Lapas attēli
PDF
ePub

University Hospital. (We were told that Western research in this area was not sufficient for Asian people, who require greater flexibility for squatting. Interestingly, the joint model we saw has been developed in stainless steel and is very heavy. I would expect that newmaterials research could provide alternative solutions.) Two new industrial services have been initiated--printed circuit board (PCB) design assembly and testing and rapid prototyping using stereolithography. The latter especially was a big hit when demonstrated at the ICCIM'91 vendor exhibit and later when it was shown to us at GINTIC. Stereolithography (SLA) was developed by the U.S. firm 3-D Systems and uses a laser on a liquid photocurable polymer to form models directly and very quickly from CAD/CAM data, so called direct three-dimensional (3-D) manufacturing. It is said to be a breakthrough because of the speeds with which prototype parts can be produced. SLA models are also heavily used for mold making, but most of the parts built this way are still only good for rapid prototyping, not real production (due to insufficient strength). GINTIC has 3-D Systems' 10-inch unit, allowing direct forming of parts up to 10 inches on a side (the company's largest unit allows 25-inch parts). Larger parts can be built in sections. (Several other U.S. and Japanese groups are also very active in this direct 3-D manufacturing technology.)

Education and training are accomplished by an emerging M.S. program in CIM just begun July 1990. Presently the program has about 40 students. There are also 2 Ph.D. students, 20 fulltime (M.S.) graduate students, and 68 part-time graduate students. It should be mentioned that this program was heavily modeled after ones at RPI and Fraunhofer and that Prof. Lester Gerhardt (RPI), has been active in collaborating with GINTIC staff in forming the new program. In fact,

A number of the research projects were presented at ICCIM'91 and these

Gerhardt is also a lecturer in their M.S. program. The list below suggests the range of papers are included in the Proceedprojects that are being studied.

At the CAD and Knowledge-Based Center

• Parametric feature-based intelligent component designer and modular feature design expert system.

• Intelligent knowledge-based OO process planning system for the manufacture of progressive dies.

[ocr errors]

Daily production scheduling system for the manufacture of PCBs.

ings. A good one (Lim, GINTIC) describes the control and communication system designed and implemented at GINTIC beginning in 1985. There is also a discussion of a C++ graphical toolkit (C. Wah et al.) developed using object-oriented techniques. At least one paper generated some healthy criticism. L. Tshon described a hierarchical approach to mechanical design based upon a tree of tasks, group leaders, and specialists. His approach allowed communication between specialists within a group but only between leaders of different groups, and even then only

Multilevel neural network job shop those at the same level. Along with scheduling system.

several other members of the audience, I doubted that this very rigid approach

At the PCB Flexible Manufacturing System would be successful at modeling real(FMS) and SLA System world systems.

[blocks in formation]

well as the Singaporean Minister for Finance presented speeches at the opening. A short history of the plans and construction of the facility was given via a slide show, and while lighthearted the message was clear; pictures of the new center were accompanied by the Hallelujah Chorus from Handel's Messiah and the theme music from "2001"!

Singapore is staking a significant fraction of its future on automation; the theme is that by the year 2000 Singapore will be an "intelligent island." To this end the JSAI Center is seen as a way of increasing awareness about AI in Singapore, learning needed skills, developing a friendly base between the two countries in a important technological field, learning about Japanese management practices, as well as building relationships and contacts. My earlier report detailed a number of AIlike projects that have already been implemented. In total about 50-60 projects have been developed; one-third are already in use, and the remainder are in prototype form. There are about 200 AI professionals now working in the country, while about three times that number are projected to be needed. The AI Center will be an important vehicle for accomplishing this goal. Japan has given the center about ¥380M (about $2.5M), mostly in the form of equipment (32 workstations, 25 386-PCs, 17 Macintoshes, etc.) and software. In addition to equipment, Japan provides software and technical expertise, while Singapore provides facilities, local personnel, and operating expenses. Six Japanese AI experts will work at the center and five or six Singaporeans will go to Japan for training with various companies (four are already there). It is also planned that about 500 training slots will be utilized within 4 years. In fact, training is a key element in the center. Three courses are planned: a 3-day course, "Intelligent Systems for

Managers"; a 2-week course, "Intelligent Systems for Business Professionals"; and a 14-day course emphasizing more detail, "Expert Systems/KnowledgeBased Systems for IT Professionals." In addition, the heart of the program is a 6-month "Prototyping Expert Systems/ Knowledge-Based Systems" course, enabling trainees the opportunity to build prototypes of "real" systems under expert guidance. The Singaporean Government is encouraging industry participation in a very direct way, by providing a 70% absentee payroll subsidy for company-sponsored employees who are either Singaporean or permanent residents.

This is a "win-win" program as I see it for both Singapore and Japan. Others should view it this way, too, as long as they appreciate that the emphasis at the center is on real-world applications-this is definitely not going to be an academic research facility. Singapore gets specific assistance as outlined above. Japanese companies get to place their equipment, software, and personnel in Singapore and encourage use of their products during training, plus perhaps get assistance in developing marketable software.

After the formal opening, there was an opportunity to browse around the center and discuss various demonstration projects. These included a Tokyo subway advisor, production line diagnosis system, myopathy consultation system, hospital operation theater prioritization system, traffic data management system, production engineering expert system, pattern optimization system, etc. Many of them were running on NEC EWS4800 systems. A few are potentially useful, but others were obviously for teaching. In the latter category was the Tokyo subway advisor. A map of the (admittedly complicated) Tokyo system is displayed and then users select an origin and destination station (in Japanese). The

system grinds (for quite a long time) while various rules are checked for the best route based on shortest time and/ or cheapest fare. I felt that most people could make a similar mental calculation in less time. Some constraints are allowed, however, and these could make such a system really useful, say at an airport. Unfortunately, the question asked by all Tokyo tourists--how do I get from here to there without going up stairs--was not available. Maybe later. For information contact

Director

Japan-Singapore AI Center
CINTECH II

75 Science Park Drive #01-01/04 Singapore 0511

Tel: +65-779-3088
Fax: +65-779-6162

The Japan-Singapore AI Center is not the only software R&D activity supported by non-Singaporean companies as Table 2 shows. The list of applications should be thought of only as representative. In addition, IBM, DEC, Sun, and probably other vendors have some software development in Singapore. Also in Singapore there is a French-Singapore (F-S) Institute and a German-Singapore (G-S) Institute on adjoining sites. Although I did not visit either place, at least one paper on FMS at ICCIM'91 was authored by a scientist from the F-S Institute (Fertin). Prof. Peter Sackett, Cranfield Institute of Technology (E-mail: sackett@ cim.cranfield.ac.uk), explained to me that the G-S Institute is more closely aligned to manufacturing than the French, which concentrates on IT issues, although both illustrate the Singapore approach by providing training, 48-week year, 2,000-h/yr classes in 2- or 3-year programs leading to a diploma (not a degree). Some institutions recognize outstanding performance on these programs as equivalent to a good degree

and valid as entry to a masters program. Sackett emphasizes that the institutes are not research oriented, although they are very well equipped and staffed, and they do undertake ambitious project work. One example is the building, again by students, of a fully operational FMS. NATIONAL UNIVERSITY OF SINGAPORE (NUS)

While I had visited NUS earlier, I was anxious to visit again in order to

see

Prof. T.R. Nanda
Dept of Mathematics
National University of Singapore
10 Kent Ridge Crescent

Singapore 0511

Tel: +65-772-2940

E-mail: matnanda@nusvm.bitnet

Nanda's work on user interface for numerical software was called to my attention by Prof. B. Datta. Prof. Nanda

has his Ph.D. from New York University (NYU) (Courant) in functional analysis has been retraining himself in applied and computational areas. For several years he has been developing a system, running on a PC and written in APL, for plotting and manipulation of mathematical entities. His system has a number of special functions of mathematical physics built in, and he is working on other capabilities such as numerical integration, solution of differential equations, etc. Nanda claims that his targets are students, but in this arena there are already a number of good products, such as Matlab, MathCad, Mathematica, etc., and so he will have a tough time against those, especially as his system is not particularly extensible. Nevertheless, it is a good piece of work and he would like to hear from potential collaborators and users. At the moment he is developing this pretty much in a vacuum and definitely needs input and suggestions from outsiders.

During my visit Nanda showed me the computational facilities of his department, which are extensive, including numerous workstations, a very excellent computerized library card catalog, and a campus area network. But he was only mildly upbeat about the research activities. He felt that NUS students work diligently, but there are very few graduate students and that it will take time for the university to develop into a true research institution. In short, whatever money can buy has mostly been purchased; the rest is up to the faculty. NUS is encouraging visiting scholars and this will definitely aid the environment. I also had an opportunity to meet one U.S. visitor spending his sabbatical at NUS, Prof. Jerome Klotz, from the University of Wisconsin's Statistics Department [E-mail: klotz@stat.wisc.edu]. Prof. Klotz echoed Nanda's assessment that NUS's students are hard working, but mostly not used to a Western style of student-professor interaction.

[blocks in formation]

CONCLUDING REMARKS

Every visitor to Singapore is impressed with the physical facilities and equipment. I mentioned this above in the context of GINTIC and again about NUS. But people are worried about this hardware-heavy/but softwareweak situation in their country. (Software here means particularly human resources.) Some Singaporeans feel that they are lacking a pool of experienced, well-trained people who can take full advantage of this wealth of hardware and hopefully lead their country to future growth. At the lower levels I have mentioned the emphasis on training already. At the higher levels this situation is claimed to be particularly bad at the universities, which are struggling to recruit good faculty members for their engineering programs. I was told that the Government has set up a special fund and laws to attract to Singapore

those Chinese students who are gradu-
ating from good U.S. or European
universities with doctoral degrees and
who do not wish to return to mainland
China. In addition to good financial
incentives, they and their family
members are granted immediate
permanent residence in Singapore.
Apparently some initial success of this
special recruiting program has already
been achieved.

Besides, the Western model is probably not appropriate in any country in Asia, and especially those with a Confucian streak that venerates age and family leadership. Singapore is free, but one doesn't step too far out of line here, anyway not if one expects to have a career and a future.

To repeat, coordination is a strong point in Singapore, strong leaders + good visions-->well planned actions->careful monitoring of action results + revision of plans and visions. Sounds reasonable, and certainly appears to work in this small country.

All scientific visitors to Singapore
must also come to grips with the reali-
ties of the political world here, which is
not as open as one might find in some
other countries. Singapore has been
run in a very authoritarian way by its ACKNOWLEDGMENT
benevolent Prime Minister Lee for many
years. The system is slowly changing,
and to be fair most Singaporeans and
many outsiders think that restrictions
are reasonable and have benefitted their
country. Nanda feels that it is still too
soon to expect a completely open soci-
ety, and perhaps this will never occur.

I wish to note my sincere thanks to Profs. Lu (University of Illinois), Gerhardt (Rensselaer), Sackett (Cranfield), and Ho (GINTIC) for reading this report and for their many helpful comments and suggestions.

FIRST KOREA-JAPAN CONFERENCE

ON COMPUTER VISION

The First Korea-Japan Conference on Computer Vision, held 10-
11 October 1991 in Seoul, Korea, is summarized and assessed.

by David K. Kahaner

INTRODUCTION AND SUMMARY

Conference chairmanship was shared describing Japanese and Korean between

The First Korea-Japan Conference on Computer Vision was held on 10-11 October 1991 in Seoul's KOEX Conference Center. Approximately 80 papers were presented (31 from Korea, 47 from Japan) in three parallel sessions during the 2 days. (KOEX, a massive convention facility, could easily have swallowed 50 such conferences.) Two invited papers summarized computer vision research in Korea and Japan, and respectively, and a panel discussed "Application of Computer Vision for Automation." There were about 200 attendees. As few Japanese or Koreans know each other's language, the conference language was English. Korean scientists are very far behind in development and industrial applications of computer vision. In basic research they are lagging a bit less because many Korean scientists have done research work abroad in the United States and Japan and are attempting to continue that in Korea. Both countries see analysis of moving images as the most important new research area, although many "standard" topics are far from solved. (The recent interest in movement stems from increasing processing speed. Without fast computers and fast data movement, researchers had to sample images every second or two, leading to very complicated relations between successive frames.)

Prof. Chung-Nim Lee
Basic Science Research Center
Pohang Institute of Science and
Technology (POSTECH)
P.O. Box 125

Pohang, Kyung Buk 790-330, Korea
Tel: +82-562-79-2041

Fax: +82-562-79-2799
E-mail: cnlee@vision.postech.ac.kr

Prof. Masahiko Yachida
Faculty of Engineering Science
Dept of Information and Computer
Sciences

Osaka University
1-1 Machikaneyama-cho
Toyonaka-shi, Osaka 560, Japan
Tel: +81-6-844-1151 x4846
Fax: +81-6-845-1014
E-mail: yachida@ics.osaka-u.ac.jp

I had met Prof. Lee almost 20 years earlier when we were both in the Mathematics Department at Ann Arbor. He was trained as a pure mathematician but is now interested in applied problems, particularly those concerned with computer vision.

The fairest things that can be said about this conference are that (1) the Korean scientists were very brave to have organized it, especially to have scheduled more or less alternating talks

research; and (2) the Japanese scientists were very gracious to have participated so enthusiastically. A preview of what was going to happen was given at the opening lectures when research in the two countries was reviewed. A literature search on computer vision uncovered fewer than one-tenth as many Korean as Japanese research papers.

Although the paper balance at the conference was much more even, it was very clear that Korean research is at a much earlier stage and that applications of vision in industry are much more limited than in Japan. There were a (very) few exceptions. For example, Prof. T. Poston in POSTECH's Mathematics Department [E-mail: tim@vision.postech.ac.kr] gave an elegant discussion of the use of singularity theory for vision applications, but at the moment this work is very far from practical realization. Also, some Korean industry is using vision techniques and, in particular, Samsung has a visual soldering inspection system that has many similarities to systems developed by Toshiba and NTT.

Computer vision is usually thought of as beginning with a real world situation (scene) that is input through a camera and then digitized and ending with a description of the scene, e.g., knowledge (perhaps leading to action which changes the scene, etc.). The inbetween steps are often modularized. Typically there is a "camera model"

« iepriekšējāTurpināt »