MT4570 Statistical Machine Learning

Academic year

2025 to 2026 Semester 2

Further information on which modules are specific to your programme.

Key module information

SCOTCAT credits

15

The Scottish Credit Accumulation and Transfer (SCOTCAT) system allows credits gained in Scotland to be transferred between institutions. The number of credits associated with a module gives an indication of the amount of learning effort required by the learner. European Credit Transfer System (ECTS) credits are half the value of SCOTCAT credits.

SCQF level

SCQF level 10

The Scottish Credit and Qualifications Framework (SCQF) provides an indication of the complexity of award qualifications and associated learning and operates on an ascending numeric scale from Levels 1-12 with SCQF Level 10 equating to a Scottish undergraduate Honours degree.

Availability restrictions

Not automatically available to General Degree students

Planned timetable

To be confirmed

This information is given as indicative. Timetable may change at short notice depending on room availability.

Module coordinator

Dr G Minas

Dr G Minas
This information is given as indicative. Staff involved in a module may change at short notice depending on availability and circumstances.

Module Staff

Dr Giorgos Minas

This information is given as indicative. Staff involved in a module may change at short notice depending on availability and circumstances.

Module description

Machine learning tools are widely used across various aspects of contemporary society. A solid understanding of the foundations that support these tools is essential for developing new methods and assessing existing ones. The aim of this module is to introduce the mathematical and statistical theory behind modern machine learning methods. This module will explore the ‘why’ and ‘how’ of machine learning methods from a theoretical perspective considering generalisation, regularisation, and optimisation. Key techniques including kernel-based methods, tree-based methods and neural networks will be discussed, along with recent developments.

Relationship to other modules

Pre-requisites

BEFORE TAKING THIS MODULE YOU MUST PASS MT2501 AND PASS MT2503 AND PASS MT2508

Assessment pattern

Coursework = 10%, Examination = 90%

Re-assessment

Oral Examination = 100%

Learning and teaching methods and delivery

Weekly contact

2.5 lectures (x 10 weeks) and 1 tutorial (x 10 weeks).

Scheduled learning hours

35

The number of compulsory student:staff contact hours over the period of the module.

Guided independent study hours

112

The number of hours that students are expected to invest in independent study over the period of the module.

Intended learning outcomes

  • Describe machine learning methods and understand their assumptions to assess their suitability for classification and regression tasks.
  • Mathematically formulate machine learning methods used in classification and regression, including kernel-based methods, tree-based methods and neural networks.
  • Derive the statistical properties of machine learning methods and utilise them to analyse, evaluate, and criticise the performance of each method.
  • Construct optimisation methods for machine learning tasks and assess their effectiveness in different settings.