Download Rule Based Expert System - Artificial Intelligence - Past Exam and more Exams Artificial Intelligence in PDF only on Docsity!
Cork Institute of Technology
Master of Science in Computing in Education
National Framework of Qualifications – Level 9
February 2005
Artificial Intelligence
(Time: 2 Hours)
Answer question any TWO questions (All questions carry equal marks).
Examiners: Mr. P. Rothwell Dr. H. Gibbons
- (a) Briefly define an expert system. [4 marks] List and explain the components of an expert system. [4 marks] Discuss ‘conflict resolution’ in a production rule-based expert system. [4 marks]
(b) Outline and explain the physical symbol system hypothesis. [7 marks] Discuss this approach to AI with reference to logic-based and Neural Network representations of knowledge. [6 marks]
(c) The Frame Problem is seen as a serious impasse for logic-based knowledge representation. Explain the frame problem using examples. [10 marks]
Hubert Dreyfus criticises attempts to represent common-sense knowledge using a formal symbolic language. Outline the argument he gives in support of his criticism. [15 marks]
[Total: 50 marks]
- (a) Using the following program and query as an example write a general description of Prolog in terms of its syntax, list handling, and the use of recursion.
prog([], L, L). prog(L, [], L). prog([H1 | T1], [H2 | T2], [H1, H2 | T3]):- prog(T1, T2, T3).
?- prog([a, b, c, d], [1, 2, 3], R). [18 marks]
(b) Write and fully explain Prolog programs to define ONE of the following list handling predicates whose behaviours are described below (you may use system defined predicates):
i) last(Item,List) to check if Item is the last element of list L. e.g.s ?- last(4, [1,2,3,4]). ?- last(2, [1,2,3,4]). yes no
ii) intersect(A, B, C) which assigns to list C the intersection of the lists A and B. E.g. ?- intersect([a,b,c], [d,e,b,c], L). L=[b,c] [14 marks]
(c) Write and fully explain a Prolog program that would solve the following problem. On Monday, Tuesday, and Wednesday last the mad hatter wore a different style of hat on each day and each hat was a of a different colour. Given the following information find out the colour of the trilby. The fez was worn on Monday or Tuesday. The trilby was worn after the red hat. The bowler is black. The beige hat was last. The trilby was not worn on Tuesday. [18 marks] [Total: 50 marks]
- (a) With the aid of examples and diagrams, fully describe the central notions of artificial neural network systems. [15 marks]
(b) With the aid of diagrams, describe the operation of backpropagation learning in artificial neural network systems. [15 marks]
(c) The delta rule for calculating the change in the weight of the i th^ connection to the j th^ output layer node is: ∆Wij = c δj Xi where δj = (d (^) j – Oj ) f’ (net (^) j ). The similar rule for the change in the weight of the i th^ connection to the k th hidden unit which feeds its output to the j th^ output unit is ∆Wik = c Σj (wkj *δj ) f’ (net (^) k ) Xk Fully explain the contribution of each element of these rules to the backpropagation algorithm described in part (b) above. [20 marks] [Total: 50 marks]
