(index.

html)

CITS5017 Deep Learning - Face

to face / Blended learning

Unit Information
Unit Code CITS5017
Title Deep Learning - Face to face / Blended learning
Level 5
Faculty Physics, Mathematics and Computing
School Physics, Mathematics and Computing
Unit Coordinator Du Huynh
Credit Points 6 points

Teaching contingency plan (In response to Covid19 restrictions)

In the event of more Covid-19 restrictions being imposed such as another lockdown, the unit will continue to run as follows:

Lectures and Practical sessions will be delivered via MS Teams. A link(s) will be circulated to all students who will be able to watch
and participate in lectures remotely.

The material associated with the lab sessions will be delivered via LMS.

Students will be provided with additional opportunities to attend consultation sessions.

Academic Information
Content This unit focuses on advanced deep learning concepts and their application. Assuming basic machine and
deep learning knowledge, and experience in related programming frameworks, it will delve deeper into the
building blocks of modern deep learning systems and their specialised applications in processing structured
and unstructured data. The unit covers in detail advanced topics in Generative Adversarial Networks (GANs),
variational autoencoders, deep reinforcement learning, policy gradient methods, and Adversarial Machine
Learning. These topics are covered with hands-on experience throughout the unit.
Unit Learning # Outcome How outcome will be assessed
Outcomes
1 apply deep neural networks to structured and unstructured final examination; laboratory
data assessments
2 apply generative adversarial networks to learn data laboratory assessments
distribution
3 analyse generative models for cross-domain data generation final examination; laboratory
assessments
4 demonstrate understanding of reinforcement learning final examination
5 compute adversarial patterns for deep models laboratory assessments
6 contrast robust deep models with non-robust models final examination

Indicative # Assessment Indicative weighting Failed component

Assessment
1 laboratory assessments 50% 0
2 final examination 50% 0

The indicative assessment structure provides an overview of how this unit is typically assessed. In order to fit
teaching requirements, this may change slightly from year to year. Restrictions to face-to-face learning
associated with COVID-19 may result in variation to the Indicative Assessment structure. The Required
Assessments table below shows how the unit is being assessed in this specific teaching period.
Required Unit grades are determined from final unit marks in accordance with the UWA Policy (http://
Assessments www.governance.uwa.edu.au/procedures/policies/policies-and-procedures?method=document&id=up15/5)
on Assessment

Contact Details
Unit Coordinator Contact Information
Name Du Huynh
Email [email protected]
Phone number +61 8 6488 2878
Room number or CSSE Room 1.22
location
Consultation
information

Unit Details
CITS5017 - Deep Learning - SEM-2-2023 - 2023 - Face to face / Blended learning

Unit Information
With an explosion of data being collected from various sensing technologies, deep learning provides core underlying theory and
techniques to data science and artificial intelligence. This unit focuses on advanced deep learning concepts and their
application. Assuming basic machine and deep learning knowledge, and experience in related programming frameworks, it will
delve deeper into the building blocks of modern deep learning systems and their specialised applications in processing
structured and unstructured data. The unit covers in detail advanced topics in Generative Adversarial Networks (GANs),
variational autoencoders, deep reinforcement learning, policy gradient methods, and Adversarial Machine Learning. These topics
are covered with hands-on experience throughout the unit.
Lecture Capture System
In accordance with UWA’s policy on lecture capture all lectures* in enabled venues will be recorded and made available to stream
and download by enrolled students via the Lecture Capture System (LCS) for the duration of the unit (including Supplementary
and Deferred Examinations period).

* Please note that in some limited circumstances lectures may not be recorded due to: ethical and/or security related issues; the
use of commercially sensitive material; the use of third-party copyright teaching material; or culturally sensitive content.
Students will be advised, in advance, when this applies.

Learning And Teaching Strategy

Apart from the face-to-face classes, students are encouraged to participate in the discussion forum set up within LMS.
Announcements will be made from time to time to the forum also.

Unit Structure
The unit comprises 12 weeks of 2-hour lectures and 11 week of lab classes (starting in week 2). All the lecture material, labsheets,
and project specifications will be available online through LMS.

Required Assessments
Assessment Assessment Weight Assessment period/ Submission Unit Learning Waiver
# Task date/s Procedure Outcome
1
Mid-semester 15% 15/7 12:00 AM Written test 1, 2 No
test

2
Programming 35% 15/7 12:00 AM Submission 1. 2. 3. 4. 5. 6 No
projects

3
Final 50% 15/7 12:00 AM Written test 1, 3, 4, 5, 6 No
examination

Where a document waiver is indicated against an assessment item no medical or other evidentiary requirement is necessary if
the request is for 48 hours or less past the assessment due date.

In accordance with the University Policy on Academic Integrity, the use of Artificial Intelligence (AI) is
permitted as an educational/study tool. It may only be used in any assessment within a unit where approval
has been granted by the unit coordinator. Improper use of AI-generated material, as set out in the Academic
Integrity policy, in assessments may lead to the occurrence of academic misconduct.

Unit Schedule
Week # Date Starting Topic Preparation Assessment Notes
1
24/07/2023 1. Introduction to Artificial Reading: No labs
Neural Networks Chapter 10
2
31/07/2023 Topic 1 cont'd Reading: work on Labsheet 1
Chapter 10

3
07/08/2023 2. Training Deep Neural Reading: work on Labsheet 2
Networks Chapter 11

4
14/08/2023 3. Deep Computer Vision Reading: work on Labsheet 3.
with CNNs Chapter 14 Programming Project 1
specification released

5
21/08/2023 Topic 3 cont'd Reading: Mid-semester test work on Labsheet 3 & Project 1
Chapter 14

6
28/08/2023 4. Processing Sequences Reading: work on Project 1
Chapter 15

7
11/09/2023 Topic 4 cont'd Reading: Programming work on Project 1
Chapter 15 Project 1 due

8
18/09/2023 5. Representation Learning Reading: work on Labsheet 4
using Autoencoders Chapter 17

9
25/09/2023 6. Convolutional, Reading: work on Labsheet 4.
Denoising, Variational AEs; Chapter 17 Programming Project 2
Attention specification released

10
02/10/2023 7. Generative Adversarial Reading: work on Project 2
Networks Chapter 17

11
09/10/2023 8. Transformers and Vision work on Project 2
Transformers

12
16/10/2023 9. Deep Reinforcement Reading: Programming work on Project 2
Learning Chapter 18 Project 2 due

Participation
Students are expected to attend the 2-hour lectures and 1 lab session per week.

Referencing Requirements
In the programming exercises in the labsheets and in the projects, students should not use codes from non-standard libraries. If a
large portion of codes is copied from the textbook, a proper reference to the textbook should be supplied. Students should do
the coding themselves and explanation/comments should be given alongside their codes.

Essential Textbooks
There are no essential textbooks set for this unit. There may be essential and recommended readings set for this unit available in
Unit Readings in the LMS.
Other
So�ware requirements:

- Python version 3.10 onward (recommended: 3.11).

- TensorFlow - this is a library for training and testing large and deep neural networks. It will be used extensively in the unit.
TensowFlow version 2.8 or above is required (recommended: 2.11).

- Keras - this is a subpackage under TensorFlow 2.x. Installing TensorFlow 2.x will automatically install this subpackage. It
provides support for some advanced TensorFlow features.

- Scikit-Learn - this is a widely used machine learning library for Python and some of the classes/functions will be used in the unit.
Other libraries that are needed for the unit are: NumPy, pandas, Matplotlib, and gym.

- Jupyter-Lab (preferred) or Notebook (https://jupyter.org/ (https://jupyter.org/)).

Late submission of Assignments

A penalty of 5 per cent of the total mark allocated for the assessment item is deducted per day for the first 7 days (including
weekends and public holidays) a�er which the assignment is not accepted. Each 24-hour block is recorded from the time the
assignment is due.

For example, if an assignment is late by three days and was given 45 out of a possible mark of 50, you would receive a mark of
37.5 out of 50 (a mark of 2.5 is deducted per day).

If there are a number of tasks within an assessment item, the late penalty may be applied to the whole assessment item a�er all
tasks have been completed.

If an assignment is graded Pass/Fail, failure to submit the assignment in time may result in the student not being permitted to
take the final exam at the end of the semester or a teaching period and being unable to progress to the next level of the course or
to graduate in the case where it is the final unit of the course;

Penalty for exceeding word limit

Where a submitted assignment exceeds the word limit, a penalty of 1 per cent of the total mark allocated for the assessment task
applies for each 1 per cent in excess of the word limit, or the marking ceases once the word limit is reached.

Academic conduct
Academic Integrity is defined in the University Policy on Academic Conduct (https://www.uwa.edu.au/policy?
policy=UP07%252F21) as "honest, ethical and responsible scholarship". UWA expects the highest degree of academic conduct
from all students.

Penalties for breach of academic conduct vary according to seriousness of the case, and may include the requirement to do
further work or repeat work; deduction of marks; the award of zero marks for the assessment; failure of one or more units;
suspension from a course of study; exclusion from the University; non-conferral of a degree, diploma or other award to which the
student would otherwise have been entitled.

Students should familiarise themselves with the information and resources available regarding academic conduct and ethical
scholarship (https://www.uwa.edu.au/library/Find-resources/ACE) at UWA.

Special Considerations
If your study has been adversely affected by illness or other significant circumstances outside of your control, you can apply for
special consideration. Significant circumstances may include but are not limited to:

• serious illness or death of a member of the student's immediate family or household or of a close friend
• serious injury
 • being a victim of a crime
• breakdown of relationship
• sudden loss of income or employment
• serious disruption to domestic arrangements

The full policy and information on applying for special consideration is available at the UWA Policy Library (https://
www.uwa.edu.au/policy?policy=UP07%252F21)

Review and Appeal of Academic Decisions Relating to Students

The full regulations governing appeal, and the process procedures are available at the UWA Policy Library (https://
www.uwa.edu.au/policy?policy=UP07%252F21)

Compulsory online modules

UWA has a wealth of resources available to support your learning. These online modules are compulsory for many students:

• Academic Conduct Essentials (ACE) introduces you to the basic issues of ethical scholarship and the expectations of correct
academic conduct at university. All students are required to complete ACE in the first semester of their course.
• Communication and Research Skills (CARS) aims to assist you to develop communication and research skills in an academic
context and is compulsory for all new coursework students.
• Indigenous Study Studies Essentials (INDG1000) introduces you to the shared learning space that UWA embodies. This
learning space includes both western and Indigenous knowledge systems. INDG1000 is compulsory for all students
completing a first year undergraduate course.

More information about these units can be found at the UWA Handbook (https://handbooks.uwa.edu.au)

Useful Contacts
Student Guild
The Guild represents all students enrolled at UWA and can help you in a number of ways. For financial, academic or welfare
support please email Guild Student Assist at [email protected] (mailto:[email protected]). For all other queries
email [email protected] (mailto:[email protected]) or visit the UWA Guild website (http://
www.uwastudentguild.com/).

University Library
Visit one of UWA's six libraries or the University Library online (https://www.uwa.edu.au/library/home) to access a wide range of
services, study spaces, information and help, including:

• Access to textbooks and readings for your units

• Individual and group study spaces
• Wireless internet and your student account
• Research, referencing and copyright support
• Printing and scanning services

Student Life
There is a wealth of material provided by UWA Student Life that can help you settle into university life and help with many other
issues you may encounter as a university student. Visit the UWA Students website (https://www.uwa.edu.au/students) to see
everything that is on offer.

Study Smarter
UWA’s Academic Skills Centre STUDYSmarter provides free academic advice, support and resources for all undergraduate and
postgraduate students at UWA. We can help you to develop the writing, research, English language, maths and stats skills you
need to excel in your university studies. Find out about drop-ins, consultations, workshops, a�er-hours advice and online
resources on the STUDYSmarter website (http://www.studysmarter.uwa.edu.au).

Student Administration
Student Admin deal with enrolments, fees and other vital services for students. They can be contacted via their website (http://
www.student.uwa.edu.au/course)