|Intelligent Systems And Their Societies||Walter Fritz|
Structure Of The Functioning Of An IS
The easiest way to present an overall view of structure is with a representative diagram. We have constructed one for an IS and present it below .
As you can see from this diagram, the IS is fundamentally a type of stimulus - response system. The stimulus is the sum of the communications entering through the senses. The brain extracts information from this and represents it as a situation.
Next, the IS selects a response rule, appropriate to the situation, and performs the response part of this rule. Here we mean by "appropriate" that performing the response permits the system to get nearer to the situation that is its objective.
The IS makes its selection of response rules from those that it finds stored in its memory. In this memory, the IS has accumulated response rules that it has generated from earlier experiences and from generalizations based on previously elaborated response rules.
"Stimulus -- Response"
Many researchers have recognized "stimulus and response" as the fundamental mechanism in animal and human activity. Some situations are found desirable, they are an objective. The perceptual control theory states that often the animal and the human being acts in such a way that it changes the situation that it perceives to the situation that is its objective (see Perceptual Control Theory) (Exterior link).
But are these also the fundamental mechanisms in the human mind? At first, it does not seem possible that all (or nearly all) of the activity of the human brain could be something as simple as a type of sophisticated stimulus - response mechanism, used to reach an objective. However, we need to keep in mind that all the incredible, and varied activity that we can observe in computer systems today are based on only a few very simple capabilities, namely: to add, subtract, compare, and jump to a different place in the program. Thus, a very complex response, based on a sequence of only a few elementary activities, is perfectly possible.
By generalizing and abstracting from our experiences we can learn more general response rules. It turns out that it is this learned mental activity, that permits us to think and do so many things. We have learned how to walk, how to do arithmetic, how to write, how to dress and to behave, and even how to create intelligent learning systems. To a large extent we have learned how to think, how to make plans, how to extrapolate. This learning process starts at birth and goes on intensively for many years until the end of schooling. Even then the learning never stops.
Need for a Complete System
Researchers in the science of "Artificial Intelligence" have investigated many areas of the mind such as pattern matching, vision, and theorem proving. However, all of these are only parts of the human mind. An intelligent system could include all of these parts, but it still would not be complete, and could not function, unless it also had senses, a method to choose responses according to its objectives and memories, and some way of performing these responses in and on its environment.
Programming versus Learning
Since it is impossible to foresee all of the many different situations in which an artificial IS may eventually find itself, we should not attempt to program its responses. It is better if, as it goes along its way through its "life", the artificial IS learns how to act based on its own past experiences.
There are, though, a few useful exceptions to this rule. For example, an IS will greatly benefit from a few of the most basic instincts. For starters, I would include the following:
For continuous reading, like a book - continue
Jump to the e-book Contents / Intelligent Systems / top of this document.
Copyright © New Horizons Press