ETC3555 · Statistical machine learning

Overview

This unit covers the methods and practice of statistical machine learning for modern data analysis problems. You will take a deep look at the procedure of learning from data with particular focus placed on how to effectively learn model parameters and methods to guard against overfitting. Topics covered will include stochastic gradient descent, deep neural networks with dropout, convolutional neural networks for image recognition, and text mining and generation with recurrent neural networks. All computing will be conducted using open source software. Introductory machine learning methods such as linear models, decision trees, random forests, and hierarchical clustering, are assumed.

Offerings

Assessment

The Handbook lists an examination for this unit.

Assessment details may change. Please refer to the assessment information in Moodle closer to the start of the teaching period.

Requisites

Learning outcomes

1. identify and understand the statistical and computational trade-offs in modern data analysis problems
2. develop computer skills for exploring modern data sets and applying state-of-the-art machine learning algorithms
3. understand and apply machine learning algorithms to solve modern data analysis problems.

Workload

Minimum total expected workload to achieve the learning outcomes for this unit is 144 hours per semester typically comprising a mixture of scheduled learning activities and independent study. Independent study may include associated readings, assessment and preparation for scheduled activities. You are expected to complete all pre-class activities prior to your scheduled class, and post-class activities should be completed after your scheduled class. Learning activities may include a combination of teacher directed, peer directed and online engagement activities.

Student experience, not official rules. Nothing here changes what the Handbook says.