The story of legends who were humble and spearheaded innovation (rather than the current "business", non technical, marvel loving morons with no sense of common sense):
The Conjecture
The document that named artificial intelligence opens with a single bold sentence: "The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it." Nobody in 1955 had established this conjecture. Nobody has established it since.
The document was a grant proposal, dated 31 August 1955 and submitted to the Rockefeller Foundation two days later. It asked for $13,500, ten participants, and two months at Dartmouth College in Hanover, New Hampshire. It introduced, for the first time in print, the phrase "artificial intelligence." The summer workshop it funded was smaller, shorter, and less conclusive than seventy years of retelling have suggested. Its name, however, has outlasted everything.
The Proposal
The proposal opened with its conjecture and then listed seven research problems: automatic computers and how to program them; how to make a computer use language; neuron nets; a theory of the size of calculations; self-improvement; abstraction; and randomness and creativity. The document did not say how anyone would know whether progress on any of these had been made. It said that significant advances were likely if a carefully chosen group worked together for a summer.
Of the four signatories, Shannon was by some distance the most famous scientist on the page: he had already founded information theory, was co-editing a volume with McCarthy, and had results in cryptography and the theory of Turing machines. Minsky, a Harvard Junior Fellow in mathematics and neurology, had built a machine for simulating learning by nerve nets and completed a Princeton PhD thesis entitled "Neural Nets and the Brain Model Problem." Rochester was Manager of Information Research at IBM, jointly responsible for the design of the IBM Type 701, who had also worked on simulation of nerve nets to test theories in neurophysiology. McCarthy, the youngest, was an assistant professor of mathematics at Dartmouth who had worked on Turing machines, the speed of computers, and the use of languages by machines. The contrast between Shannon's standing and McCarthy's juniority is worth holding in mind: the youngest and least established man in the room had picked the name and written the central conjecture.
[---]
What endured was not a research programme but a label and a concept. The name "artificial intelligence" gave a new field a clear identity distinct from cybernetics, from automata theory, from complex information processing. Departments could take the name, and did. Funding lines could adopt it, and did. Within two years, McCarthy and Minsky had started the MIT Artificial Intelligence Project; McCarthy later founded the Stanford AI Laboratory in 1962 and won the Turing Award in 1971; Newell and Simon won the 1975 Turing Award for the work that had its public debut in Hanover.40 Rochester's enthusiasm for Minsky's geometry-machine idea, taken back to IBM, set Herbert Gelernter working on a geometry theorem prover, with McCarthy as a consultant, and Gelernter presented a major paper on the result in 1959.41 None of this was the result of the summer's collective work. It was the result of those weeks of conversation, disagreement, and proximity.
No comments:
Post a Comment