Understand the basic structures and the learning mechanisms underlying neural networks within the field of artificial intelligence and examine how synaptic adaptation can facilitate learning and how input to output mapping can be performed by neural networks. Obtain an overview of linear, nonlinear, separable and non separable classification as well as supervised and unsupervised mapping. Understand the benefit of adopting naturally inspired techniques to implement optimisation of complex systems and acquire the fundamental knowledge in various evolutionary techniques. Become familiar with the basic concepts of systems optimisation and its role in natural and biological systems and entities.

(LO1) Learning  the advantages and main characteristics of neural networks in relation to traditional methodologies. Also, familiarity with different neural networks structures and their learning mechanisms.

(LO2) Appreciation of the advantages of evolutionary-related approaches for optimisation problems and their advantages compared to traditional methodologies. Also, understanding the different techniques of evolutionary optimisation for discrete and continuous configurations

(LO3) Understanding of the needs for genetic encoding and modelling for solving optimisation problems and familiarisation with the evolutionary operators and their performance.

(LO4) Understanding of the neural network learning processes and their most popular types, as well as  appreciation of how neural networks can be applied to artificial intelligence problems.

(S1) On successful completion of this module the student should be able to pursue further study in artificial intelligence as well as more advanced types of neural networks and evolutionary optimisation and bio-inspired techniques.

(S2) On successful completion of this module the student should be able to analyse numerically the mathematical properties of most major network types and apply them to artificial intelligence problems. Also, the student should be able to appreciate and understand the suitability of evolutionary optimisation in systems where classical methods cannot be effective.

(S3) On successful completion of this module the student should be able to approach methodologically artificial intelligence problems and bio-inspired algorithms in general and understand the principal mathematics of learning systems and the fundamental principles governing evolutionary optimisation techniques.

Teaching Method 1 - Lecture
Description: Part 1: Neural Networks (12) and Part 2: Evolutionary Computation (12)
Attendance Recorded: Yes

Teaching Method 2 - Tutorial
Description: Slide Presentation and Blackboard
Attendance Recorded: Yes

Standard on-campus delivery
Teaching Method 1 - Lecture
Description: Mix of on-campus/on-line synchronous/asynchronous sessions
Teaching Method 2 - Tutorial
Description: On-campus synchronous sessions

For Neural Networks (part I), 12 lectures delivering the following chapters:
Introduction: Chapter 1
Structural Aspects: Chapter 2
Learning Processes: Chapter 3
Single-Layer Perceptrons: Chapter 4
Multi-Layer Perceptrons: Chapter 5
Radial-basis Function Networks: Chapter 6
Support Vector Machines: Chapter 7
Self-Organising Maps: Chapter 8

For Evolutionary Computation (part II), 12 lectures delivering the following chapters:
Introduction: Chapter 1
Genetic Algorithms: Chapter 2 (basic elements), Chapter 3 (advanced topics), Chapter 4 (theoretical analysis)
Genetic Programming: Chapter 5 (genetic programming & gene expression programming)
Particle Swarm Optimisation: Chapter 6 (overview and extensions) Evolutionary Strategies: Chapter 7 (overview and extensions)