From: rick@cs.arizona.edu (Rick Schlichting)
Subject: INDUSTRY: Kahaner Report: Japan's ICOT (Fifth Generation Computing)
Date: 9 Jun 92 03:08:09 GMT
Message-ID: <17617@optima.cs.arizona.edu>


Crossposted from comp.research.japan:


  [Dr. David Kahaner is a numerical analyst on sabbatical to the 
   Office of Naval Research-Asia (ONR Asia) in Tokyo from NIST.  The 
   following is the professional opinion of David Kahaner and in no 
   way has the blessing of the US Government or any agency of it.  All 
   information is dated and of limited life time.  This disclaimer should 
   be noted on ANY attribution.]

  [Copies of previous reports written by Kahaner can be obtained using
   anonymous FTP from host cs.arizona.edu, directory japan/kahaner.reports.]

To: Distribution
From: 
 David K. Kahaner
 Office of Naval Research Asia
 (From outside US):  23-17, 7-chome, Roppongi, Minato-ku, Tokyo 106 Japan
 (From within  US):  Unit 45002, APO AP 96337-0007
  Tel: +81 3 3401-8924, Fax: +81 3 3403-9670
  Email: kahaner@cs.titech.ac.jp
Re: ICOT (Fifth Generation Computing) 1992 (Final conference)
4 June 1992
This file is named "icot.692"

ABSTRACT. A brief description is given of the Fifth Generation Computer
System, 1992 (final) conference, and an evaluation.

I have written several reports on the Fifth Generation Computer Systems
project (FGCS), and its associated research institute, Institute for New
Generation Computer Technology (ICOT). See for example, [icot.kl1, 28
Feb 1992].

During the week 1-5 June the 1992 FGCS international conference is being
held, with almost 2,000 people in attendance. I am hoping to receive
detailed reports from several attendees and will distribute these when
they arrive, but in the meantime I will summarize some of the important
developments very briefly.

The FGCS'92 conference is one of the best planned and executed
international meetings that I have ever attended. The materials,
lectures, and demonstrations have been organized and implemented with
great care and preparation. The conference materials include a two
volume proceedings, complete copies of all the keynote and invited
speeches, a detailed (37 page) outline of the history of the project, a
complete and detailed explanation of more than twenty demonstrations,
and a list and explanation of all the ICOT software. The printed
material, over 2,000 pages, is in English.

Within this report it is not possible to summarize fully the FGCS
project. K. Fuchi, who runs the research center in Tokyo (ICOT) gave an
excellent keynote speech. I was sufficiently impressed to produce a copy
for inclusion with the report, below. It is both readable and
interesting, and I urge you to read through it.

I discussed some of my thoughts about the project with other visiting
Westerners. The comments below are a quick summary of the general
feelings expressed to me, as well as my own.

When FGCS was established in the early 1980s, its hopes were to use two
new ideas together. First, the concept of logic programming, best known
to the computing world through the languages of Prolog and Lisp, to
solve very difficult computing problems of non-numerical type. Second,
parallel processing to provide the computational power to tackle the
huge computational needs that logic programming would require. During
the ten year period of the project, the Japanese government spent more
than 50 billion Yen (over $300 million US). ICOT has had almost 200
Japanese researchers who rotate back to their home institution every 3-4
years.  At the moment there are about 100 researchers at ICOT, almost
all under 35 years old. About 75 non Japanese researchers from 12
countries have participated in FGCS, and seven have worked at ICOT for
one year or more.

At the beginning of the project, Japan's status in terms of computer
science was very low, and in the West, project descriptions were greeted
with a combination of derision (problems are too difficult for us, no
less the Japanese, etc), and panic (if the Japanese solve these problems
before we do, they will control the world's technology for decades,
etc). No matter what the results, Western scientists would be able to
say "we told you so". Towards the latter part of the project, the
general view in the West has been that its goals have not been met. I
have consistently stated in these reports, and will do so again, that my
own opinion is that the impossibly high expectations claimed for FGCS
were Western, and not the same as those claimed by the Japanese. The
project needs to be judged against its own claims, rather than ours.
Fuchi, below, makes a similar comment. 

While FGCS was not another "sputnik", it had many significant
accomplishments. The world has changed in ten years. The key role played
by logic programming is seen by many to be reduced. Nevertheless, logic
programming systems developed at ICOT are probably the best in the
world. A variety of parallel computers have been built to test ideas,
and some of these experimental machines are as interesting as parallel
computers anywhere. The latest appears on the verge of achieving its
goals of 10^8 logical inferences per second. Whether that speed can be
attained or sustained in a real application remains to be seen. Large
numbers of young Japanese have been trained in the ideas of symbolic
computation, software and parallel computing, and there is probably as
much or more expertise within Japanese companies on these topics as
within any Western counterparts. FGCS provided both money and and focus
to successfully lubricate the start up of a knowledge processing
industry in Japan. The basic research from Japan in theorem proving and
related areas is comparable to that in the West, a situation very
different from ten years ago. Everyone I spoke to agreed on the major
role FGCS played in the infrastructure of Japanese computer science.
Also most Western attendees were more impressed than they expected to
be with the results that they heard about.

On the negative side are a lack of real applications. ICOT has
developed, and the demonstrations showed, a significant number of
interesting small (prototype) applications.  These are running on ICOT's
parallel hardware, and show good speedups, approaching linear in a few
cases.  Perhaps ICOT should have focused on a few big applications to
drive the project; these always help channel and define the problems
that really need to be solved, rather than those that the developers are
interested in solving. Perhaps ICOT hoped that industry would jump in.
In fact they did not. And given the size of the FGCS project and the
number of computer makers that participated by building hardware,
sending their researchers, etc., the direct impact of FGCS on Japanese
industry has been low. We repeatedly asked responsible industrial
officials about the impact, and consistently got a polite "not much".
It may be too early. Fuchi, below, states that we will have to wait
another five years to see realistic applications.  Also, ICOT developed
hardware and software is unused and unusable outside of Japan, and not
as much inside Japan as the ICOT people would like. A unique language
and computer operating system have been developed. Presently these need
to be run on special hardware, although ICOT is studying the possibility
of moving the software to more standard Unix workstations.

To get more users, and also to encourage more international cooperation,
MITI has announced that all the ICOT developed software will be
available free of charge in source form, without any restrictions as to
use, modification, copying, expanding, etc. Of course there are no
warranties associated with such software. This amounts to over 70 large
programs, and includes the parallel operating software, parallel logic
programming language, and all the software demonstrated at the
conference.

Actually, this concept has been discussed for about a year, but MITI has
now made it official. There is still a problem of on what machines to
run the programs, but it is an intelligent step by MITI nevertheless;
without it the intellectual products of FGCS would likely remain unused
outside of Japan. This new policy only apples to the software developed
for this project. But statements by MITI about "promote the advancement
of the technologies of knowledge information processing and parallel
processing" suggest that it will be applied to the Real World Computing
project (RWC), MITI's new ten year project.

Concerning the demonstrations, I should mention that these were
exceptionally well done. Within ten booths every major ICOT software
product was shown, with mini-lectures, hands on, video tapes, etc., with
a good combination of overview and detail. 
These included the following.
	Parallel inference systems
	Diagnostic and control expert system based on a plant model
	Experimental adaptive model-based diagnostic system
	Case-based circuit design support system
	Experimental parallel hierarchical recursive layout system
	Parallel cell placement experimental system
	High level synthesis system
	Cooperative logic design expert system
	Parallel LSI router
	Parallel logic simulator
	Protein sequence analysis program
	Model generation theorem prover
	Parallel database management system
	Knowledge representation language
	Parallel legal reasoning system
	Experimental motif extraction system
	Concurrent program development system
	Parallel constraint logic programming system
	Experimental system for argument text generation
	A parallel cooperative natural language processing system
	Experimental discourse structure analyzer


As part of the program materials for this conference, we were given a
questionnaire. Results of the questionnaire will influence the future of
the project. Some downsizing will occur, although staff will be retained
to deal with the issue of distributing software. Fuchi mentioned that
the topic of parallel information processing is still considered an
important topic and might be the subject of another national project.


                         Launching the New Era

                             Kazuhiro Fuchi
                        Director, Research Center
         Institute for New Generation Computer Technology (ICOT)
             4-28, Mita 1-chome, Minato-ku, Tokyo 108, Japan
(Presented as a keynote speech at the FGCS 1992 Conference, June 1992, Tokyo)

[In brackets below are remarks made by Fuchi during the presentation of
his talk. Whenever names are given of ICOT scientists, these are usually
accessible via electronic mail by addressing in the form 
`last-name@icot.or.jp', such as `kurozumi@icot.or.jp'.]

Thank you for coming to FGCS'92. As you know, we have been conducting a 
ten-year research project on fifth generation computer systems. Today is 
the tenth anniversary of this project and marks the founding day of our 
research center.  

The first objective of this international conference is to show what we 
have accomplished in our research during these ten years.  

A variety of innovative studies, in addition to our own, have aimed at 
future generations of computers and addressed the future of information-
processing technologies. These projects are currently in progress in many 
parts of the world.  Another objective of this project is to offer an 
opportunity for researchers involved in FGCS-related advanced research to 
present the results of their work and to exchange ideas.  

I constantly use the phrase "parallel inference" as the keywords to 
simply and precisely describe the technological goal of this project.  
Our concept holds that parallel inference technology will provide the 
core for those new technologies in the future - technologies that will be 
able to go beyond the framework of conventional computer technologies.  

During these ten years I have tried to explain this idea whenever I have 
had the chance. One obvious reason why I have repeated the same thing so 
many times is that I want to make its importance known to the public.  
But I have another less obvious reason.  

When this project started, an exaggerated image of the project was 
engendered. This image seems to persist even now.  For example, some 
people said that we were trying in this project to solve in a mere ten 
years some of the most difficult problems in the field of artificial 
intelligence (AI), or to create a machine-translation system equipped 
with the same capabilities as humans.  

In those days, we had to face criticism based upon the false impression 
that it was a reckless project trying to tackle impossible goals. Now we 
see criticism from inside and outside the country claiming that the 
project has failed because it has been unable to realize those grand 
goals.  
 
The reason why such an image was born appears to have something to do 
with FGCS'81 - a conference we held one year before the project began.  
At that conference we discussed many different dreams and concepts. The 
substance of those discussions was reported as sensational news all over 
the world.  Any vision with such ambitious goals, however, can never be 
materialized as a real project in its original form. If a project is 
started in accordance with the original vision, it cannot be managed and 
operated within the framework of an effective research scheme.  
Therefore, our plans had become much more modest by the time the project 
was launched.  

For example, the development of an application system, such as a machine-
translation system, was removed from the list of goals. It is impossible 
to complete a highly intelligent system in ten years. It requires a 
preliminary development stage to enhance basic studies and to improve 
computer technology itself. We decided that we should focus our efforts 
upon this preliminary stage. There was, also, another reason. At that 
time in Japan, some companies had already begun independently to develop 
realistic, low-level machine-translation systems, and developmental 
competition was sure to ensue. A large part of our pattern-recognition 
research was also eliminated, because a national project called "Pattern 
Information Processing" had already been conducted by the Ministry of 
International Trade and Industry for ten years. We also found that the 
stage of the research did not match our own.  
 
Does all this mean that FGCS'81 was deceptive? I do not think so. First, 
in those days, a pessimistic outlook predominated concerning the future 
developments of technological researches. For example, there was a 
general trend that research on artificial intelligence would be of no 
practical use. In that sort of situation there was considerable value in 
maintaining a positive attitude toward the future of technological 
researches whether this meant ten years or fifty. I believe that this was 
the very reason why we received remarkable reactions, both positive and 
negative, from the public.  

Second, the concept of parallel inference was presented in a clear-cut 
form at FGCS'81. Let me show you a diagram (Figure 1). This diagram was 
the one I used for my speech at FGCS'81. The draft for this was completed 
in 1980, but I had come up with this idea four years earlier.  After 
discussing the concept with my fellow researchers for four years, I 
finally completed this diagram.  

Here, you can clearly see our concept that our goal should be a parallel 
inference machine.  [The figure Fuchi mentions clearly shows that the
program goal is to build a parallel inference machine.]

I would appreciate it if you would compare this diagram with our planning 
for the final stage of this project, which Deputy Director Kurozumi will 
show you later. I would like you to compare the original structure 
conceived 12 years ago and the present results of the project so that you 
can appreciate what has been accomplished and criticize what is lacking 
or what was immature in the original idea.  

Some people tend to make more of the conclusions drawn by a committee 
than the concepts and beliefs of an individual. It may sound a little bit 
beside point, but I have heard that there is a proverb in the West that 
goes, "The horse designed by a committee will turn out to be a camel." 
 
The preparatory committee for this project had a series of enthusiastic 
discussions for three years before the project's launching. I thought 
that they were doing an exceptional job as a committee. Although the 
committee's work was great, I must say that the plan has become a camel.  
It seems that their enthusiasm has created some extra humps.  

This is not the first time I have expressed my opinion about this issue.  
I have, at least in Japanese, been saying this in public for the the past 
ten years. I think that I could have been discharged at anytime my 
opinions were inappropriate.  

As the person in charge of this project, I have pushed my opinion in line 
with the parallel inference policy based upon my own beliefs.  Although, 
I have been criticized as still being too ambitious, I have always been 
prepared to take responsibility for that.  

Since it is a national project, it goes without saying that it should not 
be controlled by one person. I have had many discussions with variety of 
people for more than ten years.  Fortunately, the belief has not remained 
just one person's belief but has become a common belief shared by many 
researchers and research leaders involved in the project.  

Assuming that this project has proved to be successful, as I believe it 
has, this fact is he biggest reason for its success.  For a research 
project to be successful it needs a favorable work environment that meets 
many conditions [including financial]. But what is most important is
that he research group involved has a common belief and a will to reach
its goals.

So much for introductory remarks. I wish to outline in terms of parallel 
inference the results of our work conducted over these ten years. I 
believe that the remarkable feature of this project is that, by focusing 
upon one language, it has succeeded in experimenting with the development 
of hardware and software on a large scale, based upon that language.  

>From the beginning, we envisaged that we would take logic programming and 
give it a role as a link that connects the highly parallel machine 
architecture and the problems to be solved concerning applications and 
software. Our mission was to find a programming language for parallel 
inference.  

A research group led by Deputy Director Furakawa was responsible for this 
work. As result of their efforts, Ueda came up with language model, GHC, 
[Guarded Horn Clause] at the beginning of he intermediate stage of the
project. The precursors to this were such as Parlog and Concurrent
Prolog. He enhanced and simplified hem to make this model. Based upon
GHC, Chikayama designed a programming language called KL1.
 
This KL1, a language derived from the logic programming language concept, 
provided a basis for the later half of our project.  Thus, all of our 
research plans in the final stage were integrated under a single 
language, KL1.  

For example, we developed a hardware system, the Multi-PSI, at the end
of the intermediate stage. We demonstrated it at FGCS'88. After the
conference we made copies and since then have used them as the
infrastructure for software research.

In the final stage, we made a few PIM prototypes, a "parallel" inference
machine that has been one of our final research goals on the hardware
side. These prototypes are being demonstrated at this conference.

Each prototype has a different structure including an internal network
architecture, and each machine itself is a subject of research. Looking
from the outside, however, all of them are KLl machines.

Division Chief Uchida and Laboratory Chief Taki will show you details on 
this later. What I want to emphasize here is that all of these machines, 
including their internal chip designs even, are designed with the 
assumption that KL1, a language equivalent to other very high level 
languages, is a machine language.  

On the software side, our research topics were also integrated under the 
KLl language. All the application software and basic software, including 
operating systems, were to be written in KL1.  

We demonstrated an operating system called PIMOS at FGCS'88, which was 
the first operating system software written in KL1. At that point, it was 
still immature. It has been improved since then, and the fully matured 
version of PIMOS securely backs the demonstrations being shown at this 
conference.  

Details will later be given by Laboratory Chief Chikayama, but I wish to 
emphasize that not only have we succeeded in writing software as 
complicated and huge as an operating system entirely in KL1, but also 
that we have proved through our own experience that KLl is much more 
appropriate than conventional languages for creating the voluminous 
software required for an operating system.  

One of the major challenges in the final stage was to demonstrate that 
KLl is effective not only for basic software, such as operating and 
language-processing systems, but also for a variety of applications. As 
Laboratory Chief Nitta will report later, we have been able to 
demonstrate KLl for various software issues including LSI-CAD, genetic 
analysis, and legal inference. These are closely related to issues in the 
real world and have a scale nearly sufficient for practical use.  But, 
again, what I wish to emphasize is that the objective of those 
developments has been to demonstrate the effectiveness of parallel 
inference.  
 
In fact, in the first stage we tried this type of approach to develop a 
project focusing upon one particular language, although it was a stage in 
which research was conducted using sequential techniques. We used ESP, an 
expanded version of Prolog, as a basis.  

Assuring that ESP could play a role as KLO, a sequential personal 
inference machine called PSI was designed as hardware.  We decided to use 
the PSI machine as a workstation for our research. Some 500 PSIs, 
including modified versions, have so far been produced and used in the 
project.  

SIMPOS, the operating system designed for PSI,- is written solely in ESP. 
In those days, this was one of the largest software volumes written in a 
logic programming language. Up to the intermediate stage of the project, 
we used PSI and SIMPOS as the infrastructure to conduct research on 
expert systems and natural language processing.  [Fuchi mentioned that
there have been criticisms of this approach.]

Our project, though it is conducted on a large scale, is still
considered basic research.  Accordingly, it is supposed to be conducted
in a free, unrestrained atmosphere so as to bring about innovative
results. Some of you may be wondering whether above policy restrains the
freedom and diversity of research? But this is also based upon my, or
our, philosophy.  I believe that research is a process of "assuming and
verifying an hypothesis." If this is true, the hypothesis must be as
pure and clear as possible. If not, you cannot be sure what you are
trying to verify.

Another thing is that we had a strong belief that our hypothesis had 
sufficient scope for a world of rich and free research. Even if the 
hypothesis were restrained, we believed that it was a creative 
constraint.  

I would be a liar if I was to say that there was no resistance among our 
researchers when we decided upon the policy. KLl and parallel processing 
were a completely new world to everyone. It required a lot of courage to 
plunge headlong into this new world.  But once everyone overcame the 
psychological barrier, the researchers set out to create new parallel 
programming techniques one after another.  

Among the many other results we obtained in the final stage was the 
development of a fast theorem-proving system, or prover.  Details will be 
given in Laboratory Chief Hasegawa's report. I think that this research 
will lead to the resurrection of theorem-proving researches.  This prover 
is also used in the inference engine of our legal inference system.  
 
The research on programming languages has not ended with KL1. For 
example, a constraint logic programming language called GDCC has been 
developed as a language higher than KL1. We also have a language called 
Quixote.  

>From the beginning of this project, I have advocated the idea of 
integrating three types of languages-logic, functional, and object-
oriented languages-and two worlds: the programming world and the database 
world. In the Quixote language, this idea has been materialized. It is 
something we call a deductive object-oriented database language. [My
discussions with researchers at the conference suggested that Fuchi's
view of object oriented languages is relatively recent.]

Another language, CIL, was developed by Mukai in the process of studying 
natural-language processing. CIL is a semantic-representation language 
that is also designed to deal with situation logic. CIL is incorporated 
into Quixote in a natural form. Quixote therefore has the characteristics 
of a semantic-representation language. As a whole, it shows one possible 
future form of knowledge-representation languages.  

More details, along with the development of a distributed parallel 
database-management system, Kappa-P, will be given by Laboratory Chief 
Yokota.  

Thus far I have outlined, albeit briefly, the final results of our ten-
year project. Recalling what I envisaged ten years ago and what I have 
dreamed and hoped would materialize for 15 years, I believe that we have 
achieved as much or more than what I expected, and I am quite satisfied.  

Naturally, a national project is not performed for mere
self-satisfaction. The original goal of this project was to create the
core of next-generation computer technologies. Many different elemental
technologies are required to materialize future computers or tomorrow's
information processing. Although it is impossible for this project alone
to provide all of the elements, we are proud to be able to say that we
have created the core, or at least provided an example.

The results of this project, however, cannot be commercialized as soon
as the project is finished, which makes the fact that it is a national
project more meaningful. I estimate that we need another five years,
which could be called "a maturing period for the technology", for the
results to actually take root in society.  I had this prospect in mind
at the beginning of this project ten years ago, and have kept relating
it in public right up until today, but I still have the same idea now
that our project is nearing its end.

It often happens that there is a gap of ten or twenty years between the
basic research stage of a technology and the day it appears in the
business world. UNIX, C, and RISC, the technologies involved in the
popular trend toward downsizing, are good examples of this.

>From the beginning of our project, we tried to think of technologies in 
the distant future that would go beyond these.  

[Fuchi commented that many researchers are of the opinion that Japanese
industry has already caught up with ICOT research.]

A movement toward parallel computers has been gaining momentum as a
technology leadinto the future. But skeptical opinions were dominant ten
years ago. The situation was not very much different five years ago, and
skepticism still remains.  But the trend seems to be rapidly changing.
In the background, there exists the fact that it is becoming easier to
make a parallel machine by placing many chips together, because
semiconductors technology constantly makes progress.

In most cases, people still focus upon supercomputers for scientific and 
engineering  use, and their ideas tend to be [naive and] vague regarding 
the software side [sic!]. Despite this a new age is dawning.  

The software issue might not be a big problem as long as scientific and
engineering computations remain merely scaled-up matrix calculations
[and maybe Fortran would suffice], but software will certainly become a
great problem in the future. If this problem could be solved, would it
allow us to deal with massive problems that have complicated overall
structures? If this is realized, we would have something like a
general-purpose capability that is not limited to the range of
scientific computations. We might then be able to replace the mainframe
computers we are using now.

The scenario mentioned above is one possibility leading to new, future 
computers that are equivalent to current mainframes. If you think about 
what type of technology will then be required, the answer should be the 
parallel inference technology which we have been pursuing.  

I will not press this idea upon you. But I anticipate that if anybody 
starts research without knowing our ideas, or if that research is based 
upon a philosophy that is quite different from ours, after many twists 
and turns that person will reach more or less the same concept as us with 
only small differences such as different names.  

It may be valuable for researchers to be independent from what has 
already been done, for them to struggle through a process of research by 
themselves, and to find that they have followed the same course as 
somebody else. But it is more efficient to build upon what has been done 
in this FGCS project and divert energy to moving forward from that point.  

This project will be finished at the end of this year. As for "maturation 
of the parallel inference technology", I think we will need a new pattern 
for our research activities. There is a concept called "distributed 
cooperation" in computation models. I expect that, like in that kind of 
pattern, the seeds generated in this project will spread both inside and 
outside the country and sprout in many different parts of the world.  

For this to be realized, the results of this project must be freely 
accessible and available worldwide. This means that it is essential to 
disclose our accomplishments, for example in the software area including 
the source codes, and to make them "international common public assets." 

We have tried for ten years to encourage international exchange in this 
project. As a result we have enjoyed opportunities to meet with many 
researchers involved in advanced studies in various parts of the world.  
They have given us much support and cooperation. This project could not 
have been completed without this generous support and cooperation.  

In that regard, and also considering that this is a Japanese national 
project that aims at making a contribution, though a small one, toward 
the future of mankind, we believe that we are responsible for leaving our 
research accomplishments as a legacy to future generations and to the 
international community in the most suitable form.  

Although this project is about to end, the end is just another starting 
point. The advancement of computers and information-processing 
technologies is closely related to the future of human society. For the 
purpose of launching a new age, I fervently hope that the circle of those 
who share our passion for a bright future will continue to expand.  Thank 
you.  
 
--------------------------------END OF REPORT---------------------------