ALEX ANDREW graduated B.Sc. (Hons.) from Glasgow University in Mathematics and Natural Philosophy, his course having been interrupted for three years by direction to wartime electronics research. He joined the Departments of Physiology and Psychological Medicine at Glasgow University as Research Assistant, and from there spent eighteen months on leave of absence in 1954-55 with Warren McCulloch's group in the Massachusetts Institute of Technology. His work in Glasgow was largely on electronic instrumentation for neurophysiology, and also on physiology as such, and in MIT was on visual system of the frog. He graduated Ph.D. from Glasgow with a thesis on work related to neurophysiology.
He worked on machine learning at the National Physical Laboratory, Teddington, in the Division headed by A.M. Uttley, and gave a paper at the first IFAC Congress, Moscow 1960. Then he went to the management consultancy firm SIGMA, managed by Stafford Beer, for two years, followed by a year in Heinz von Foerster's group at the University of Illinois, and then by a Lectureship in Cybernetics in the University of Reading from 1965 to 1982.
After taking early retirement from Reading, he set up the firm Viable Systems with his wife Joyce, and continued his research interests. The academic year 1987-88 was spent in the University of Aegean, Izmir, Turkey, teaching Artificial Intelligence.
He was appointed Director for International Affairs of WOSC (World Organisation of Systems and Cybernetics), and elected to Honorary Fellowship in 1997. Also acting as WOSC Webmaster, he is now also Internet Editor and author of the regular feature Internet Commentary of the WOSC official journal Kybernetes, and an active book reviewer for that journal.
A paper written in 1965, while with Heinz von Foerster's group in Illinois, introduced a principle of 'significance feedback' of which one version is equivalent to that of 'backpropagation of error', which came later and is the basis of most of the work on artificial neural nets in recent decades. The paper is:
A.M. Andrew, Significance Feedback in Neural Nets. Report of Biological Computer Lab., Univ. of Illinois, Urbana, 1965. Reprinted in: Int. J. Systems Research and Info. Science, vol. 6, pp. 59-67, 1993.
A particular theme of interest of his has been the place of metrical continuity in A.I. and in biological information processing. This began with the early work on machine learning in the National Physical Laboratory, where the only way to solve certain problems seemed to be by use of continuous variables, although this was regarded as uninteresting by enthusiasts for A.I. Two papers that marked the beginning of this interest were:
A.M. Andrew, "Learning machines". In: Mechanisation of Thought Processes, HMSO, London, 1959, pp. 473-505.
A.M. Andrew, "Self-organizing control mechanisms and some principles for more advanced learning machines." In: Automatic and Remote Control (IFAC Moscow Congress), Butterworth, London, 1961, pp. 818-824.
He has written the following books:
Brains and Computers, Harrap, London, 1963
Artificial Intelligence, Abacus, Tunbridge Wells, 1983
Computational Techniques in Operations Research, Abacus, Tunbridge Wells, 1985
Self-Organizing Systems, Gordon and Breach, New York, 1989
Continuous Heuristics, Ellis Horwood, Chichester, 1990
A Missing Link in Cybernetics: Logic and Continuity, Springer 2009
Chapter "Parallel and Distributed Processing in Cybernetics and Applied Systems ed Constantin Virgil Negoita, Marcel Dekker Inc, New York, 1992
Recent papers given at the Society Loose Ends in Cybernetic Thinking and Notes on the continuation of discussion his 2007 talk "In Defence of Strong AI"
Further links: Tiscali Homepage. Robert Vallee net (Past President of WOSC)
Eulogy delivered by Dr. G. K. Wallace, Senior Lecturer (retired) at Reading Crematorium, All Hallows Road, Monday June 10th 2013.