Artificial Intelligence Methods (G5BAIM)

Course Feedback - Academic Year 1999/00

This course is run at Nottingham University. It is a second/third year option. This page summarises the feedback from the students that attended the course. All comments are listed. A response from the lecturer is also given and at the end there are some notes with regards to the examination.

The lecturer is Graham Kendall. Feel free to email me or visit my home page.

The home page for the course (including a full set of course handouts) can be found here.

Comments from Course Feedback Forms

Total Questionnaires Returned = 63

Number of students who sat exam = 119


Lecturers Response

Overall I am plesaed with this course feedback. Next year, this course may change as it may be split into two modules. If this is done I will take the opportunity to update my powerpoint slides and try to develop the areas where I have been criticised (e.g. more interaction, more examples, more hands on).

Examination Report


118 people took the exam for this course. 10 people failed to hand in the coursework.
The figures for the number of candidates who attempted the questions are based on those questions that contributed towards their marks. That is, if the student attempted five questions, only four are counted in this report.
The overall average for this course was slightly above what was expected.


The coursework involved writing various search algorithms to solve the 8-puzzle problem.
From the feedback I received the coursework was well received and most people found it challenging but an enjoyable assignment.
The overall average for the coursework was below the expected average but, if the 10 people who did not hand in their assignment are excluded from the calculation then the average is very close to the expected value.

Question 1

This question was attempted by 110 candidates. The question was similar to previous exam questions (A* algorithm) and its use had been explained at various times throughout the lectures.
The question achieved an above average mark, although many students found difficulty in producing the correct search tree and stating the correct route.

Question 2

This question was attempted by 110 candidates. It mainly involved genetic algorithms as well as considering other population based approaches. The students were divided between those that could apply the algorithms correctly and those that really had no idea.
The second part of the question was not so well answered with not many students giving their own ideas, rather they quoted material from the lectures.
The question achieved an above average mark.

Question 3

This question was attempted by 67 candidates. The question asked the students to consider two neighbourhood search algorithms (hill climbing and simulated annealing).
In general, the students were able to describe the two search algorithms but found difficulty in describing the relationship between the two, without giving two separate algorithms.
The question also asked the student to show an example of how likely simulated annealing is to accept a worse move. This was, in general, poorly answered.
This question achieved a below average mark.

Question 4

This question was attempted by 95 candidates. It asked the students to describe the Turing Test and The Chinese Room Experiment and then comment on the issues raised by these two experiments.
The students gave good, reasoned descriptions and arguments for these topics and the question achieved an above average mark.

Question 5

This question was attempted by 10 candidates. This question asked the students to give their opinion as to what constitutes artificial life and the problems it would pose. In addition, the students were asked to describe their ideas as to how a computer program could learn how to play a card game (blackjack).
The low number of students who attempted this question, I assume, shows they thought it was a difficult question.
This is probably the case as it achieved the lowest average mark of all the questions.

Question 6

This question was attempted by 83 candidates. The question asked the students to state what functions could be learnt by a perceptron and then comment on how two types of perceptrons (but identical in operation) could be built.
This question is very similar to questions from previous exams. This is probably one reason why it achieved the highest average mark and was generally well answered by those attempting it.