November 2005

(20 marks)
3. Define a linguistic variable and it’s value. How are linguistic variables determined and used in fuzzy expert systems?

A linguistic variable is a fuzzy variable. For example a the lingustic variable Tall in the context of “is person tall?”, represents a persons membership of that fuzzy set. They may be widely considered tall if their height is greater then say 175cm. In a fuzzy system in regard to membership, if John is 185cm then “if john is tall” may be more true then someone who is 180cm.
A linguistic variable is inherently a “fuzzy” variable. It represents data that is inexact or uncertain to some extent.
The value of a linguistic variable is how true it is, how accurately the word tall represents john’s height.

In the context of fuzzy expert systems:
When for an expert system the data and rules in the knowledge base requires it be represened with uncertainty or inexactness, a fuzzy hybrid system can be used for this task.
For example in a medical diagnosis fuzzy expert system, the symptoms of a patient may be describe with linguistic variables like “very painful”, “somewhat painful”, “dark color”, “light colour”.
Since an standard expert system requires exact logic, it would not adequately account for the uncertainty and vagueness of the data and rules. So a fuzzy system is used to evaluate all the inferencing processes.

(20 marks)
2. List and describe five main players in an expert system development team. Describe their roles in knowledge based system development respectively.

1. Knowledge Engineer:
   Collaborates exensively with the expert to extract the knowledge. They work with the expert and the end user to understand the most appropriate scheme for the expert system. They create the representation of the experts knowledge, and design the knowledge based system (eg, frame based, rule based, case based reasoning, etc).
2. Domain Expert:
   Usually a person who has worked extensively in a specific domain who has vast knowledge of it. The task of the knowledge based system is to emulate the expert to their level of competency in a narrowly defined task. The knowledge aquisition from an expert is considered a bottleneck in the development cycle.
3. End User: The person or group of people or things, that the expert system will be use by on completion. The knowledge engineer works with them to understand and build the definition of what is and isn’t required and possible for the expert system. They are an important source for testing and evalutaion during the developement cycle.
4. Programmer: The people who implement the the system based on the architecture specifications designed by the knowledge engineers. From the backend to the user interface. They implement the designs.
5. Manager: The person responsible for overseeing the project as a whole. Making sure the project deadlines are met, and collaboration between all project members is maintained. Keeps the project running towards the goal.

(20 marks)
1. Define Intelligence and Knowledge. Describe the main advantages of applying neural networks, fuzzy logic, and genetic algorithms in intelligent systems.

Intelligence is the ability to understand and learn concepts and ideas, and to solve problems.

Advantages of :-

Neural Networks:
Can be taught, developing the networks understanding of a problem on its own. Can find and use patterns, allowing large data sets to be interpreted without prior rules known.

Fuzzy logic:
Can represent ambiguity and uncertainty in data. Can give excellent results when “near enough” is more then enough. Can represent human reasoning and understanding with linguistic variables.

Genetic algorithms:
Can provide approximate solutions to extremely complicated problems, which may not be solvable at all algorithmically. Using evolutution you can ‘evolve’ a solution without having a vast rule base or prior case base (expert systems), or training data (neural networks).

1 exam to go. It’s going to be the clincher. This is rather important, and I wasted a whole week doing not much.
Maybe I’ll ask myself why when I get the time.

Last night Australia beat Uruguay and qualified for the 2006 soccer world cup in Germany.
It’s the first time Australia has qualified in 32 years.

Just shows to go you that John Saffran getting that curse removed was a good idea.

I’ve given up on planning to start my NaNoWriMo.. I don’t need the stress during this exam month. I think it would have been stressful even if I hadn’t delayed. But my exams had been even more pressing at the start of this month.

Ah well maybe next year?