AI Becomes An Industry 1980

The first successful commercial expert system, R I, began operation at the Digital Equipment Corporation (McDermott, 1982). The program helped configure orders for new computer systems; by 1986, it was saving the company an estimated $40 million a year.

 By 1988, DEC 7 s AI group had 40 expert systems deployed, with more on the way. Du Pont had 100 in use and 500 in development, saving an estimated $10 million a year. Nearly every major U.S. corporation had its own AI group and was either using or investigating expert systems.

In 198 1, the Japanese announced the  " Fifth Generation "  project, a 10 - year plan to build intelligent computers running Pro log. In response the United States formed the Microelectronics and Computer Technology Corporation (MCC) as a research consortium designed to assure national competitiveness. In both cases, AI was part of a broad effort, including chip design and human - interface research. However, the AI components of MCC and the Fifth Generation projects never met their ambitious goals. In Britain, the Alvey report reinstated the funding that was cut by the Light hill report.

Overall, the AI industry boomed from a few million dollars in 1980 to billions of dollars

in 1988. Soon after that came a period called the  "A1  Winter, "  in which many companies

suffered as they failed to deliver on extravagant promises.

The Return of Neural Networks: 

Although computer science had largely abandoned the field of neural networks in the late 1970s, work continued in other fields. Physicists such as John Hop field (1982) used techniques from statistical mechanics to analyse the storage and optimisation properties of networks, treating collections of nodes like collections of atoms.

Psychologists including David Rumelhart and Geoff Hnton continued the study of neural - net models of memory. As we discuss in Chapter 20, the real impetus came in the mid - 1980s when at least four different groups reinvented the back - propagation learning algorithm first found in 1969 by Bryson and Ho.

 The algorithm was applied to many learning problems in computer science and psychology, and the widespread dissemination of the results in the collection  Parallel Distributed Processing  (Rumelhart and McClelland, 1986) caused great excitement.

These so - called  connectionist  models of intelligent systems were seen by some as direct competitors both to the symbolic models promoted by Newell and Simon and to the logicist approach of McCarthy and others (Smolensky, 1988). It might seem obvious that at some level humans manipulate symbols - in fact, Terrence Deacon's book  The Symbolic Species  (1997) suggests that this is the  de jining characteristic  of humans, but the most ardent connectionists questioned whether symbol manipulation had any real explanatory role in detailed models of cognition. This question remains unanswered, but the current view is that connectionist and symbolic approaches are complementary, not competing.