Articles in This Issue
Introduction
Now, we are right in autumn in Japan.
Japanese has many expressions about autumn; the season of appetite, reading,
colored leaves, and recreation are often used. Among many other expressions,
the season of harvest is specially impressive.
Witnessing the sowing of the seeds in spring, growing of crops through hot
summer, and the harvesting in autumn can bring a great joy to people.
Recently, we have been excited with the news that some ICOT OB projects have
been adopted into some R&D programs sponsored by the government.
I believe the success of these ICOT OB projects means that FGCS is now
producing a significant harvest.
In this issue, we would like to introduce the projects and achievements of
ICOT OBs adopted to government sponsored programs, and share our joy of this
"FGCS harvest " with you.
Let's start with the headlines of No. 14.
(Chie Takahashi)
AITEC NEWS Headlines
1. "KLIC Programming Contest" Report
We have closed the program submission for the KLIC Programming Contest and are
now on the final stage of evaluation. Some comments from applicants indicated
that this year's subjects were more difficult than last year's.
The Contest Committee had decided to make the subject matter more sophisticated
than last year so applicants could use unique KLIC functions more essentially.
However, subjects were a little difficult and we had less applicants than last
year. Considering KLIC dissemination, it is more meaningful to have more
participants; we will have to rethink the direction of the contest next year.
We are going to introduce the award nominees in the "KLIC Programming Contest
CYBER Ceremony" on the AITEC Homepage as soon as they are decided. We hope you
enjoy it.(http://www.icot.or.jp)
(Kiyoshi Kaneda)
(No Referenced Article)
2. ICOT OB Researchers Received Adoption in Governmental Sponsored Programs
Recently, the Ministry of Education, Science, Sports and Culture and the
Ministry of International Trade and Industry along with other related agencies
have increased the number of open R&D programs. The budget for each program is
enlarging from around 100 million to 300 million yen, which is considered
larger than average for a usual project budget for a university-level project.
The project term is 2 to 3 years, relatively short, and the themes can be
chosen by researchers, which is why this kind of program is getting popular.
The qualification for admittance depends on the programs. Some programs call
for projects from only university and national institute researchers, others
from groups of such researchers, and still others from groups of corporate
manufacturer researchers. Each program is highly competitive, with an
acceptance ratio usually around 10 percent.
In such competitive programs, several proposed projects from some groups
including ICOT OBs, ICOT related researchers of manufacturers and/or ICOT
related overseas researchers have been adopted. So far, we have been informed
that four projects were adopted for government sponsored programs. The
following is the project lineup in order of date to be adopted:
1) Prof. Kazuo Taki Group (Kobe University)
From: New Energy and Industrial Technology Development Organization (NEDO)
Title: A VLSI Design Automation Software for Low Energy Consumption
VLSI's Utilizing Parallel Processing
Term: November, 1995 - March, 1997
(Article No. 14-1)
2) Prof. Kazumasa Yokota Group (Okayama Prefectural University)
(Prof. Kazumasa Yokota moved from Kyoto University in April, 1997)
From: Ministry of Education, Science, Sports and Culture
Title: Research and Development on Agent-Based Database Systems
for Integration of Media and User Environments
in Scientific Research on Priority Areas
Term: April, 1996 -
(No Referenced Article)
3) Prof. Takashi Chikayama Group (University of Tokyo)
From: The Information-technology Promotion Agency, Japan (IPA)
Title: Development of Integrated Parallel Processing Technologies
Term: October, 1996 - February, 1998
(Article No. 14-2)
4) Prof. Ryuzo Hasegawa Group (Kyushu University)
From: The Information-technology Promotion Agency, Japan (IPA)
Title: Advanced Argumentation System Based on Theorem Proving Technologies
Term: October, 1997 - January, 1999 (tentative)
(Article No. 14-3)
These project themes are based mainly on products of the FGCS project and
further developed products. We at AITEC, who are disseminating FGCS technology,
are especially delighted with the success of ICOT OBs and related researchers.
We strongly hope that their success further increases and contributes to the
development of IT R&D and the IT industry.
In this issue, three researchers will introduce their projects.
(Edited by Shunichi Uchida)
3. The Second IFS Seminar at Keio University
On October 22 and 23, 1997, the second IFS Seminar for this fiscal year was
held at the Department of Science and Technology at the Yagami Campus of Keio
University. The seminar concentrated on KLIC, and had about 30 participants
from the university and outside.
We demonstrated the "multiple sequence alignment using a parallel iterative
aligner" by accessing a parallel machine at AITEC from notebook PCs through a
portable phone. In the demonstration, from time-to-time we indicated the
computed results returned from the parallel machine at AITEC which was
responding to the commands from the notebook PCs.
Since it was our first time to demonstrate in this way, we carefully prepared
and rehearsed for the demonstration.
At the seminar, we proved that we could do the demonstration with a narrow
bandline of a portable phone, and we had a lively discussion with the
participants.
The seminar schedule was as follows:
October 22 (Wednesday)
13:30 Explanation about IFS
Dr. Akira Aiba (AITEC)
13:50 IFS Demonstration of genetic information processing
Ms. Reiko Tanaka (IMS)
< break >
14:45 Introduction to KL1
Dr. Takashi Chikayama (Univ. of Tokyo)
< break >
16:45 KLIC Seminar
Mr. Yasuyuki Shirai (MRI)
October 23 (Thursday)
13:00 KLIC Programming (1)
Mr. Miyuki Koshimura (Kyushu Univ.)
< break >
14:30 KLIC Programming (2)
Mr. Tetsuro Fujise (MRI)
15:30 Exercise
16:30 Closing
If you are interested in the IFS seminar, or in hosting the seminar, please
send mail to:kaneda@icot.or.jp, or fax to: +81-3-3456-1618 (Kaneda)
(No Referenced Article)
4. The Research Funding Program in FY'97
The projects for the AITEC research funding program in FY'97 for
parallel/knowledge processing software have been decided.
For the FY'97 program, we extended our call for proposals to overseas
universities hoping more researchers than last year would use this opportunity
to conduct software R&D.
As a result, we received 37 proposals from 35 researchers from 6 countries
including Japan. The projects were judged, as in the FY'96 program, by review
from an outside refereeing committee. It was very competitive, and the
committee members had careful discussions to decide the projects.
Finally, 19 projects were adopted from 6 countries: Japan, USA, UK, France,
Austria and Australia. We are delighted with such a variety of countries
participating in this program, and we are looking forward to their
achievements.
Also, 3 projects were selected as continuing projects from FY'96. In total,
22 projects have started as funded projects for FY'97.
For details of each project, please look at the AITEC Homepage.
(http://www.icot.or.jp)
(Akira Aiba)
(No Referenced Article)
5. KLIC-3.001 Released and Mirror Site Opened at University of Tokyo
KLIC-3.002 was released. This is essentially the bug fixed version of
KLIC-3.001.
The file is available at the following URL:
ftp://klic.icot.or.jp/pub/klic/v3.0/klic-3.002.tgz (primary site)
klic.tgz
ftp://ftp.logos.t.u-tokyo.ac.jp/klic/klic-3.002.tgz(mirror site)
klic.tgz
/pub/klic/v3.0/klic-3.002.tgz (mirror site)
klic.tgz
It is noted that the distribution sites increased and changed from the
previous distribution sites. For people affiliated with Japanese universities,
it is recommended to get the file from the second site
(ftp.logos.t.u-tokyo.ac.jp) via SINET.
Please enjoy it.
Caution: Some serious bugs were STILL found in a shared memory parallel
version. The bugs will be fixed in the next version. Thus, if you are using
the previous (2.00x) version of shared memory implementation, DO NOT update
your program.
(Daigo Sekita - Mitsubishi Research Inst., Inc., KLIC task group)
(No Referenced Article)
6. AITEC Homepage Renewal
We have renewed the layout and structure of the AITEC Homepage. Because the
old AITEC Homepage had become very complicated and consequently had a long URL,
we revised the whole structure and improved some English pages. Most URLs of
the AITEC Homepage have been changed in this renewal, so please renew your
bookmarks of our homepage. We are sorry for this inconvenience, and if you
have any questions about URLs please contact us at:www-admin@icot.or.jp
(Akira Aiba)
(No Referenced Article)
Articles In This Issue
Report of the Success of ICOT OBs and Related Researchers in Recent
Government Sponsored Programs
Recently, besides the Ministry of Education, Science, Sports and Culture; the
Ministry of International Trade and Industry (MITI) and other related agencies
have started more R&D funding programs. The recently most attractive ones have
been programs based on national supplementary budget, especially the ones with
budgets over 100 million yen and are open to both universities and private
companies.
Regarding fields receiving funding, an Information Technology (IT) field is
considered more important and replaced the public works field. Funding for
universities and national institutes has higher priority than before, and a
part of the fund was utilized for IT R&D programs, a number of researchers and
engineers of universities and manufacturers had applied to them.
Although such open and competitive programs have been widely made in the US
and Europe, they had not been common in Japan. So those programs mentioned
above are kind of experimental programs in Japan which may change the
fundamentals of government sponsored R&D systems.
Several groups of ICOT OBs and related researchers applied for the programs.
As far as AITEC knows, 4 projects have been adopted including research effort
on the "Scientific Research on Priority Areas" as we mentioned in the headline.
Here, we introduce the projects' outlines and comments written by the project
researchers. We hope their experience will be useful to you.
We have asked 3 researchers to outline their projects as follows:
1) Prof. Kazuo Taki Group Project (Kobe University)
- Theme: A VLSI Design Automation Software for the Low Energy Consumption VLSIs with a Utilization of Parallel Processing
- Research Members: Kazuo Taki, Masahiro Numa, Hironori Nakajo,
Masahiro Yasugi (Kobe University)
- Term: November, 1995 -- March, 1997
- Outline:
It is necessary to develop VLSI chips of low energy consumption for
quickly advancing portable IT devices of which market is rapidly growing.
In order to develop such VLSI chips, a new logic circuit structure for
very low energy consumption and relevant technology on highly advanced design
automation software are essential.
The objective of our project is to research and develop highly advanced
design automation software to create digital LSI for a level of low energy
consumption which current CMOS technology has never achieved. The project
features two points of technical innovation.
The first innovative point is to make it possible to design LSI featuring
low energy consumption (power reduction ratio: 75% - 90% of present energy
consumption). We have proven that a new logic circuit (SPHL), revised
path-transistor logic, could nearly achieve our targeted low energy
feature. However, as we could not get sufficient results with the
manual design, we needed highly advanced design automation software systems.
In this project, to get the best results for our low energy consumption
feature, we tried to develop design automation software based upon the new
design method. The software unifies logic composition and layout to achieve
the first technical resolution.
The second innovative point is to develop the above mentioned design
automation software as parallel software in order to greatly shorten the
duration of LSI design. So far, it has been pointed out that a parallel
computer can hardly make the run-time of this type of software shorter.
However, in this project, we could successfully apply the FGCS technology to
achieve a rapid parallel processing ability, scheduling characteristic and
load balancing improvement for achievement of the second technical resolution.
This software outputs the results of LSI design, including all the
information necessary for producing it, and enables simple evaluation and
technology transfer.
The experimental LSI for evaluation was so far fully compatible with the
external specification with a market standard built-in 8-bit micro-processor
(Zilog-Z80), and the internal architecture was newly designed for low energy
consumption. We could design over 60% of the VLSI with this new design
automation software, and confirm its operation and performance. Although the
clock speed of the experimental LSI was just as fast as the other on the
market,the energy consumption was approximately one forth of the other
commercial product. This LSI is now being redesigned using the most advanced
process, and scheduled to be commercialized for portable IT devices, such as
portable phones, etc., from the Sharp Co.
For both low energy consumption LSI technology and design automation
software, we had to start from the very beginning. The pioneering spirits
we gained at ICOT moved us toward this challenge to develop this
significant opportunity on a new frontier. This challenge is regarded as a
mini-ICOT project for me, and as a result we have advanced hardware technology
and parallel software technology to support the hardware design in this
project because we did the same at ICOT.
2) Prof. Takashi Chikayama Group Project (University of Tokyo)
- Theme: Development of Integrated Parallel Processing Technologies
- Research Members: Takashi Chikayama, Hirokazu Murao (Univ. of Tokyo),
Hidetsune Kobayashi (Nihon Univ.)
Kazunori Ueda (Waseda Univ.), Naoto Niki (Science Univ.
of Tokyo), Yu Inamura (NEC Information Systems),
Tetsuro Fujise (Mitsubishi Research Institute)
Steve Gregory, Rong Yang (Univ. of Bristol)
- Term: October, 1996 - February, 1998
- Outline:
Most of the parallel computer systems so far in practical use have been
designed based upon both data parallel technology, and task parallel systems
have been limited only to experimental systems. For high performance parallel
processing, it is necessary to integrate both kinds of parallel processing,
complementing deficiencies and making synergy effects. This project is
proposing for streamlined integration of both data and task parallel
processing, providing design and implementation of language constructs for
synchronization and data passing.
To put it more concretely, the project uses a concurrent and parallel logic
programming language, KL1, developed in the FGCS project as a coordination
language, making data-parallel programs written in other languages such as
Fortran or C cooperate in a task-parallel way.
To develop this, we will extend KLIC, an implementation of KL1 for general
purpose computers developed in the FGCS follow-on project, with features to
handle arrays with distributed elements as a single array, to make automatic
redistribution of elements on demand, to call foreign language routines in
parallel, and to distribute such processing loads automatically. Upon this
implementation, we will construct experimental application software for
large-scale combinatorial optimization and large-scale symbolic formula
manipulation, so as to evaluate appropriateness and efficiency of the proposed
mechanism.
Many of the technologies developed in the FGCS project are useful not only to
knowledge information processing systems which were the primary objectives of
the project, but also to wider areas of applications. This project aims at
making use of an advantage of KL1 in flexibly describing concurrent processing
and further extending it. Researchers who engaged in the design and
implementation of KL1 and KLIC joined the project together with others who
recognized the possibilities of the technology.
FGCS technologies have a little different basis from conventional computer
technologies which are apt to be looked on as something peculiar and
inconsistent with conventional computer technologies. In proposing this
project, we had difficulties in persuading people that FGCS technologies can
be a key in constructing truly general-purpose large scale computation
technology.
We, who had taken part in the FGCS project, felt very happy that this proposal
was accepted, as the FGCS technologies were recognized to have possibilities
of having high values in liaison with large-scale numerical processing
technologies. We expect that, as parallel processing and knowledge
information processing technologies get higher importance in future computer
technologies, tangible and intangible bequests of the FGCS will continue to
grow up in synergy with technologies of other areas of computer science.
3) Prof. Ryuzo Hasegawa Group Project (Kyushu University)
- Theme: Advanced Argumentation System Based on Theorem Proving
Technologies
- Research Members: Ryuzo Hasegawa (Kyushu Univ.),
Katsumi Nitta (Tokyo Institute of Technology),
Hiroshi Fujita (Kyushu Univ.), Katsumi Inoue (Kobe Univ.)
Yasuyuki Shirai (Mitsubishi Research Institute)
Reiner Haehnle (Karlsruhe Univ.),
Hiroshi Komatsu (Law Offices of Homma & Komatsu)
- October, 1997 - January, 1999 (tentative)
- Outline:
The new HELIC-II developed at ICOT has been highly rated by many researchers
involved in the argumentation system for legal reasoning. However the project
has also revealed hurdles which have to be dealt with in order to make
practical use of the system. The issue concerns such aspects as theoretical
bases for reasoning, the relation to law theory, the extension of inference
models, user interfaces and so on.
MGTP, also developed at ICOT, has successfully achieved a high level of
performance, including parallel execution which is not only for solving
mathematical problems, but also for some constraint satisfaction problems.
However one of the great achievements of the system is its success in solving
open problems in the field of finite algebra. However, insufficiency in the
area of actual applications has lead the MGTP to be appreciated only by the
people involved in theorem proving research.
This proposal is aiming at achieving the following two items:
- Combining the two technologies and the know-how which was acquired through
the development process to improve its capability in real situations.
- Developing a legal knowledge-base and some new techniques, such as a rule
preference module, heuristic search module, search control module, and
constraint solving module, which are needed to formalize practical legal
reasoning.
Theorem proving technology provides the argumentation system with not only the
theoretical bases for advanced functions used in legal reasoning, but also
makes large scale argumentation possible, thanks to its great efficiency.
We are very glad to know that the technology of the argumentation system for
legal reasoning and theorem proving, the theme of the long committed project
at ICOT, have been publicly recognized as a promising technological basis.
However in order to develop those technologies to a level where they can bear
realistic usage and be promoted; the combining of element technologies in the
development process and showing the public its ability in real situations is
absolutely vital.
We would like to augment the appeal of the usefulness of ICOT products through
this project.
The proposal emphasizes the following points:
- The technology of argumentation systems for legal reasoning and theorem
proving developed at ICOT, is undoubtedly at the top level in the world and
has also been rewarded with a high reputation.
- Argumentation systems are equipped with a multi-agent negotiation function,
thus it can be applied to broad areas in industries.
- Theorem proving systems are a useful problem solving tool for not just
solving mathematical problems.
- The project should not only deal with theoretical toy problems, but should
also have a good vision of what would be necessary for the technology to be
applicable in the real world.
The members of this project have been maintaining a close cooperative
relationship which was established through working together at ICOT. We are
hoping to make this project fruitful with excellent teamwork.
(Edited by Shunichi Uchida)
Message From the Editorial Desk
We hope you enjoyed AITEC NEWS No. 14.
Did you find our report of ICOT OBs' research projects for government
sponsored programs useful? We hope so.
We would like to keep reporting a variety of activities, research, and
achievements relating to AITEC on AITEC NEWS. If you have any news and
information you would like to introduce in this publication, please send it
to us: aitec-news@icot.or.jp
All the AITEC staff wish you a Merry Christmas and a very Happy New Year!
**********************************************************************
* *
* A I T E C N E W S Issue #14 *
* AITEC NEWS Editorial Team: *
* Makiko Sato, Chie Takahashi, Akira Aiba *
* Kazumi Kasai, Kouichi Takeda, Yoshiharu Torii *
* Hiroshi Sato, Shunichi Uchida *
* AITEC NEWS English Version Team *
* Masayo Fukushima, Shunichi Uchida *
* Issued on: November 26, 1997(Japanese Version) *
* January 14, 1998(English Version) *
* By: Research Institute for Advanced Information *
* Technology (AITEC), a subcenter of *
* Japan Information Processing Development *
* Center (JIPDEC) *
* 2-3-3, Minato-ku, Shiba, Tokyo 105, Japan *
* Tel: +81-3-3456-3191 Fax: +81-3-3455-4877 *
* E-mail: aitec-news@icot.or.jp *
* http://www.icot.or.jp *
* *
**********************************************************************
www-admin@icot.or.jp
AITEC NEWS Home
AITEC Homepage