- Hits: 4922
Donald Michie (see also entry for Ann McLaren for more on Michie and for these source details) was born in Rangoon on born November 11 1923. He attended Rugby school and was awarded an open scholarship to study classics at Balliol College, Oxford, in 1942. Instead, in 1943, inspired by his father to do "something unspecified but romantic" behind enemy lines in China, Michie attempted to enrol on a course for intelligence officers in Japanese. On arrival at the School of Codes and Ciphers in Bedford, he was told that the course was full, and decided instead to take up training in cryptography. A fast learner, he was soon recruited to Bletchley Park in Buckinghamshire, and was assigned to the "Testery", a section working on solving the German high-level teleprinter cipher, code-named Fish. The work at Bletchley Park has now been celebrated in book and film.
There, he developed code-breaking techniques which led to effective automatic deciphering of German high-level ciphers. Declassification of once secret documents now makes clear how profoundly important this wartime research was. In April 1944 he invented a technique for using the Colossus computer, developed at Bletchley, to automatically decode the secondary wheel of the Lorentz machine, which the Germans used for encoding Fish.
The innovation, tested by Michie and Jack Good, endowed the machine with a degree of general-purpose programmability and led to a radical last-minute enhancement in the construction of Colossus II. The results were dramatic. Texts which previously had taken days to decipher could now be completed in hours, allowing repeated effective interception of enemy attacks.
During this period at Bletchley, Michie held frequent lunchtime discussions with Alan Turing on the possibility of building computer programs that would display intelligence. Before the war, Turing had developed the mathematical basis for modern digital computation, and was applying the principles he had developed in the decoding efforts at Bletchley. Both Michie and Turing were interested in programming computers to play chess, as well as developing programs which could learn automatically from experience.
He was subsequently a pioneer in the field of machine intelligence - loosely, creating "computers that can think." He developed the earliest variations, at one time with matchboxes and beads, of the machines upon which millions across the world now daily depend.
Following the end of the war, Michie decided to take up his offer from Oxford. His wartime experience had diverted his former interest in classics into a passion to study science. Supported by a Balliol College war memorial studentship, he received his MA in human anatomy and physiology in 1949. During his subsequent DPhil degree at Oxford, Michie put his boyhood hobby of breeding pet mice to work in a series of genetic studies published in the journal Nature.
His first marriage, to Zena Davies, had ended in divorce in 1949 but, in Oxford, he married his fellow student Anne McLaren (see separate entry) in 1952 and, the following year, received his DPhil in mammalian genetics. Donald and Anne shared a lifelong commitment to socialism: both joined the Communist Party during the cold war, and from the world peace congresses of the 1950s to the recent anti-war demonstrations they were always there, often together. Donald was for years in the 1950s the Daily Worker's science correspondent, drawing on his broad-based understanding of the sciences, as well as the Marxism that was always his intellectual template. Although he left the Communist Party in 1956, Marxism remained his intellectual template and he was never other than partisan on every scientific or social question to the end of his life. Donald wrote a fierce polemic against Jack Straw's Iraq war lies in 2003, published by the Stop the War Coalition.
In their early days together, he and Anne went on to work on techniques related to in vitro fertilisation, first at London University and then at Edinburgh. In the 1950s, he and Anne worked on pioneering techniques which were fundamental in the development of in vitro fertilisation but they were divorced in 1959. Even so, their lives would wind back and fro together over the next decades even unto their joint deaths in a car accident in 2007.
While working at the department of surgical science in Edinburgh, Michie co-wrote one of the first introductory textbooks on the new science of molecular biology but, now separated from Anne, he moved his area of interest and subsequently became one of the founders of the field of artificial intelligence, to which he devoted the remainder of his academic career.
From 1960, his attention returned to his wartime discussions with Turing, and in particular the question of whether computers could be programmed to learn from experience. For demonstration purposes, he developed a noughts-and-crosses playing machine called Menace, for which he developed a general-purpose learning algorithm called Boxes. Since no computers were then available to him, he hand-simulated the Boxes algorithm, using a device made from an assembly of matchboxes. By 1963, Michie had assembled a small artificial-intelligence research group at Hope Park Square in Edinburgh. With the support of the Edinburgh vice-chancellor, Sir Edward Appleton, Michie established the experimental programming unit in 1965.
In 1966 he was joined in Edinburgh by Richard Gregory and Christopher Longuet-Higgins, both interested in the development of a brain research institute. The following year, he was appointed to a personal chair of machine intelligence and became the first director of the department of machine intelligence and perception. The period up to 1973 is widely perceived as one of the most fertile in the history of artificial intelligence research, and its history is documented by the frequently cited Machine Intelligence book series of which Michie was editor.
His crowning achievement was the development, under a team he led, of Freddy II, the world's first demonstration of a laboratory robot capable of using computer vision feedback in assembling complex objects from a heap of parts. Unfortunately, a series of events conspired to bring this period of rapid achievement to an end. Disagreements concerning the priorities of the field broke out between Michie, Longuet-Higgins and Gregory. At the same time, the growing economic crisis at the beginning of the 1970s was cutting into the budget of the Science Research Council, which was starting to look for savings.
Sir James Lighthill, a well-known British fluid dynamicist, was commissioned by the Science Research Council to analyse the prospects for the high-cost robotics project in Edinburgh. The resulting report, published in 1973, called a halt to artificial intelligence research in all but two areas.
The robot program was discontinued with knock-on effects for similar research programs in the US. The resulting dissolution of Michie's research group in Edinburgh left him isolated in the research unit. There he continued his research studies into computer chess and machine learning for the remainder of the 1970s.
By the early 1980s, automated assembly robots in Japan were outstripping traditional methods of manufacturing in other countries including the UK. Additionally, computer systems which imitated the decision-making of human experts were becoming increasingly successful. As a consequence, governments in the UK, Europe and US resumed large-scale funding of artificial-intelligence projects in response to the Japanese Fifth Generation project.
In 1986, as head of the Turing Trust in Cambridge, Michie founded the Turing Institute in Glasgow, in honour of his former colleague's key contributions to the field. Under Michie's leadership, the institute conducted advanced, industrially oriented research in machine learning, robotics and computer vision.
Following his retirement in the early 1990s, he continued actively in research on machine learning with his third wife, Jean Hayes-Michie. They had married in 1971, but she died from cancer in 2002, after which he resumed his friendship with Anne, with whom he died in a car accident on July 7th 2007. Michie, who was 83 when he died, was survived by his son Chris, from his first marriage, and by his daughters Susan and Caroline and son Jonathan from his marriage to Anne.