Behavior-Based Intelligence

Computer Science An Overview _J. Glenn Brookshear _11th Edition

Early work in artificial intelligence approached the subject in the context of explicitly

writing programs to simulate intelligence. However, many argue today that human

intelligence is not based on the execution of complex programs but instead by simple

stimulus-response functions that have evolved over generations. This theory of

“intelligence” is known as behavior-based intelligence because “intelligent” stimulus-

response functions appear to be the result of behaviors that caused certain individu-

als to survive and reproduce while others did not.

Behavior-based intelligence seems to answer several questions in the artificial

intelligence community such as why machines based on the von Neumann architec-

ture easily outperform humans in computational skills but struggle to exhibit com-

mon sense. Thus behavior-based intelligence promises to be a major influence in

artificial intelligence research. As described in the text, behavior-based techniques

have been applied in the field of artificial neural networks to teach neurons to behave

in desired ways, in the field of genetic algorithms to provide an alternative to the

more traditional programming process, and in robotics to improve the performance of

machines through reactive strategies.