Advertisement
RayBiotech
RayBiotech

Computerized Creativity

Scientific Discovery: Computational Explorations of the Creative Processes. Pat Langley, Herbert A. Simon, Gary L. Bradshaw and Jan M. Zytkow. The MIT Press, Cambridge, MA, 1987. 346 pp. $25 HB, $9.95 PB. One of the most commonly heard objections to artificial intelligence (Al) runs: "Well, you may be able to get a computer to play chess or diagnose illnesses, but a computer will never do anything really creative like write a good play or discover the theory of relativity." Scientific discovery

By | January 26, 1987

Scientific Discovery: Computational Explorations of the Creative Processes. Pat Langley, Herbert A. Simon, Gary L. Bradshaw and Jan M. Zytkow. The MIT Press, Cambridge, MA, 1987. 346 pp. $25 HB, $9.95 PB.

One of the most commonly heard objections to artificial intelligence (Al) runs: "Well, you may be able to get a computer to play chess or diagnose illnesses, but a computer will never do anything really creative like write a good play or discover the theory of relativity." Scientific discovery seems to involve a mysterious blend of intuition and insight that no precise computer model could capture. How could we begin to formalize the methods by which scientists discover new laws, theories and concepts?

This exciting book shows the fertility of a computational approach to scientific discovery. There is nothing mysterious about creativity for its authors. They view discovery as the result of the same general kinds of information processing that researchers in Al and cognitive science have studied in more mundane tasks such, as problem solving. The authors describe a series of computer programs that have been used to "discover" many of the most important scientific laws, including ones associated with the names of Boyle, Ohm, Kepler, Coulomb, Snell, Black and Joule. The BACON programs generate quantitative laws such as ones that describe patterns of chemical reactions.

The heuristics used by these programs are described in detail but without technicalities, so no special background is presupposed. Since the basic Al ideas that underlie the program are lucidly explained, the book provides a splendid entry for a scientist curious about the burgeoning field of artificial intelligence.

Although the authors' application of computational methods to many scientific discoveries demonstrates the potential of the in-formation processing approach, they are aware that much remains to be done. The data used by BACON and the other programs to generate laws are pre-selected by the programmer who knows what heuristics will operate on them in order to produce what laws. BACON has not made novel discoveries all on its own; rather, it is used to show how scientists might have made the discoveries that they did.

This "might" is quite weak, since little psychological or historical evidence is presented to show that the heuristics BACON uses to produce discoveries were ones used by actual scientists. My own reading of some of the historical cases is that the discoveries involved more complex processes than the ones these programs implement. Scientists often search for laws with the guidance of theories and suggestive analogies, rather than simply deriving laws from data.

Nevertheless, the authors provide an excellent guide to what has been done and to much of what remains to do in the Al approach to scientific discovery. Readers interested in the creative aspect of science should not miss this path-breaking study.

Thagard is research pathologist in the Cognitive Science Laboratory
at Princeton University, Princeton, NJ 08544.

Advertisement

Follow The Scientist

icon-facebook icon-linkedin icon-twitter icon-vimeo icon-youtube
Advertisement
RayBiotech
RayBiotech

Stay Connected with The Scientist

  • icon-facebook The Scientist Magazine
  • icon-facebook The Scientist Careers
  • icon-facebook Neuroscience Research Techniques
  • icon-facebook Genetic Research Techniques
  • icon-facebook Cell Culture Techniques
  • icon-facebook Microbiology and Immunology
  • icon-facebook Cancer Research and Technology
  • icon-facebook Stem Cell and Regenerative Science
Advertisement
The Scientist
The Scientist
Advertisement
The Scientist
The Scientist