Name: ___________________

Student No: ___________________

Faculty: ___________________

Signature: ___________________

THE UNIVERSITY OF QUEENSLAND

FINAL EXAMINATION, SEMESTER II, NOVEMBER 2000

IC231

LABORATORY INTRODUCTION TO MODELS IN COGNITIVE SCIENCE

Time: Two (2) hours for working

Ten (10) minutes for perusal before examination begins

General Directions to Candidates

There are twelve questions on the paper. You should attempt all questions. Each one is worth 10 marks (Total marks 120).

The examination has been set to take the full two hours, so answer the questions that you know first, and only after answering them, go back to the questions that you are less certain about.

Answer all questions on the examination paper. If you need additional space use the back of the paper and clearly label each answer.

Question 1

marks

a) Consider an interactive activation and competition (IAC) network that represents pools of features for {coffee, tea or OJ}, {croissants or pancakes}, {banana or mango}. Draw an IAC network that will represent the patterns for (1) alice’s breakfast {coffee, croissants, mango} and (2) bob’s breakfast {OJ, pancakes, banana}.

5

b) How many positive and negative weights does your network have? Explain your answer.

2

c) If another 10 breakfast patterns were added to your IAC network, it would be possible to use the network to generate prototypes. Explain how you would use the network to generate the typical breakfast of coffee drinkers.

3

Question 2

marks

a) At a general level, what roles can simulation play in theories of real-world phenomena?

3

b) What aspect of selective attention does the Stroop effect illustrate?

2

c) Explain why the following statement is incorrect:
“Cohen, Dunbar and McClelland’s (CDM) Stroop model shows that the major cause of the Stroop effect is the massive amount of practice that children get in reading words compared to the little practice they get at naming colours.”

3

d) What conclusions can validly be drawn from CDM’s Stroop model?

2

Question 3

marks

a) List two advantages and one disadvantage of distributed representations in cognitive modelling.

2

b) What aspects of Hebbian networks (or matrix models) are useful for modeling human memory?

3

c) Describe the Hebbian learning rule.

2

d) What outputs does a hebbian network give when trained on (i) orthogonal and (ii) non-orthogonal patterns? Explain your answer.

3

Question 4

marks

a) Could the perceptron convergence procedure learn the XOR function? Justify your answer.

3

b) Explain how backpropagation differs as a learning algorithm from the perceptron convergence procedure.

4

c) What is a local minima in weight space?

3

Question 5

marks

a) Self-organizing maps (SOMs) have been used to study the organization of orientation detectors in the visual cortex. Explain how a SOM network could learn to represent oriented lines from a simple retina of 3x3 units.

6

b) What effect does changing the neighbourhood size have on performance of a SOM?

2

c) Can a SOM represent patterns that it has not seen during training? Justify your answer.

2

Question 6

marks

a) What cognitive principles was Copycat designed to illustrate?

3

b) Is the parallel terraced scan in Copycat a strategy that searches breadth-first, depth-first or neither? Justify your answer.

1

A possible solution derived by Copycat to the problem abc : abd, ccbbaa : ? is ddbbaa. This solution is driven by the fact that abc is a successor group to the right and ccbbaa is a successor group to the left, allowing a correspondence to be formed between them. For the solution ddbbaa to be formed in the Workspace:

c) what initial and modified rule would allow this to occur ?

2

initial rule: replace ________________________ by _______________________

modified rule: replace ________________________ by ________________________

d) The slippages between the concepts used in the initial and modified rule are either implied by or incorporated in the correspondence that is built between the groups abc and ccbbaa. Denote what descriptions would be attached to the two groups, the slippages that occur when the two groups are mapped, and what implied slippages are required to derive the modified from the initial rule above.

1

Copycat question continued:

e) In the following diagram fill in the descriptions attached to the letters, the bonds that form between the letters, and the correspondences between letters that are consistent with your answers above.

3

a	B	c
group
successor group
right
whole

c

c

b

b

a

a

Question 7

marks

This question refers to the Buck and Buck firefly model, and the simulation from the lab class.

a) Explain how flash synchronization occurs in a population of simulated fireflies.

4

b) On average, how many cycles would it take for an initially randomized firefly to synchronize its flashing with a population of synchronized fireflies?

3

c) In a population of fireflies, what behaviour would be observed if each firefly flashed immediately its flasher unit reached a value of 1?

3

Question 8

marks

a) What is the difference between genotypic and phenotypic variability in a population?

2

b) List two mechanisms for generating genotypic variability?

2

c) Given an initial population of 100 individuals which are tested on a task and each assigned a fitness value. Explain how roulette wheel selection would be used to generate the next generation.

2

d) Define the term “evolutionarily stable strategy”, “evolutionarily stable state”.

2

e) Explain why game theorists consider that any strategy based on pure altruism would not be an evolutionarily stable strategy.

2

Question 9

marks

This question refers to Hinton and Nowlan’s simulation of the Baldwin effect.

a) In Hinton and Nowlan’s model, explain each of the following terms: chromosome, initial population, number of guesses, allele, fitness function, generation.

6

b) In Hinton and Nowlan’s oringinal simulation, the initial population was 1000, and the number of guesses was 1000. What changes in outcome would you expect if the population size was (i) increased to 100000; and (ii) decreased to 100?

2

c) In the lab class, the simulator presented the results of each run as a graph. Explain what the x- and y-axes represent; and what is plotted on the graph.

2

Question 10

marks

a) Describe the Prisoner’s Dilemma (PD) and Iterated Prisoner’s Dilemma (IPD) problems in game theory. What critical factors distinguish IPD from PD? Include a sample payoff matrix and explain the relationships between the values in each cell of the matrix.

4

b) Give real world examples that have been considered as good examples of PD and IPD (one example of each). What conclusions for agents (societies as a whole, individual humans, animals or synthetic) that trade together can be drawn from PD and IPD?

2

c) In simulations of IPD, define the terms “strategy”, “round”, “generation” and “fitness”.

2

d) Consider an initial population of 30 Nice, 30 Nasty and 30 TFT. What differences would you expect in the number of “nice” if the simulation had 100 rounds per generation compared to one that had 3 rounds per generation?

2

Question 11

marks

a) Explain the reasoning by evolutionary psychologists such as Tooby and Cosmides that brains/minds designed by evolutionary processes result in a collection of adapted modules. [nb. To answer this question, you don’t need to agree with T&C. You do need to cogently summarize their reasoning.]

4

b) What types of questions in the origins and evolution of language can be addressed by computational modeling?

3

c) Why are other questions about the origins and evolution of language so difficult to answer?

3

Question 12

marks

a) What was Allan Newell’s goal in developing SOAR?

4

b) Describe briefly how the goals, architecture and functioning of SOAR differ from the neural network models studied in IC231.

2

c) Explain what an "impasse" is in Soar and how it deals with one.

2

d) What are the benefits and limitations of Soar's chunking procedure?

2

End of Exam

Question	Marks /10
1
2
3
4
5
6
7
8
9
10
11
12
Total/120