Donald Michie (11 November 1923 – 7 July 2007) was a British researcher in artificial intelligence. He is best known for his pioneering work in machine learning, including the development of MENACE, one of the first learning programs, and for founding the Department of Machine Intelligence at the University of Edinburgh. His career spanned cryptography, genetics, and AI, and he remained active in research until his death in 2007.
Michie's contributions to Artificial intelligence were foundational, particularly in the areas of Machine learning and inductive reasoning. He was a proponent of the idea that machines could learn from experience, a concept that was not widely accepted in the early days of the field. His work at Bletchley Park during World War II placed him at the heart of early computing history, and he later became a leading figure in British AI research.
Early life and education
Michie was born in Rangoon, Burma, to British parents. He attended Rugby School, where he excelled academically, and won a scholarship to study classics at Balliol College, Oxford. In early 1943, seeking to contribute to the war effort, he attempted to enroll in a Japanese language course in Bedford for intelligence officers. On arrival, he discovered the course was not as expected, and instead he trained in cryptography, showing a natural aptitude. Six weeks later, he was recruited to Bletchley Park, the British codebreaking center, and assigned to the 'Testery', a section that tackled the German teleprinter cipher known as 'Tunny'. During his time at Bletchley Park, he worked alongside Alan Turing, Max Newman, and Jack Good, and he and Good were among the initial staff of the 'Newmanry', a section dedicated to machine-based codebreaking.
After the war, Michie returned to Oxford, studying at Balliol College from 1945 to 1952. He received his Doctor of Philosophy (D Phil) degree in 1953 for research in mammalian genetics, a field far removed from his later work in AI.
Career and research
In 1961, Michie developed the Matchbox Educable Noughts And Crosses Engine (MENACE), one of the first programs capable of learning to play a perfect game of noughts and crosses (also known as 'tic-tac-toe'). Since computers were not readily available at the time, Michie implemented the program using about 304 matchboxes, each representing a unique board state. Each matchbox was filled with colored beads, each color representing a different possible move in that board state. The quantity of a particular color indicated the 'certainty' that playing the corresponding move would lead to a win. The program was trained by playing hundreds of games, updating the bead quantities in each matchbox based on the outcome of each game. This early example of reinforcement learning demonstrated the potential of machines to improve their performance through experience.
Michie became director of the University of Edinburgh's Department of Machine Intelligence and Perception (previously the Experimental Programming Unit) from its establishment in 1965. The machine intelligence unit predated the university's computer science unit, making Edinburgh one of the earliest centers for AI research in the UK. Michie remained at Edinburgh until 1985, during which time he supervised numerous students and contributed to various aspects of AI, including Neural network research and the development of inductive learning algorithms.
In 1982, Michie left Edinburgh to found The Turing Institute in Glasgow, alongside Peter Mowforth and Tim Niblett. The institute was named in honor of Alan Turing and focused on AI research and applications. In 1984, the institute worked under contract from Radian Corp to develop code for the Space Shuttle auto-lander. The code was developed using an inductive rule generator called Rulemaster, which learned from training examples provided by a NASA simulator. This project was a notable early application of Machine learning to a real-world engineering problem. The Turing Institute was also involved in a series of advanced robotics projects, including robot navigation, robot sensing, and using robots to learn the naïve physics of handling objects through random play.
Michie remained active in the research community into his eighties. He devoted the last decade of his life to the UK charity The Human Computer Learning Foundation, working with Stephen Muggleton, Claude Sammut, Richard Wheeler, and others on natural language systems and theories of intelligence. In 2007, he was completing a series of scientific articles on the Sophie Natural Language System and a book manuscript entitled 'Jehovah's Creatures'. Michie also invented the memoisation technique, a method for optimizing computer programs by storing the results of expensive function calls and reusing them when the same inputs occur again.
Awards and honours
Michie was awarded numerous fellowships and honours during his career, including:
- Fellow of the Royal Society of Edinburgh (FRSE) in 1969
- Fellow of the British Computer Society (FBCS) in 1971
- Fellow of the Association for the Advancement of Artificial Intelligence in 1990
- Foreign Honorary Member of the American Academy of Arts and Sciences in 2001
- Corresponding Fellow of the Slovenian Academy of Sciences and Arts in 2005
These honours reflect his international reputation and his contributions to the field of artificial intelligence.
Personal life and death
Michie was married three times. His second marriage was to biologist Anne McLaren from 1952 to 1959. He had four children: one by his first wife, and three by McLaren, including economist Jonathan Michie and health psychologist Susan Michie. Michie and McLaren remained friends after their divorce and became close again after the death of his third wife. On 7 July 2007, Michie and McLaren were killed in a car crash when their car left the M11 motorway while travelling from Cambridge to London. Their deaths were a significant loss to the scientific community.
Legacy
The Donald Michie Papers are housed at the British Library, and can be accessed through the British Library catalogue. These papers provide a valuable resource for historians of computing and AI, documenting Michie's work and the development of the field. Michie's contributions to Machine learning and his early vision of machines that could learn have had a lasting impact, and his work is still cited in contemporary AI research. His founding of the Department of Machine Intelligence at Edinburgh helped establish the UK as a major center for AI research, and his legacy continues through the institutions and researchers he influenced.