James Cook University Subject Handbook - 2024

MA3832 - Neural Network and Deep Learning

Subject Description

This subject in a continuation from MA3405/MA5405. In this subject, students will learn the mechanics of modern neural networks, essential skill in cloud comping and deploy the model onto cloud platform. The concepts of regularisation, pooling, optimisation momentum and convolution are introduced and applied in the field of computer vision neural networks. In this subject, Amazon is used to apply machine learning optimisation to machine learning tasks. This methodology of using machine learning for optimisation greatly enhances the outcomes of model development and removes some of the intuition needed for model development. At the conclusion of the subject, students will provide a detailed computer vision project plan using industry standard planning method.

Learning Outcomes

• demonstrate and apply advanced theoretical and technical knowledge of neural network to an industry or research problem
• develop multi-layer perceptron and convolution neural networks
• programmatically interact with AWS using Python Jupyter Notebooks
• develop and deploy machine learning models on AWS

Subject Assessment

• Written > Essay (including multi-draft) 1 - (25%) - Individual
• Written > Case study analysis - (25%) - Individual
• Written > Project report - (50%) - Individual

Note that minor variations might occur due to the continuous subject quality improvement process, and in case of minor variation(s) in assessment details, the Subject Outline represents the latest official information.

Availabilities