From: sarah.student@gmail.com   
      
   Ted Dunning writes:   
   > On May 1, 6:43 am, sarah  wrote:   
   > > hello there   
   > > I am a beginner in AI. I don't know what does knowledge extraction   
   > > tool means. In addition I don't know the meaning of Individual in AI.   
   > > I want to know the meaning of Rule Discovery,too. Can any one explain   
   > > these words to me in simple sentences?   
   > >   
   >   
   > Ahh my....   
   >   
   > My guess is that nobody will be able to fully explain these terms in   
   > simple sentences because AI is still a very formative field even after   
   > 50 years or more.   
   >   
   > The terms you are asking about come from a subset of AI in general   
   > that is concerned with the encoding of apparently intelligent behavior   
   > in terms of rules.  These rules generally take the form of "if   
   > something then something else".  The something can be any condition   
   > that might be observed by the computer, the something else can be an   
   > action taken by the computer, or simply a conclusion that could be   
   > used later in other rules.   
   >   
   > The simple fact is, these rule systems are hard to build.  That is,   
   > they are easy to start, but hard to maintain once they get to a   
   > certain size.   There are many reasons for this, but the primary one   
   > is that rules based on hard and fast inference are highly susceptible   
   > to problems when contradictions are found.  As well, there can be   
   > subtle inferences that drive surprising (and often totally wrong)   
   > inferences.  It is incredibly hard to build a rule system that is not   
   > subject to these problems and which still has interesting and   
   > substantial uses.  Ultimately, all previous systems of this sort have   
   > been subject to a break-even point were every additional fix to   
   > improve behavior in some situation causes as many problems in other   
   > situations.  Once this point is reached, it is essentially impossible   
   > to improve the system by simply adding or modifying rules.   
   >   
   > The tools that people use to build these systems have as their primary   
   > goal to make this break-even point be as high as possible and to   
   > develop a set of rules that has useful behavior as quickly as   
   > possible.  Somewhat pretentiously, the discipline of building these   
   > systems is known as "knowledge engineering" and the process of   
   > eliciting rules from previously observed data or from domain experts   
   > is called knowledge extraction.  A knowledge extraction tool is   
   > something that makes knowledge extraction more feasible.  Rule   
   > discovery is usually used to describe the extraction of rule-based   
   > knowledge from observed data rather than from experts.   
   >   
   > Keep in mind, though, that knowledge-based tools are simply one kind   
   > of AI of many.   
   >   
   > Two good starting points are the wikipedia article on AI and Peter   
   > Norvig's book on the subject.   
   >   
      
   Dear friend   
   Thanks a lot for the explanation. But can you give me an example to   
   help me understand the worlds better?   
   By the way can you add the term "Building block approach for genetic   
   programming" to my list too?   
      
   [ comp.ai is moderated ... your article may take a while to appear. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)