Areas Of AI
There are many areas of AI that don't belong anywhere else on this site and they will be explained here.
- Fuzzy Logic
Fuzzy Logic
Fuzzy Logic was created to account for the grey area of Boolean logic. Where in Boolean logic we can only have two values - True or False ( also represented as 1's or 0's ) - Fuzzy Logic represents the grey area between these two values.
Fuzzy Logic allows the computer to use common sense, in real life an answer does not always only have a true or false answer, so fuzzy logic allows the machine to account for this "partial truth" answer.
Natural Language Processing
Voice Recognition programs not only allow people with poor typing abilities to use computers more efficiently but they also let the disabled use computers without keyboards.Voice recognition programs can only dictate what the human is saying, the AI in the machine translates what it hears, not understand what is said. For the computer to understand what is being said is a different field of AI called Natural Language Processing is used.
The Goal of Natural Language Processing (NLP) is to build a computer that will understand the natural languages (such as English or French) that humans can understand so effortlessly, to create a computer that we can talk to just like another human being.
Just like all other AI fields NLP is still a challenge to researchers. The sentence "She is good looking" to many humans is that a particular woman is pretty, but computers could read that sentence as "She is good at looking" as the computer looks at looking as sighting or seeing, not the way she is presented to males. To overcome this problem researchers teach the computer to look at the existing text along with a online dictionary and encyclopaedia.
There is also a Formal Natural Language which is a subset of our natural language, this helps the computer to understand words a bit better, this along with teaching the computer to read existing text and a online dictionary helps the computer to understand what a human can say and brings us on step closer to overcoming NLP.
Machine Learning
Teaching a machine to learn form past experiences and other means means that the machine can continually self-improve and then increase efficiency and effectiveness, just like humans do.I believe machine learning will need to combine with neural networks, robotics and expert systems and use their theories to make a machine learn, and therefore trying to make a machine learn before neural networks has been accomplished is pointless, and of course every AI area is no where near perfect and probably wont be for years to come.