Machine Learning for Engineers
| Code | School | Level | Credits | Semesters |
| MTHS3004 | Mathematical Sciences | 3 | 10 | Spring UK |
- Code
- MTHS3004
- School
- Mathematical Sciences
- Level
- 3
- Credits
- 10
- Semesters
- Spring UK
Summary
This is an introduction to machine learning with an emphasis on practical use in engineering applications, such as control theory or decisions under uncertainty. The module is structured around a subset of the following topics: Inference, or model-driven predictions; Supervised learning: the basics, optimization and beyond least squares; Unsupervised learning; Neural networks and deep learning; The Bayesian approach and Gaussian processes; Generalized linear models: going beyond normality.
Target Students
BEng and MEng students in the Faculty of Engineering and PGT Students in the Schools of Mathematics, and Physics and Astronomy. Numbers may be restricted by the availability of computer laboratory places.
Classes
- One 1-hour workshop each week for 11 weeks
- One 2-hour lecture each week for 11 weeks
Each week there will normally be 2 lectures to introduce key mathematical knowledge/ideas/techniques on module topics. Alternate weeks 1 hour of worked examples to facilitate solving of problems/tutorial/problem class or provide students with the opportunity to gain individual help understanding module topics, clarification of lecture notes or support in developing problem solving skills.
Assessment
- 50% Coursework 1: Coursework. Reassessment: 1 coursework covering all learning outcomes of the module.
- 50% Coursework 2: Coursework. Reassessment: 1 coursework covering all learning outcomes of the module.
Assessed by end of spring semester
Educational Aims
To introduce computational and statistical methods that are becoming increasingly important in engineering for the solution of complex problems in a data-driven framework; to tailor the methods to specific applications in engineering and to understand which method to select given a specific problem; to implement these techniques in a computer language to generate numerical solutions as an aid to understanding and validating the introduced techniques.Learning Outcomes
A student who completes this module successfully should be able to:
L1 - Understand the key concepts of machine learning and when to use it.
L2 - Harness the power of statistical learning approaches to gain insight into large data sets.
L3 - Understand the practical advantages and disadvantages of the different methods and how to balance them for a given problem.
L4 - Develop a computational implementation of the different algorithms in a suitable computational environment.
L5 - Analyse and interpret the output of a data-driven method.