social and nonverbal interactions
 Supporting social and nonverbal interactions
 Archiving the logs of long-term community activities
 Analyzing usages and effects of nonverbal expressivity



Make a model of multimodal social interaction
Calculate social evaluations for involved information
Summarize or make reutilize the involved information
Hideaki Takeda / National Institute of Informatics
e-kyoshitu(e-classroom) Project
Toru Takahashi, Yasuhiro Katagiri, H. Takeda
Trial Use: e-教室(e-classroom) Project


e-教室(e-classroom) Project:
 Run by NPO
 Distance learning for children (mainly junior-high school, 1215yrs)
 Several classrooms (math, economics, CG, etc)
TelMeA for e-教室
 Experimental use of TelMeA
 Classroom for
 Leaning “agent” as new technologies by using
 Communicating to each other (“BBS” for participants)
 (demo)
Hideaki Takeda / National Institute of Informatics
TelMeA for e-教室
Hideaki Takeda / National Institute of Informatics
The current status of “TelMeA for e-教室”




Period: c.a. 4 month (2003.1.16-)
Login users: 64
Posted users: 24
Post No.: 297, Post thread No.: 22
Hideaki Takeda / National Institute of Informatics
Summary


Information technologies, in particular AI can offer new
opportunities for communities
 Reducing constraints of the real world
 Time, space, etc
 new communication ways
 Knowing new related people, communication via agents etc
They will change meaning or roles of communities
 e.g,
 Very weak communities
 Quick life cycle of communities
 Belonging so many communities
Hideaki Takeda / National Institute of Informatics
Summary

Challenges
 Support of life cycle of communities
 Create, maintain, diverse, merge, disappear
 Trust
 Trust is very difficult
 Trust may be more complicated than the real world…
Hideaki Takeda / National Institute of Informatics