Note: This unit version is currently under review and is subject to change!

COMP5328: Advanced Machine Learning (2019 - Semester 2)

Download UoS Outline

Unit: COMP5328: Advanced Machine Learning (6 CP)
Mode: Normal-Evening
On Offer: Yes
Level: Postgraduate
Faculty/School: School of Computer Science
Unit Coordinator/s: Liu, Tongliang
Session options: Semester 2
Versions for this Unit:
Site(s) for this Unit:
Campus: Camperdown/Darlington
Pre-Requisites: None.
Brief Handbook Description: Machine learning models explain and generalise data. This course introduces some fundamental machine learning concepts, learning problems and algorithms to provide understanding and simple answers to many questions arising from data explanation and generalisation. For example, why do different machine learning models work? How to further improve them? How to adapt them to different purposes?

The fundamental concepts, learning problems, and algorithms are carefully selected. Many of them are closely related to practical questions of the day, such as transfer learning, causal inference, and learning with label noise.
Assumed Knowledge: COMP5318. Students are strongly advised to complete the assumed knowledge unit prior to enrolment as this unit builds on concepts taught in COMP5318.
Timetable: COMP5328 Timetable
Time Commitment:
# Activity Name Hours per Week Sessions per Week Weeks per Semester
1 Lecture 2.00 1 13
2 Tutorial 1.00 1 13
3 Project Work - own time 3.00 10
4 Independent Study 6.00 13

Learning outcomes are the key abilities and knowledge that will be assessed in this unit. They are listed according to the course goal supported by each. See Assessment Tab for details how each outcome is assessed.

(6) Communication and Inquiry/ Research (Level 4)
1. Presentation of the design and evaluation of a machine learning algorithm, describing the design processes and evaluation.
2. Understanding of the variance and bias trade-off in machine learning algorithms.
(4) Design (Level 5)
3. Ability to understand and analyse some machine learning algorithms and have some knowledge to further improve them.
4. Ability to understand and analyse some machine learning problems and have some knowledge to adapt the existing machine learning models to different purposes.
(2) Engineering/ IT Specialisation (Level 5)
5. Experience to implement machine learning algorithms from peer-reviewed papers.
6. Understanding of the nature of the statistical foundations of designing or adapting learning algorithms.
7. Knowledge of the introduced machine learning models and the relative strengths and weaknesses of each and their most appropriate uses.
(1) Maths/ Science Methods and Tools (Level 4)
8. Knowledge of methods to analyse machine learning algorithms, such as hypothesis complexities and generalisation bounds.
Assessment Methods:
# Name Group Weight Due Week Outcomes
1 Quiz* No 10.00 Week 4 1, 2, 6, 8,
2 Assignment 1 Yes 20.00 Week 9 (Thursday) 2, 3, 5, 6, 7, 8,
3 Assignment 2 Yes 20.00 Week 13 (Thursday) 1, 2, 3, 4, 6, 7, 8,
4 Final Exam No 50.00 Exam Period 2, 6, 7, 8,
Assessment Description: Quiz*: Tests in the lab

* indicates an assessment task which must be repeated if a student misses it due to special consideration.

Assignment 1: TBA. Due in Week 9. (It should be done in groups of 2 or 3).

Assignment 2: TBA. Due in Week 13. (It should be done in groups of 2 or 3).

Final Exam: Written Examination

The University has authorised and mandated the use of text-based similarity detecting software Turnitin for all text-based written assignments. This unit of study may also use MOSS for specialised detection of software code similiarity.

Assignments submitted electronically are to be consistently due at 23.59 on the submission day.

Consistent penalty of 5% per day late, e.g.,

a) A “good” assignment that would normally get 9/10, and is 2 days late, loses 10% of the full 10 marks, ie new mark = 8/10.

b) An average assignment, that would normally get 5/10, that is 5 days late, loses 25% of the full 10 marks, ie new mark = 2.5/10

Assignments more than 10 days late get 0.
Grading:
Grade Type Description
Standards Based Assessment Final grades in this unit are awarded at levels of HD for High Distinction, DI (previously D) for Distinction, CR for Credit, PS (previously P) for Pass and FA (previously F) for Fail as defined by University of Sydney Assessment Policy. Details of the Assessment Policy are available on the Policies website at http://sydney.edu.au/policies . Standards for grades in individual assessment tasks and the summative method for obtaining a final mark in the unit will be set out in a marking guide supplied by the unit coordinator.
Minimum Pass Requirement It is a policy of the School of Computer Science that in order to pass this unit, a student must achieve at least 40% in the written examination. For subjects without a final exam, the 40% minimum requirement applies to the corresponding major assessment component specified by the lecturer. A student must also achieve an overall final mark of 50 or more. Any student not meeting these requirements may be given a maximum final mark of no more than 45 regardless of their average.
Policies & Procedures: IMPORTANT: School policy relating to Academic Dishonesty and Plagiarism.

In assessing a piece of submitted work, the School of Computer Science may reproduce it entirely, may provide a copy to another member of faculty, and/or to an external plagiarism checking service or in-house computer program and may also maintain a copy of the assignment for future checking purposes and/or allow an external service to do so.

Other policies

See the policies page of the faculty website at http://sydney.edu.au/engineering/student-policies/ for information regarding university policies and local provisions and procedures within the Faculty of Engineering and Information Technologies.
Recommended Reference/s: Note: References are provided for guidance purposes only. Students are advised to consult these books in the university library. Purchase is not required.
  • A Probabilistic Theory of Pattern Recognition
  • Convex Optimization
  • Foundations of Machine Learning
Online Course Content: USyd eLearning

Note that the "Weeks" referred to in this Schedule are those of the official university semester calendar https://web.timetable.usyd.edu.au/calendar.jsp

Week Description
Week 1 Introduction to Machine Learning Problems
Week 2 Loss Functions and Convex Optimisation
Week 3 Hypothesis Complexity and Generalisation
Week 4 In class quiz
Assessment Due: Quiz*
Week 5 Dictionary Learning and NMF
Week 6 Sparse Coding and Regularisation
Week 7 Learning with Noisy Data
Week 8 Domain Adaptation and Transfer Learning
Week 9 Learning with Noisy Data II: Label Noise
Assessment Due: Assignment 1
Week 10 Mixture Proportion Estimation
Week 11 Causal Inference
Week 12 Multi-task Learning
Week 13 Review
Assessment Due: Assignment 2
Exam Period Assessment Due: Final Exam

Course Relations

The following is a list of courses which have added this Unit to their structure.

Course Year(s) Offered
Bachelor of Advanced Computing/Bachelor of Commerce 2018, 2019, 2020
Bachelor of Advanced Computing/Bachelor of Science 2018, 2019, 2020
Bachelor of Advanced Computing/Bachelor of Science (Health) 2018, 2019, 2020
Bachelor of Advanced Computing/Bachelor of Science (Medical Science) 2018, 2019, 2020
Bachelor of Advanced Computing (Computational Data Science) 2018, 2019, 2020
Bachelor of Advanced Computing (Computer Science Major) 2018, 2019, 2020
Bachelor of Advanced Computing (Information Systems Major) 2018, 2019, 2020
Bachelor of Advanced Computing (Software Development) 2018, 2019, 2020
Graduate Diploma in Computing 2019, 2020
Master of Data Science 2018, 2019, 2020
Master of Information Technology 2019, 2020
Master of Information Technology Management 2019, 2020
Master of IT/Master of IT Management 2019, 2020

Course Goals

This unit contributes to the achievement of the following course goals:

Attribute Practiced Assessed
(6) Communication and Inquiry/ Research (Level 4) No 22.5%
(7) Project and Team Skills (Level 3) No 0%
(5) Interdisciplinary, Inclusiveness, Influence (Level 5) No 0%
(4) Design (Level 5) No 10%
(2) Engineering/ IT Specialisation (Level 5) No 42%
(3) Problem Solving and Inventiveness (Level 5) No 0%
(1) Maths/ Science Methods and Tools (Level 4) No 25.5%

These goals are selected from Engineering & IT Graduate Outcomes Table 2018 which defines overall goals for courses where this unit is primarily offered. See Engineering & IT Graduate Outcomes Table 2018 for details of the attributes and levels to be developed in the course as a whole. Percentage figures alongside each course goal provide a rough indication of their relative weighting in assessment for this unit. Note that not all goals are necessarily part of assessment. Some may be more about practice activity. See Learning outcomes for details of what is assessed in relation to each goal and Assessment for details of how the outcome is assessed. See Attributes for details of practice provided for each goal.