Conference report and a view of cybernetics in Russia

Conference report and a view of cybernetics in Russia

Cybernetics ­ expectations and results, Moscow

A two-day public event, termed a"jubilee reading" was held in the Polytechnical Museum, Moscow, during 24-25 November 1998, to mark the 50th anniversary of the publication of Norbert Wiener's book Cybernetics, as well as the 40th of its appearance in Russian translation. The event was sponsored jointly by the Polytechnical Museum and the Russian Association for Artificial Intelligence (RAAI).

The main effective organiser was Prof. Vadim ("Dima") L. Stefanuk, of the Institute for Information Transmission Problems of the Russian Academy of Science, who is Vice-President of RAAI. The intention was that he would organise and chair the meeting jointly with Prof. Dmitri A. Pospelov, but the latter was unable to participate because of illness. With the approval of Dmitri Pospelov, the sessions were co-chaired by the present writer, though my knowledge of the Russian language is such that I depended on help from Dima Stefanuk to follow the gist of the presentations.

Nominally, the first day was also chaired by the Director of the Museum, Prof. G.G. Grigoryan, who gave a well-received opening address stressing the importance of new ideas and viewpoints, and then excused himself because of other duties in the Museum. Then I gave a short talk (translated sentence-by-sentence by Dima Stefanuk), of which the text is appended, referring to my own impressions of cybernetics in Russia.

The presentations on the first day were mainly in the nature of reminiscences and reviews of past work. Several speakers took up a point I had mentioned in my talk, namely that for a time the topic of cybernetics was unacceptable in the then USSR. Some of the early workers risked disfavour in following the lines they did.

Prof. Z.M. Benenson recalled wartime work on electronic predictors for anti-aircraft gun control, as carried out in Russia in parallel with efforts in the Lincoln Laboratory and elsewhere. The showing of slides of circuits using thermionic valves (vacuum tubes) was nostalgic. There was also pleasant nostalgia in the reminiscences of another speaker, Prof. A.V. Shileyko, recalling the excitement of developing an early digital computer, when no one on the project thought of tearing himself away and going home before 10 p.m.

Other speakers reviewed the contributions of particular workers including M.A. Gavrilov and his school, A.A. Lyapunov, A.I. Kitov, A.I. Berg and V.M. Glushkov. For Glushkov, author of the well-known text Principles of Cybernetics and former Director of the Institute of Cybernetics in Kiev, it was noted that the 75th anniversary of his birth had recently passed. The daughter of A.A. Lyapunov was listed as one of two speakers to give accounts of his work, but in the event she preferred to leave it to the other one. The daughter, N.A. Lyapunova, is a research worker in the Institute of Human Genetics and did not feel she could contribute usefully to the review of essentially mathematical topics.

Prof. V.L. Stefanuk reviewed the work of a further group of early workers in the area, including I.A. Poletaev,M.G. Haase-Rapoport, M.L. Tsetlin and M.M. Bongard. His talk was well illustrated by slides from original photographs, including some showing Bongard as a mountaineer, a hobby that led to his untimely death.

The presentation by Stefanuk was meant to be the first following the break for refreshment, but he gladly changed places with Prof. Yu. A. Mikheev who was due to speak later in the afternoon about the contributions of V.M. Glushkov. Prof. Mikheev's time was limited because was actually absenting himself from a meeting of the Duma, the Russian Parliament, in order to speak at this meeting. He acknowledged the relevance of cybernetics to government, and to Russia's present difficulties.

A particularly interesting presentation was by Prof. Yakov I. Fet, from the Siberian Division of the Russian Academy of Science, Novosibirsk. He is an editor, jointly with Dmitri Pospelov, of a valuable collection of essays on the history of computer science in Russia. The term is interpreted widely and the material includes a great deal that is relevant to cybernetics, including quite a lot about the visit of Norbert Wiener in 1960.

Prof. Fet later gave me a copy of the volume (Pospelov and Fet, 1998) that has recently been published. It is a substantial work of 662 pages. A second volume is in preparation. The main text is in Russian, but the Preface and table of contents are given in both Russian and English. Authors represented include Berg, Glushkov, Kantorovich,Kolmogorov, Lyapunov and many others. It is an extremely valuable work of reference. Prof. Fet, as well as Dima Stefanuk, was very interested to see the copy of the Norbert Wiener biography by Masani that I had taken with me.

The first day's session was in a medium-sized auditorium entered from the main exhibition area of the Museum, and participants were finally invited to move into a hall of the exhibition area for demonstrations and informal discussion. This part of the proceedings was especially delightful, including an excellent buffet donated by Siemens-Nixdorf. Dima Stefanuk demonstrated a prized possession ­ an example of the "music box" devised as an electronic composing device by Marvin Minsky in the very early days of Artificial Intelligence.

Relevant items of the museum's permanent collection were shown and demonstrated, including a player piano and other antique automatic musical instruments. The famous Russian "bionic hand"­ one of the first (if not the first) prosthetic device to be actuated by natural bioelectric potentials ­ was shown and demonstrated. We then watched two computer programs play draughts (checkers) against each other, with the pieces moved on a real draughts board by robotic manipulators. There was also the chance to look at other exhibits of this very comprehensive Museum, of which a number relating to space exploration were nearby. The Museum was founded in 1872.

On the second day the meeting was in a large auditorium, with a balcony, in another part of the Museum building. This was where Wiener spoke in 1960. The occasion was recalled by the man who interpreted for him, E.L. Napelbaum, now of the Institute for Systems Analysis. He remembered how the auditorium was jam-packed, with people sitting in the aisles and anywhere a niche could be found.

The auditorium, or at least the downstairs part of it, was fairly full for the recent meeting. Many of the participants were students from the Moscow State Engineering Physics Institute(Technical University). The head of its Cybernetics Department, Dr Sergey V. Sinitzyn, was one of the speakers. The focus, this time, was on relatively recent achievements and current work.

The proceedings began with a request for a few moments' silence in respect of a colleague who had recently died ­ G.S. Pospelov, of the same Institute as Dmitri Pospelov but no relation. Then Napelbaum spoke about the first IFAC Congress and the visit of Norbert Wiener in 1960, and confirmed my impressions of the marvellous spirit that prevailed.

Other speakers reviewed Russian contributions to Artificial Intelligence as well as to Robotics and legged locomotion, with cine film illustrating the latter topic. Dima Stefanuk spoke about the International AI Conference in Tbilisi in 1975, of which he was one of the instigators.

The current situation in computer chess was reviewed by M.V. Donskoy, who was responsible for an earlier prize-winning chess program. He regretted the fact that computer chess has now become a commercial matter, such that algorithms are not published and the programmers responsible for innovations are not accessible for questioning at meetings.

I was invited to say some words at the close of the meeting and I said I was glad to see that cybernetics was alive and well in Russia, and I conveyed the greetings of WOSC.

These notes only cover a small part of the content of the two days, but it is hoped they convey something of their flavour. There is a plan to set up a Website embodying the material of the meeting, probably with English-language versions also available.

Alex M. Andrew

Reference

Pospelov, D.A. and Fet,Ya. I. (Eds) (1998), Essays on the History of Computer Science in Russia,Scientific Publishing Center of the Siberian Division of the Russian Academy of Science, Novosibirsk, ISBN 5-7692-0101-0.

Appendix

Cybernetics in Russia ­ a personal view

First I want to say that I am honoured to be invited to speak at this event. This year sees the 50th anniversary of the publication of Wiener's book in English, and the 40th anniversary of the publication in Russian. Two years after the publication in Russian, in 1960, Norbert Wiener visited Moscow and spoke here on the topic"Science and society". His visit was at the time of the first Congress of IFAC­ the International Federation for Automatic Control ­ held in Moscow,and he attended some of its sessions.

The IFAC Congress was the occasion of my own first visit to the then USSR, and it was a great privilege and a wonderful experience to be present. Workers from east and west found they had much more in common than either side had previously imagined, and barriers of mistrust were broken down. Many enduring friendships and plans for collaboration were formed. The participants from the West were impressed by the warmth of their welcome in Russia, as well as by the enthusiasm with which the hosts had made the arrangements for the Congress, including a special issue of postage stamps and a magnificent reception in the Kremlin.

The publication of Wiener's book in translation, and the invitation to Norbert Wiener to visit Moscow, indicated the breaking down of another barrier of mistrust, since there had been an earlier period in which the topic of Cybernetics did not find favour in the Soviet Union. It is difficult to say why this was so, and a large amount of research was in progress in the USSR that was relevant to Cybernetics, though not presented under its heading. In particular, the famous Institute of Automation and Remote Control in Moscow was a centre for pioneering studies of self-optimising control systems, close to the idea of versatile and effective machine learning that was pursued by many workers in different countries. This area of study has a truly cybernetic inter-disciplinary character, since consideration of principles of machine learning naturally leads to comparison with the biological example.

The period following Wiener's visit ­ the 1960s and 1970s ­ was notable for fruitful international collaboration in the field of Artificial Intelligence. There were numerous academic exchanges between the Soviet Union and USA, as well as with the UK and elsewhere in Western Europe. Moscow and Novosibirsk were particularly active sites in the Soviet Union, as were Edinburgh and Sussex Universities in Britain. This happy collaboration was certainly helped by contacts made in 1960. One such contact, which had a great effect and led to many others, was with Prof. Anatoli Napalkov of the Biology Faculty of Moscow University. He took the initiative in making contact with me in 1960, and I learned about the interesting work of his group in applying cybernetic ideas in the study of animal learning of complex tasks.

It is interesting to note that the study of machine learning, along with its implications for the understanding of human and animal learning, is a "hot topic" at the present time. An important development has been a new view of "reinforcement learning"discussed in a recent book by Richard Sutton and Andrew Barto (1998). In this the environment is represented in the formalism of automaton theory, such that interaction with it produces both an immediate return and a change of state. An important new feature is the assignment of backed-up values to states in a way suggested by Bellman's "dynamic programming".

These developments can be seen as a breakthrough in the application of machine learning to a wide range of types of problem, but they stem from earlier work in various countries. That of Tsetlin in Russia was influential, among many others. There is a connection with the theories of the Napalkov group, since they also represented the learning process as a sequence of transitions between states, and thus as a finite automaton.

Of course, the study of machine learning is just one aspect of Cybernetics and Artificial Intelligence,and many others will be treated in reports in this meeting.

It should not be forgotten that the significance and implications of Cybernetics go well beyond automatic control and Artificial Intelligence, and the study of animal and machine learning, important though these topics are. Norbert Wiener, as well as Warren McCulloch and other pioneers of the subject, were concerned with very general issues affecting the quality of life, or what has often been called the "human condition". In particular, Norbert Wiener was very much concerned about the adverse social and ecological effects of modern technology, especially automated manufacture. He discussed the dangers at length, notably in his book The Human Use of Human Beings (1950), and it is significant that the title of the talk he chose to give in Moscow was "Science and society".

I would like to close by referring to another aspect of Wiener's work which is stressed in the definitive biography by Pesi Masani (1990). This is his generalisation of the second law of thermodynamics, according to which entropy or disorder tends to increase, so as to apply in many contexts including that of society. Masani has in fact referred to Wiener as "the first scientist-philosopher of the stochastic age".

It has been argued, for example by Adami (1998), that the defining characteristic of life is that it preserves the information in the genetic code over periods of time that exceed by many orders of magnitude the time a single DNA molecule would be likely to survive. Human society can similarly be seen as maintaining its characteristics in the face of a tendency to disorder. At the present time, when there is evidence of many slides towards chaos in the world situation, Norbert Wiener's analysis from the stochastic viewpoint is worthy of attention.