- Created by: Vicky
- Created on: 10-05-12 18:08
Artificial intelligence is a reasoning process performed by computers, which allows the computer to:
- draw deductions
- produce new information
- modify rles or write new rules
the computer, just like a human, is able to learn as it stores more and more data.
Neural networks are biological system that are used by the brain for learning new things. By understanding how the brain works, scientists can develop ICT systems making use of artificial neural networks that mimic the way the brain works. The main advantage of artificial neural networks is that they can learn by example just like the human brain. Which means they are useful for pattern analysis or data classification
Ordinary ICT sytems are good at :
- fast processing of data
- obeying a set of instructions given as the program code.
Ordinary ICT systems are not good at:
- adapting to circumstances
- dealing with data in paralled (I.E computers like to process data in linear fashion)
- dealing with data that is imprecise or contains errors.
Neural network advantages
- they are good where algorithms cannot be developed i.e. where it is difficult to develop a computer program because the nature of the problem being solved is not understood enough.
- they are good where there are plenty of examples that the system can learn from.
- they are good where a structure can be identified from exsisting data
Ict systems making use of neural networks have the following disadvantages , these are :
- they are only suited to certain tasks
- the examples used to teach the system must be chosen carefully, otherwise time is wasted and the system can produce unpredictable results
- because the neural network learns on its own, it's operation can be unpredictable.
What is an expert system?
An expert system is an ICT system that uses artificial intelligence to make decisions based on data supplied in the form of answers to questions. This means that the system is able to respond in the way that a human expert in the field would come to a conclusion. A good expert system is one that can match the performance of a human expert in the field.
There are 3 main components of an expert system.
- knowledge base - a huge organised set of knowledge about a particular subject. It contains facts and also judgemental knowledge, which gives it the ability to make a good guess, like a human expert
- inference engine - a set of rules on which to base decisions and most of these rules have the "if-then" structure. It is the part of the expert system that does the reasoning by manipulating and using the knowledge in the knowledge base.
- A user interface - the user interface presents questions and information to the operator and also recieves answers from the operator.
How expert systems work
expert systems use a problem solving model that organises and contols the steps that need to be take to solve the problem. one part of this model is the use of "IF-THEN" rules to form a line of reasoning. these problem solving methods are provided as part of the inference engine.
The knowledge base is all the information an expert uses when arriving at a decision or diagnosis, the more knowledge and experience a human expert has in the field. Then the more knowledge can be added to the knowledge base. Because knowledge is often incomplete or uncertain, a rule can sometimes have a confidence factor or weight.
A user of the expert system is asked a series of questions which may include the results of tests and on the basis of the answers the expert can arrive at an answer or diagnosis
Advantages and disadvantages of expert systems
- consistency - they provide consistent answers for repetitive decisions
- cheaper - they are cheaper than using a human expert such as a doctor or consultant
- the expert system can consult a much larger pool of knowledge compared to a human expert
- the expert system is available 24/7 and is never on holiday or sick when needed.
- the computer uses all the information that it has, unlink a human expert who may forget and make mistakes
- can make absurd errors - if data is incorrectly input, for example a person's age and their weight are swapped around- then absurd doses of drugs could be given.
- not able to create a creative repsonse. Human experts can produce creative responses in certain situations,where as an expert system would be unable to do this.
- not being able to realise when no answer is available to a problem.