Glushkov: Introduction to Cybernetics
In 1964 the publishing house of the Ukrainian Academy of Sciences in Kyiv issued Victor Glushkov's Vvedeniye v kibernetiku (Introduction to Cybernetics), 324 pages. It brings together the theory of algorithms and automata, Boolean algebra, self-organising systems and perceptron training, digital computers and Algol-60, predicate calculus and the automation of proofs. Two English translations appeared in 1966.
Why it matters
The Kyiv school set out in one textbook what the atlas spreads across several lines: the perceptron as a machine that learns to recognise, and automatic theorem proving as the start of automating scientific creative work. English translations made the book available outside the USSR within two years.
The Kyiv original of 1964 itself was read: its title page names the Academy of Sciences of the Ukrainian SSR and its Scientific Council on Cybernetics, and the colophon on the last printed page, 324, gives printing by resolution of that council, a print run of 15,000 and signing to press on 7 December 1963. The original is in Russian; the publisher is in Kyiv. The full English translation by the US Air Force Foreign Technology Division (FTD-TT-65-942, 17 February 1966, 469 pages) was read as well. A second, separate translation was made by Scripta Technica for Academic Press: New York, 1966, 322 pages. The text contains a theory of training discrete alpha-perceptrons citing Rosenblatt, heuristics that cut the search in proof finding, and the Institute of Cybernetics' experiments in teaching a machine to tell meaningful from meaningless phrases. What the record does not claim: that the book proved human thinking could be modelled, as the research report behind the candidate said. The phrases artificial intelligence and thinking occur zero times in the translation, though its abstract names the automation of thought processes among the branches of cybernetics. Of many results in mathematical logic and algorithm theory the abstract says they are given in summary form, without detailed proofs and sometimes without any proof. The report also gave Library of Congress number 66017287; the catalogue record has 66016438.