Intelligent Systems And Their Societies Walter Fritz

Response Rules

 

Response rules are possibly the most important part of an IS. They are the knowledge that permits an IS to respond to a stimulus, and thus, to act in a given way to a given situation

In biology, a response rule is called a "response arc". It consists of a mapping unit that links a stimulus and corresponding response. In expert systems or knowledge bases, response rules are often called productions. Here they consist of one part that indicates when and where the response rule is applicable, and another corresponding part that contains the answer the computer should give in that case. So we see, response rules are used in both biological and in artificial ISs.
For a discussion on whether we think with words, see Do we think with words? ( Enter for continuous reading, like a book).

"During its existence, an IS continuously experiences: It finds itself in a certain situation and it acts. So that the IS has a chance of acting better in the future, it stores these experiential pairs of situation and response along with the results it observed from its response in a structure called a response rule. These are concrete response rules.

The "results of the action" of natural ISs are the emotions that are felt as a result of the response. They are the result of communications received by the brain from the various receptors of the body and consist of such messages as pain, hunger, tiredness, or the absence of these. In artificial ISs, however, the "emotions" that the brain receives can come from two sources. They can be signals from a person who is observing the responses of the IS, or they could be signals that are calculated from the summary of body variables such as a lack of power (fuel), the impact of bumping into an object, or even the awareness of a malfunction.
ISs, whether human, animal, or artificial, learn not only from their own direct experiences, but also from observing and copying the action of others.

In artificial ISs, response rules consist mainly of a situation statement and a response. There are often complementary features besides these two. For instance: the moment when the IS used the response rule for the last time; the pleasure or pain resulting from the application of the response rule; the positive or negative weight of the concepts in the situation statement.

Instincts are a particular subset of response rules that are present at the start of the existence of the IS. In humans they are present at birth and include such actions as holding on and sucking, both of which are responses that babies needed to stay alive, at least at some period of evolution. In artificial ISs, curiosity and other general response rules are sometimes seen as beneficial and are incorporated into the original programming of the system.

 

For continuous reading, like a book - continue here.
Jump to the top of this document / Details of IS / Intelligent Systems /
A Scientific Philosophy / E-book Contents.


Last Edited 22 July 2013 / Walter Fritz
Copyright © New Horizons Press