Unit Overview

Description

This unit examines the mathematics behind modern artificial intelligence (AI), with a specific focus on the analysis of neural networks and deep learning. The emphasis is on foundational theory rather than the application of AI. Topics may include linear algebra formalisms, the multi-layer perceptron, the universal approximation theorem for neural networks, expressivity and the role of depth, backpropagation/automatic differentiation, gradient-based (stochastic) optimisation, the geometric structure of piecewise linear networks, random initialisation and signal propagation, benign overfitting, and reservoir computing as a dynamical systems approach to recurrent networks.

Credit
6 points
Offering
AvailabilityLocationModeFirst year of offer
Not available in 2026UWA (Perth)On-campus
Details for undergraduate courses
  • Level 3 elective
Outcomes

Students are able to (1) demonstrate an understanding of the mathematics underlying neural network models and related AI methods; (2) apply core results from approximation theory, optimisation, and linear algebra for the analysis of AI methods; and (3) articulate the theoretical role of depth, nonlinearity, and parameterisation in modern learning systems.

Assessment

Indicative assessments in this unit are as follows: (1) assignments; (2) tests or quizzes; and (3) final examination. Further information is available in the unit outline.



Student may be offered supplementary assessment in this unit if they meet the eligibility criteria.

Unit rules
Prerequisites
Successful completion of
MATH1011 Multivariable Calculus
or MATH1013 Mathematical Analysis
and
MATH1012 Mathematical Theory and Methods
or MATH1014 Algebra
Advisable prior study
MATH2064 Numerical Methods and either STAT2062 Fundamentals of Probability with Applications
or STAT2063 Probabilistic Methods and their Applications
Contact hours
lectures: 3 hours per week
workshops: 2 hours per week
  • The availability of units in Semester 1, 2, etc. was correct at the time of publication but may be subject to change.
  • All students are responsible for identifying when they need assistance to improve their academic learning, research, English language and numeracy skills; seeking out the services and resources available to help them; and applying what they learn. Students are encouraged to register for free online support through GETSmart; to help themselves to the extensive range of resources on UWA's STUDYSmarter website; and to participate in WRITESmart and (ma+hs)Smart drop-ins and workshops.
  • Visit the Essential Textbooks website to see if any textbooks are required for this Unit. The website is updated regularly so content may change. Students are recommended to purchase Essential Textbooks, but a limited number of copies of all Essential Textbooks are held in the Library in print, and as an ebook where possible. Recommended readings for the unit can be accessed in Unit Readings directly through the Learning Management System (LMS).
  • Contact hours provide an indication of the type and extent of in-class activities this unit may contain. The total amount of student work (including contact hours, assessment time, and self-study) will approximate 150 hours per 6 credit points.