School of Engineering, Technology and D e s i g n

Assignment Guidelines

Title of the module: AI for data intelligence.

Title of the assignment: Assignment 2
Module Team Dr. Scott Turner

Dr Amina Souag
Assignment Deadline: 15th May 2023
Expected Feedback Date:
Location of Feedback Via Turnitin on Blackboard
Assessment Type Portfolio
Where to Submit Blackboard Submission Tool [1]
What to Submit Written Report
[1] If you experience any problems with this system, then please contact the
Computing Administration Team ([email protected])

Overview
You will produce a single report incorporating three reports on the specified three examples.

Guideline only word count: Around 2000 words (this is a guide) not including tables, figures,
references, abstract in the count. Do not worry if you go over the word count there is not a
direct penalty.

Page 1 of 6
Generated from
 School of Engineering, Technology and D e s i g n

Assignment Instructions Assignment Guidelines

You are to produce a short report with the following three activities listed below and
also shown in Appendix B

 KP1 – Logic/Activity 4.1 Deterministic Logic

 Activity 9.3 Playing with a Deep Learning Network
 In Class/Activity 10.3 Combining ANN and GA

With each activity, you are expected to write a short paragraph that shows evidence of
reflection on how the technique used could be applied in other context, limitation and
strength of the technique. For higher grades you should be more reflective, for example
how does what you have done related in different applications.

In all three parts you are expected to follow the guidance in Formatting of the
Submission – not following this will affect the reporting writing part of the assignment
(see Rubric)

Page 2 of 6
Generated from
 School of Engineering, Technology and D e s i g n
Assignment Guidelines

Submission Guidelines
This is an individual assignment and must be submitted through ’Assignments’ area in the
module on Blackboard https://learn.canterbury.ac.uk.
The submission itself has ONE part:
1. Report (see assignment instructions for guidance)

The submission deadline is 14:00 on 15th May 2023

CCCU Late Submission Guidelines

In the case of late submission of work for a component of assessment (including dissertations,
extended essays, and projects) without an approved extenuating circumstance, the work will
be penalized.

The standard penalty will apply to all components of assessed work for all programmes,
unless:

EITHER:
(i) a special regulation has been approved by Academic Board;

OR:
(ii) the work is marked on a pass/fail basis and it is not possible to give a numeric mark.
The standard penalty to be applied by all programs will be 5% of the eligible marks, per day,
for up to 7 days, after which a mark of 0 will be recorded.
The reduction of 5% will be applied to the total mark that the student can receive for the com-
ponent of assessment and not to the mark that the student has earned. If, therefore, the com-
ponent of assessment is marked out of 100 and the student has scored 50%, the student will
lose 5 marks and receive 45%.

Mark Scheme/Rubric
Please see Appendix A for a detailed rubric.

Page 3 of 6
Generated from
 School of Engineering, Technology and D e s i g n
Assignment Guidelines

Formatting of Submission

Single report with three parts.

For each portfolio item/part in the report; structure it in the following way
 Introduction: Introduction to the whole task. Explain the goals and purpose of the
activities.
 Method and Results: Describe what you did and what it showed. Imagine you were giving
the example for someone else to follow. What would the need to know to repeat your
work? What results did you get? What do they show?
 Discussion and Conclusion: What did you see happen? What more general conclusions
can you draw from the work done? In some of the tasks there are questions within the
tasks to help you with this, or there is a part marked Reflection and Feedback to help with
this section.
 Bibliography (optional): For any sources you used but not needed to reference/cite.
 Reference: Any reference you need to include in the main text, for each part there is
expected to be at least one reference. A separate reference list per portfolio item/part.

Resources You May Find Useful

The majority of the resources you may need are accessible via the module on Blackboard at
http://learn.canterbury.ac.uk.

Page 4 of 6
Generated from
 School of Engineering, Technology and D e s i g n

Learning Outcomes Assessed (Fully or

Assignment Guidelines

Partially)

LO Description
1. Demonstrate comprehensive understanding of
the principles and techniques applicable to the
study of artificial intelligence and machine
learning and apply this knowledge to authentic
problems
2. Critically analyse the contemporary application of
artificial intelligence methods and algorithms
taking into account ethical, legal and social
considerations
3. Design artificial intelligence systems for data
science using data that may be complex,
contradictory or incomplete
4. Critically apply the Conceive, Design, Implement
and Operate Methodology to a data intelligence
problem

Page 5 of 6
Generated from
 School of Engineering, Technology and D e s i g n

APPENDIX A: MARK RUBRIC Assignment Guidelines

KP1 – Logic/Activity 4.1 Deterministic Logic (total 30%)

0-14% Below threshold including poorly written reports, including not
following the Formatting of Submissions requirements, or incomplete
tasks
15-18%– complete all the tasks from Task 1 to Task 4.2
19-22% Complete all the tasks including Task 4.3
23-30% – All of the task completed, and the Prolog code has been
developed further including finding ways to make expansion of
categories easier. including
 Change the facts structure to make it easier
 Add rules for Aunt, Uncle and Cousin

Activity 9.3 Playing with a Deep Learning Network (total 30%)

0-14% Below threshold including poorly written reports, including not
following the Formatting of Submissions requirements, or incomplete
tasks
15-18% – completed all parts including Reflection and Feedback.
19-22% As in Pass criteria but with referencing to support the
arguments throughout.
23-30% – As in merit put also including a second example and
comparing the two examples with appropriate referencing and
documentation.

In Class/Activity 10.3 Combining ANN and GA (total 40%)

0-19% Below threshold including poorly written reports, including not
following the Formatting of Submissions requirements, or incomplete
tasks.
20-23% – complete Task 1 and Task 2 part A.
24-27%- complete Task1, Task 2 parts A and B
28-40% all of the above plus Task 2 C with further discussion on based
on published literature on controlling a robot using neural networks
and genetic algorithm.

Page 6 of 6
Generated from
 School of Engineering, Technology and D e s i g n
Assignment Guidelines

APPENDIX B: Copies of the Tasks

KP 1 - Logic/Activity 4.1 Deterministic Logic

Task 1:

Using a Prolog tool (Swish will do) Enter the program below.

mother(sarah, jane).%sarah is the mother of jane

father(tom, jane). mother(victoria, sarah).

father(henry, sarah). mother(victoria, ben).

mother(jane,thomas).

father(thomas,time).

parent(X,Y):- father(X,Y).

parent(X,Y):- mother(X,Y).

descendant(X,Y):- parent(Y,X).

descendant(X,Y):- parent(Y,Z), descendant(X,Z).

brother(X,Y):- mother(Z,X),mother(Z,Y).

Task 2:

Find if jane is a descendant of Victoria. How do you ask the question?

Task 3: Page 7 of 6
Generated from
Make sarah the mother of someone, and then test who that person is a descendant of. You
School of Engineering, Technology and D e s i g n
might need to write the descendant rules to cope with this.
Assignment Guidelines

Task 4.

There are a number of problems with the brother rule at the moment

Task 4.1

Identify the issues.

Task 4.2

Identify possible ways to solve this.

Task 4.3

Challenge: Make the changes to make the program to see if you can address the problems.

Page 8 of 6
Generated from
 School of Engineering, Technology and D e s i g n
Assignment Guidelines

Week ANN 4 Bring it together/In class activities/Activity 9.3

Playing with Deep Learning

Purpose Online activity to look at Deep Learning using TensorFlow

Task Questions to be answered by the end of the session (i and ii)

1. Can you see potential applications of DeepLearning and Image Processing? If so where?
2. What do you understand about the term overfitting?
3. Advanced: What can we do to potentially improve the prediction on the test data, minimise overfitting with this
example?
4. Click on

https://colab.research.google.com/github/margaretmz/deep-learning/blob/master/
fashion_mnist_keras.ipynb#scrollTo=QwNmlfIC0YxM

2. run on CoLab version

3. We can get Google to do a lot of the work for us here. Google allows you access to their specialist Tensor servers
temporarily to handle so the heavy work.
 School of Engineering, Technology and D e s i g n
4. To follow the worksheet look for the [ ] symbols that should turn into play buttons when your cursor is over them. Click
on the play button at each stage to carry out the action. You can ignoreAssignment Guidelines
the warnings that occur.

5. Now play with the parameters a bit? What do you find?

Can you now answer these questions

Reflection
and 1. Can you see potential applications of DeepLearning and Image Processing? If so where?
feedback 2. What do you understand about the term overfitting?
3. Advanced: What can we do to potentially improve the prediction on the test data, minimise overfitting with this
example?

In Class/Activity 10.3 Combining ANN and GA

The code below uses a Genetic Algorithm to train a single neuron.

Task 1

Run the code for different expected outputs? Did it train?

https://trinket.io/python3/274544f496
 School of Engineering, Technology and D e s i g n
Assignment Guidelines

Task 2

A. How could you use an evolutionary approach in Week ANN 1/In Class/Activity 6.3 to find the weights?

B. Could you adapt the code so that it finds both sets of weights just by running the code once?

C. How can you what we learnt through the module to find a set of weights to control a robot?