Assessment Brief: ICT5357 Problem-Solving and Decision-Making with Machine Learning

Trimester 3, 2024

Assessment Overview
Type Weight Length Due ULOs
Assessment Task
ing Assessed
Assessment 1: Supervised Machine Individual 20% 1000 (500 Week 4 ULO1
Learning Project words for ULO2
Implement a machine learning report + ULO4
model to perform a simple 500 words
classification or regression task. for
Write a report outlining steps model)
taken.
Assessment 2: Laboratory Individual 30% 2000 Weeks ULO1
Practicum words 6, 8 ULO2
Regular invigilated practical tests, (1200 ULO3
Invigilated words
implementing some machine ULO4
learning model or machine learning for
analysis to solve a problem. report +
800
words
for
model)
Assessment 3: Unsupervised Group 25% 3000 Week 7 ULO1
Machine Learning Project words ULO3
Perform clustering and (2000 ULO4
dimensionality reduction on a words
complex dataset and implement a for
machine learning model to perform report +
1000
a classification/regression task on
words
the reduced dataset. Write a report
for
outlining steps taken. model)
Assessment 4: Computer Vision Individual 25% 2000 Week ULO1
Case Study words 12 ULO2
Apply strategic problem-solving in a (1500 ULO4
domain involving visual data words for
analysis. Write a report outlining report +
steps taken. 500 words
for
model)

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 1
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
equiv. – equivalent word count based on the Assessment Load Equivalence Guide. It means this assessment is
equivalent to the normally expected time requirement for a written submission containing the specified
number of words.

Note for all assessments tasks:

● Students can generate/modify/create text generated by AI. They are then asked to
modify the text according to the brief of the assignment.
● During the preparation and writing of an assignment, students use AI tools, but may
not include any AI-generated material in their final report.
● AI tools are used by students in researching topics and preparing assignments, but
all AI-generated content must be acknowledged in the final report as follows:

Format
I acknowledge the use of [insert the name of AI system and link] to [describe how it
was used]. The prompts used were entered on [enter the date in ddmmyyy:] [list the
prompts that were used]
Example
Tools
I acknowledge the use of ChatGPT https://chat.openai.com to create content to plan
and brainstorm ideas for my assessment. The prompts used were entered on 18
March, 2023:
● What are some key challenges in running an online business?

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 2
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
Assessment 1: Supervised Machine Learning Project

Due date: Week 4

Group/individual: Individual
Word count/Time provided: 1000 (500 words for report + 500 words for model)
Weighting: 20%
Unit Learning Outcomes: ULO-1, ULO-2, ULO-4

Assessment 1 Detail
You will be provided with a number of simple raw datasets that could be used for
classification or regression. You will choose one dataset and submit a report that outlines:

● A problem that could be addressed with the dataset

● A machine learning algorithm that could be used to train AI to answer this question
● Any preprocessing of data and other preparation that you would do for this training

You will also submit one or more models trained with the data. You are required to submit:

1. A document (1000 words) that includes:

○ What kind of problem could be solved and/or what kind of decisions could be
made, if an appropriate machine learning model were trained on this dataset
○ What such a model would do, either as a classification task or a regression
task
○ An explanation of how the dataset relates to the classification or regression
task
○ Any specific learning algorithms that might be particularly suitable or
unsuitable, given the dataset, and why
○ What preprocessing would you perform on the dataset, prior to machine
learning
○ An evaluation of the performance of the model
2. The project used to develop your model(s).
3. The deployed model(s).

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 3
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
 Assessment 1 Marking Criteria and Rubric
The assessment will be marked out of 100 and will be weighted 20% of the total unit mark.

Marking Not satisfactory Satisfactory Good Very Good Excellent

Criteria
(0-49%) of the criterion (50-64%) of the (65-74%) of the (75-84%) of the (85-100%) of the
mark criterion mark criterion mark criterion mark criterion mark

Problem Unable to identify and Able to identify and Able to identify Able to identify and Able to identify and
Solving connect problems to connect some and connect many connect problems connect problems to
(20 marks) machine learning problems to problems to to machine learning machine learning
models and training machine learning machine learning models and training models and training
data. models and training models and data. data, and provide
data. training data. clear explanations for
these connections.

Data Explanation of the Dataset is Explanation

Explanation is Explanation is
(20 marks) dataset is missing or described with provides a clear
thorough and exceptional, offering
contains significant basic details but overview of the
demonstrates a detailed insights into
inaccuracies. No lacks clarity or dataset, including
strong grasp of features, target
modifications or pre- depth. Basic features, target
dataset variable, and
processing are modifications or variable, and
characteristics and dataset's role in the
performed, or pre-processing relevance to the
relevance. task. Modifications or
explanations are steps are task. Modifications
Modifications or pre- processing are
unclear. mentioned, but or pre-processing
pre-processing are exceptional, showing
their rationale is are adequately
well-reasoned and deep understanding
unclear. explained and
effectively of their impact on the
justified.
contribute to model task.
development.
Machine Inappropriate Basic algorithms Appropriate Algorithms are well- Algorithm selection
Learning (20 algorithms or are chosen, but algorithms are suited to the task, and hyperparameter
marks) hyperparameters are their suitability and selected, and tuning are
chosen, or hyperparameters relevance to the hyperparameters exceptional, showing
explanations are are inadequately task is justified are carefully tuned deep understanding
missing. explained. with reasonable with clear of model choices.
hyperparameters. reasoning.
Models (20 No model results Basic model results Provides some Model outcomes Model performance
marks) provided or results do are mentioned but meaningful are explained with is explained clearly,
not address the lack context or insights, and insight, their interpretation shows
problem. interpretation. results are implications are deep understanding.
described clearly. well interpreted.

Clarity and Answers are unclear, Answers are Answers are Answers are clear Answers are clear
Presentation disorganised, and somewhat clear, mostly clear and and organised. and organised, easy
(20 marks) difficult to understand. but lack organised. to understand, and
organisation. engaging or
insightful.

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 4
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
 Assessment 2: Laboratory Practicum
Due date: Weeks 6, 8
Group/individual: Individual
2000 words (1200 words for report + 800 words for
Word count/Time provided:
model), 1000 words each
Weighting: 2x 15% = 30%
Unit Learning Outcomes: ULO-1, ULO-2, ULO-3, ULO-4

Assessment 2 Detail
During one hour of the weekly seminar in Weeks 6 and 8, you will be provided with some
problem or decision that can be addressed with an appropriate machine learning approach.
You will be required to provide written answers, and in some weeks you will also be
required to provide one or more machine learning models to address the problem.

These assessments are held during two relevant weekly seminar and they are invigilated.

Assessment 2 Marking Criteria and Rubric

Each assessment will be marked out of 100 and will be weighted 15% of the total unit mark,
totalling a combined weight of 2 x 15% = 30%.

Marking Not satisfactory (0- Satisfactory Good Very Good Excellent

Criteria 49%) of the criterion (50-64%) of the (65-74%) of the (75-84%) of the (85-100%) of the
mark criterion mark criterion mark criterion mark criterion mark
Problem Unable to identify Able to identify Able to identify Able to identify and Able to identify and
Solving (25 and connect and connect and connect connect problems connect problems to
marks) problems to machine some problems many problems to to machine machine learning
learning models and to machine machine learning learning models models and training
training data. learning models models and and training data. data, and provide
and training training data. clear explanations for
data. these connections.

Data and No relevant data Some relevant All/most relevant All relevant data All relevant data
Machine preprocessing, data data data preprocessing, preprocessing, data
Learning (25 visualisations, preprocessing, preprocessing, data visualisations, visualisations,
marks) clustering, data data clustering, clustering,
dimensionality visualisations, visualisations, dimensionality dimensionality
reduction or time clustering, clustering, reduction or time reduction or time
series techniques dimensionality dimensionality series techniques series techniques are
are used. reduction or time reduction or time are used and used and explained
series series techniques explained well. with excellent insight.
techniques, but are used and
submission explained.
contains errors

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 5
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
 or omissions in
explanations.

Orange / Machine learning Machine learning Machine learning Machine learning Machine learning
AutoML project is project achieves project is project is well project is
techniques disorganised and goals but with somewhat organised and exceptionally
(25 marks) difficult to follow. little organised but easy to follow. organised with a
organisation. contains one or logical workflow.
more errors in the
workflow.

Clarity and Answers are Answers are Answers are Answers are clear Answers are clear
Presentation unclear, somewhat clear, mostly clear and and organised. and organised, easy
(25 marks) disorganised, and but lack organised. to understand, and
difficult to organisation. engaging or insightful.
understand.

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 6
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
Assessment 3: Unsupervised Machine Learning Project
Due date: Week 7
Group/individual: Group
3000 words (2000 words for report + 1000 words for
Word count/Time provided:
model)
Weighting: 25%
Unit Learning Outcomes: ULO1, ULO3, and ULO4

Assessment 3 Detail
You will be provided with a number of complex high-dimensional raw data sets. You will
choose one of these datasets and submit a report that outlines:

● A problem that could be addressed with the dataset

● How methods could be used to reduce the dimensionality of the dataset and uncover
structure in the dataset
● What kind of classification or regression could be performed with the data once the
complexity in the dataset has been reduced

You will also submit several models that have been trained with the data, including:

● A model trained on the raw dataset

● A model trained on the dataset after applying some dimensionality reduction
techniques
● A model trained on the dataset after manually choosing a selection of input features
in the training dataset

You are required to submit:

1. A document (3000 words) that includes:

○ What kind of problem could be solved and/or what kind of decisions could be made, if
an appropriate machine learning model were trained on this dataset
○ What such a model would do, either as a classification task or a regression task
○ An explanation of how the dataset relates to the classification or regression task
○ Insights gained from the raw data, including from visualisations of the data
○ An explanation of any dimensionality reduction techniques that could be used to
reduce the complexity of the raw data
○ Any specific learning algorithms that might be particularly suitable or unsuitable, given
the dataset, and why
○ What preprocessing would you perform on the dataset, prior to machine learning
○ An evaluation of the performance of the models
2. The project used to develop your models.
3. The deployed models.

Assessments 3 Marking Criteria and Rubric

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 7
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
 The assessment will be marked out of 100 and will be weighted 25% of the total unit mark.

Marking Not satisfactory (0- Satisfactory Good Very Good Excellent

Criteria 49%) of the criterion (50-64%) of the (65-74%) of the (75-84%) of the (85-100%) of the
mark criterion mark criterion mark criterion mark criterion mark
Problem Unable to identify Able to identify Able to identify Able to identify Able to identify
Solving (10 and connect and connect some and connect many and connect and connect
marks) problems to machine problems to problems to problems to problems to
learning models and machine learning machine learning machine learning machine learning
training data. models and models and models and models and
training data. training data. training data. training data, and
provide clear
explanations for
these
connections.

Data (10 Explanation of the Dataset is Explanation Explanation is Explanation is

marks) dataset is missing or described with provides a clear thorough and exceptional,
contains significant basic details but overview of the demonstrates a offering detailed
inaccuracies. No lacks clarity or dataset, including strong grasp of insights into
modifications or pre- depth. Basic features, target dataset features, target
processing are modifications or variable, and characteristics variable, and
performed, or pre-processing relevance to the and relevance. dataset's role in
explanations are steps are task. Modifications Modifications or the task.
unclear. mentioned, but or pre-processing pre-processing Modifications or
their rationale is are adequately are well-reasoned pre- processing
unclear. explained and and effectively are exceptional,
justified. contribute to showing deep
model understanding of
development. their impact on the
task.
Data Creates little or no Creates Creates relevant Creates relevant Clearly explains
Visualisation visualisations of visualisations of visualisations of and helpful how visualisations
(15 marks) data, and/or offers data, but offers on data. Presents visualisations of of data connect to
no explanation or limited explanation some insights data. Presents insights into the
understanding of of how these based on these good insights problem space.
their significance. visualisations visualisations. based on these
connect to insights visualisations.
into the problem.

Dimensionality No dimensionality Some Explains some Explains Clearly explains

Reduction (15 reduction techniques dimensionality dimensionality dimensionality dimensionality
marks) are used and/or no reduction reduction reduction reduction
rationale given for techniques are techniques used techniques used techniques used
these. used. Little and offers some and the rationale and the rationale
rationale given for rationale for these. for these. Shows for these. Shows
these. Shows some good insight into excellent insight
insight into dimensionality into dimensionality
dimensionality reduction. reduction.
reduction.

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 8
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
Machine Inappropriate Basic algorithms Appropriate Algorithms are
Algorithm
Learning (10 algorithms or are chosen, but algorithms are well-suited to the
selection and
marks) hyperparameters are their suitability and selected, task, and hyperparameter
chosen, or hyperparameters relevance to the hyperparameters
tuning are
explanations are are inadequately task is justified are carefully tuned
exceptional,
missing. explained. with reasonable with clear showing deep
hyperparameters. reasoning. understanding of
model choices.
Models (10 No model results Basic model Provides some Model outcomes Model
marks) provided or results results are meaningful are explained with performance is
do not address the mentioned but insights, and insight, their explained clearly,
problem. lack context or results are implications are interpretation
interpretation. described clearly. well interpreted. shows deep
understanding.

Orange / Machine learning Machine learning Machine learning Machine learning Machine learning
AutoML project is project achieves project is project is well project is
techniques (15 disorganised and goals but with little somewhat organised and exceptionally
marks) difficult to follow. organisation. organised but easy to follow. organised with a
contains one or logical workflow.
more errors in the
workflow.
Clarity and Answers are Answers are Answers are Answers are clear Answers are clear
Presentation unclear, somewhat clear, mostly clear and and organised. and organised,
(15 marks) disorganised, and but lack organised. easy to
difficult to organisation. understand, and
understand. engaging or
insightful.

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 9
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
 Assessment 4: Computer Vision Case Study
Due date: Week 12
Group/individual: Individual
2000 words (1500 words for report + 500 words for
Word count/Time provided:
model)
Weighting: 25%
Unit Learning Outcomes: ULO-1, ULO-2, ULO-3, ULO-4

Assessment 4 Detail
You will be provided with a number of image libraries. You will choose one image library and
submit a report that outlines:

● A problem that could be solved with the image dataset

● How AutoML tools may help to analyse the imagery
● A strategy to solve the problem with the image library and an appropriate machine
learning model

You will also submit one or more models that you have trained. You are required to submit:

1. A document (2000 words) that includes:

○ A realistic fictional scenario in which the image library could help solve a problem
○ An outline of a strategic approach to solving the problem with the use of the image
library and a machine learning model trained from it
○ Explorations and insights gained from the image library
○ Image preprocessing used on the data, including any advanced pre-processing that
might be conducted with access to advanced AutoML tools
○ An evaluation of the performance of any models
2. The project used to develop your model(s).
3. The deployed model(s).

Assessments 4 Marking Criteria and Rubric

The assessment will be marked out of 100 and will be weighted 25% of the total unit mark.
Marking Not satisfactory (0- Satisfactory Good Very Good Excellent
Criteria 49%) of the criterion (50-64%) of the (65-74%) of the (75-84%) of the (85-100%) of the
mark criterion mark criterion mark criterion mark criterion mark
Strategic No relevant scenario Some Adequate Good explanation Excellent
Problem or problem domain identification of explanation of the of the scenario explanation of the
Solving (20 is identified. No the scenario and scenario and and problem scenario and
marks) relevant machine problem domain. problem domain. domain. An problem domain,
learning objective is An objective of the An objective of the appropriate and the objective
identified or machine learning machine learning machine learning of the machine
explained. project is project is project objective is learning project.
identified, but not identified, but may identified and
explained. not be well explained.

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 10
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
 explained.

Data (20 Does not explain Able to somewhat Able to somewhat Able to explain the Able to clearly
marks) how the image explain the explain the connection explain the
library connects to connection connection between the connection
the machine learning between the between the image library and between the
objective. image library and image library and the objective of image library and
the objective of the objective of the the machine the objective of
the machine machine learning learning project. the machine
learning project, project. Shows Shows some learning project.
but no mention of little knowledge of knowledge of how Shows great
how further data how further data further data insight into how
should be should be should be further data
collected or collected or collected or should be
augmented. augmented. augmented. collected or
augmented.
Image No image pre- Basic image pre- Image pre- Image pre- Image pre-
Processing (20 processing is processing is processing is processing is processing is
marks) performed. performed, but mostly appropriate performed and completely
with no to the task, but adequately appropriate and
explanation. with little explained. More clearly explained,
explanation. More advanced image showing deep
advanced image pre-processing is understanding of
pre-processing is adequately its role in image
mentioned but not explained in analysis. More
explained. relation to AutoML advanced image
tools. pre-processing is
clearly explained
in relation to
AutoML tools.
Models (10 No model results Basic model Provides some Model outcomes Model
marks) provided or results results are meaningful are explained with performance is
do not address the mentioned but insights, and insight, their explained clearly,
problem. lack context or results are implications are interpretation
interpretation. described clearly. well interpreted. shows deep
understanding.
Orange / Machine learning Machine learning Machine learning Machine learning Machine learning
AutoML project is project achieves project is project is well project is
techniques (15 disorganised and goals but with little somewhat organised and exceptionally
marks) difficult to follow. organisation. organised but easy to follow. organised with a
contains one or logical workflow.
more errors in the
workflow.
Clarity and Answers are Answers are Answers are Answers are clear Answers are clear
Presentation unclear, somewhat clear, mostly clear and and organised. and organised,
(15 marks) disorganised, and but lack organised. easy to
difficult to organisation. understand, and
understand. engaging or
insightful.

Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 11
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning
Asia Pacific International College Pty Ltd. Trading as Asia Pacific International College Page 12
PRV12007; CRICOS 03048D ICT5357 Problem-Solving and Decision-Making with Machine Learning