Artifical Intelligence + Expert Sytem

Question 1

What is Artificial Intelligence

Answer 1

Machines can model the type of intelligence displayed by humans.
-Learning
-Decision Making
-Creativity
These allow the machines to make decisions automatically

Question 2

PPQ - Explain what is Artificial Intelligence (3)

Answer 2

Artificial Intelligence considers all aspects of intelligence and models these using computer systems
The aim is to create intelligent machines
… which can function/react like humans
… By learning and adapting

Question 3

PPQ - Explain what is the Turing Test

Answer 3

A test to determine if a machine behaviour is indistinguishable to a humans response.
The test involves a player, a human and a machine
The player does not know which is the human and which is the machine.
The player asks questions of the human and the machine and tries to discover which is the human

Question 4

What are the features of a Neural Network (Feedforward Network)

Answer 4

-Models the human brain
Units will be arranged in layers …Input, multiple hidden and output layers
-The number of units can vary
-Input layer units will receive information from the outside world e.g. units could be pixels on a camera, input=colour/brightness
-The hidden units will attempt to process inputs from the input units … each unit passes its value to every unit in the next layer
-Units use the input from the previous layer, along with a weighting for each input to perform a calculation to pass onto the next layer
-Units may decide not to pass its output onto the next layer
-Backwards propagation may be used for learning

Question 5

What is Backward Propagation

Answer 5

It gives the neural network a set of inputs, paired with the correct output value
The network runs inputs and compares its outputs to correct outputs
When don’t match, the network runs backwards from output to input … Modifying values in network
Instantly repeated until a minimum amount of errors
…machine has learned

Question 6

What is an Expert System

Answer 6

• In artificial intelligence, an expert system is a computer system that emulates the decision-making ability of a human expert.
• Expert systems are designed to solve complex problems by reasoning through bodies of knowledge, represented mainly as if-then rules rather than through conventional code.
• The three phases of an expert system are the user interface, Inference Engine and Knowledge Base

Question 7

Explain the three phases of an expert system

Answer 7

User Interface 
-Allows users to input queries
... Text boxes, buttons, options
-Where the user gets results
... Possibilities / decision
... Reason / Justifications

Inference Engines
Uses the user input to interrogate the knowledge base and draw conclusions
…Passed to interface
Can use fuzzy logic (yes, no, maybe)

Knowledge Base

• All world knowledge about a specific domain (medical, car engine fault diagnostic and life insurance)
• Contains all expert knowledge, Factual Knowledge, heuristics, rules etc…

Question 8

What is heuristical knowledge

Answer 8

This is knowledge gained through experience which won’t be found in a manual or textbook
Rule of thumb

Question 9

How could a mechanic help with the creation of a new expert system for engine diagnostic?

Answer 9

Could be interviewed

..Get heuristical information gained from experience which is inputted into the knowledge base

Question 10

What is Fuzzy Logic

Answer 10

Normal logic deals only with true/false
Fuzzy logic ca deal with probability
..How likely something is to happen
..Value between 0.0 and 1.0
E.g coin toss = heads = 0.5

Question 11

What is a Shell

Answer 11

The shell contains the structure for creating an expert system
…An empty expert system without a knowledge base
… The knowledge engineer can enter appropriate rules and facts to generate an expert system (forgiven domain)
…obtained from experts

Question 12

What is a Knowledge Engineer

Answer 12

Collect information for the knowledge base from the experts and other sources (books)
Formulate the knowledge into the rules and facts
…to be entered into the knowledge base
Technical experts who create expert systems

Question 13

What are the general advantages and disadvantages of an Expert Systems

Answer 13

Advantages

• All up to date knowledge … Some things humans may not know or remember
• Knowledge of experts can be recorded and used before they move on
• Provide consistent decisions (accurate)
• Expert advice is always available
• Decisions are free of human emotion
• Used as a training aid
• Does not get tired/overworked
• Provide reasons for decisions

Disadvantages

• Machines don’t learn from mistakes (unless Ai is being used)
• Humans become over-reliant on the system …leads to deskilling of humans
• Do not have human intuition
• Works best with a specific problem