Herbert Simon
(1916-2001)
Herbert Simon won a Nobel-prize in economics in 1978, but he was also a pioneer in artificial intelligence, decision making, cognitive psychology and computer science. With Allen Newell he built computer programs to simulate human cognition, especially ones that would solve problems with a trial and error ("generate and test") system that they called the General Problem Solver.
It was similar to a
two-stage model of free will and was part of the inspiration for
Daniel Dennett's "Valerian" model of decision making.
Newell and Simon also created a "blackboard" architecture, in which various items could be written to a blackboard which was visible to other systems that could read and write on the blackboard. As a model for human consciousness, the blackboard resembles the Theater of Consciousness and Global Workspace Theory of
Bernard Baars. A blackboard architecture is an important part of most computer "expert systems."
Simon studied how people learn. In the extreme cases of geniuses and idiot savants, he argued that over a ten-year period, the brain could memorize vast numbers of patterns that would make them instant experts. He estimated that chess masters had learned 50,000 chess board positions using the same memory that allows humans to recognize individual faces.
Simon was an optimist about intelligent machines. He predicted in the mid-1950's that computers would excel humans at chess within a decade, something that did not happen until the end of the century.
Satisficing
In 1958 Simon introduced the motion of "satisficing" in decision-making as an alternative to the principle of maximizing utility that was standard in neo-classical economics.
The model of
homo oeconomicus was a rational being who maximized utility. But economic decisions are always made with limited time and incomplete access to all the possible information, so it is expedient to be
satisfied with a choice that suf
fices some criteria that are easily identified.
In the age of digital computing, it became clear that some problems were too difficult to calculate solutions in reasonable computing times. Simon showed that many natural problems are computationally intractable or lack some information, either of which preclude mathematical optimization.
The Architecture of Complexity
For Teachers
To hide this material, click on the Normal link.
For Scholars
To hide this material, click on the Teacher or Normal link.
Thomas Sturm reseach on Computationalism
|
