This 10-credit Level 9 Certificate provides an overview of the main concepts and techniques in machine learning. Modern AI systems use machine learning to learn from data, to build flexible representations from that data, and to use those representations to detect fraud, to classify blood samples as diseased or not, to discover relationships between criminals, to develop new antibiotics, to recommend content in social media feeds, and to answer prompts and thus to power AI agents. Machine learning draws upon statistics, probability, numerical optimisation and linear algebra, combined with computing concepts like data management, networking and programming. Compared to traditional programming, with its tightly defined specifications, machine learning follows general principles which are configured to match the characteristics of the data. Choosing, applying and validating these principles effectively is the main aim of this programme.
You will be introduced to the data model building pipeline, covering exploratory data analysis, feature engineering, model selection and optimisation and validation, underpinned by visualisation. Each machine learning task is introduced, its strengths and weaknesses are highlighted, and guidance for its use is given. Such tasks include clustering, classification, regression, and rule learning. The programme is Python-based and uses pandas for data processing, scikit-learn for most of the machine learning tasks, and matplotlib/seaborn for visualising both the data and the models that we derive from this data. The weekly Python practicals show how machine learning works in practice, on real data sets. Assessment is based on being able to investigate data effectively using the tools from the programme.
Students taking this Certificate should have some programming experience (and hence an understanding of control flow constructs, variables, invoking other modules, etc.). Experience of Python programming is highly desirable but not essential. Some familiarity with basic statistics (mean, variance, etc.) is desirable but a review of such concepts is provided at the start.
Unique Features
This Certificate provides an overview of the main concepts and techniques in machine learning. Modern AI systems make extensive use of machine learning.
Delivery
Lectures will be classroom based but remotely accessible. Students may attend lab classes in person or get remote support from the lecturers. Students will be required to attend in person for a module assessment.
Machine Learning (10 credits)
Applicants will normally require a second class honours award in an honours degree in computer science, or equivalent. Exceptions to this are considered on a case by case basis where the applicant has substantial relevant work experience. Applicants whose previous higher education took place in a location where English is not the first spoken language must provide certification to meet SETU's English Language requirements.
Credit from this Certificate is transferable to our MSc in Computer Science programme.
Course Leader