|
| |
Techniques for modeling knowledge in software
Artificial Intelligence
If we want software tools that question and advice the
users, the knowledge must be formally coded in a knowledge base, mainly using techniques
from Artificial Intelligence.
Roughly, one can say that there are 3 important kinds of formal knowledge expressions:
- Language.
These are modelled in formal language and symbolic schemes. We give an example of this:
IF the screen DOES NOT WORK
AND the power OF the screen IS CONNECTED TO the
computer
AND the computer WORKS
THEN the screen IS DEFECT
- Databases of examples of mental reflexes.
Neural networks is the technology to drill software to learn from
examples (recognising colours for example). For simple cases, statistics such as linear
regression can also be used.
- Some knowledge is expressed in vague language, e.g.
"When there is a sudden obstacle, you must gear quickly but moderately".
These fuzzy terms are difficult to explain in words. However, one can give examples of
what is quick, not quick, almost quick, etc. Fuzzy logic is
the technology to model those vague words.
Examples:
|
