Module Title:   Neural Networks and Fuzzy Systems

Module Credit:   20

Module Code:   CM-0340D

Academic Year:   2015/6

Teaching Period:   Semester 1

Module Occurrence:   A

Module Level:   FHEQ Level 6

Module Type:   Standard module

Provider:   Computer Science

Related Department/Subject Area:   SCIM (Dept of Computer Science)

Principal Co-ordinator:   Dr Y Peng

Additional Tutor(s):   Dr A Csenki

Prerequisite(s):   CM-0318L

Corequisite(s):   None

Aims:
To understand neural networks and fuzzy logic theory; to gain knowledge of neural networks and fuzzy system development; to apply the techniques for solving real-life problems, such as signal processing, control, and pattern recognition.

Learning Teaching & Assessment Strategy:
Formal lectures will outline the theoretical principles. Practical exercises will be carried out to link the theories with practical implementation in laboratory by providing sample codes of typical neural networks and fuzzy logic systems. The module will be assessed through a project to apply the learned knowledge to solve a real-life problem. Several assessed exercises will be provided for testing the understanding of theoretical knowledge and the practice of implementation skills. The supplementary assessment is to repair deficiencies identified in the original assessment. Oral feedback will be given during laboratory classes. Students who have an acceptable body of knowledge equivalent to the pre-requisite shown, CM-0318L, will be permitted to study this module.

Lectures:   20.00          Directed Study:   152.00           
Seminars/Tutorials:   4.00          Other:   0.00           
Laboratory/Practical:   24.00          Formal Exams:   0.00          Total:   200.00

On successful completion of this module you will be able to...

demonstrate detailed knowledge and systematic understanding of essential facts, concepts, principles and theories relating to computing and computer applications of neural networks and fuzzy systems.
Use such knowledge and understanding in the modelling and design of computer-based neural networks and fuzzy systems.
Demonstrate advance knowledge and understanding of mathematical principles necessary to underpin the module and the ability to apply mathematical methods, tools and notations proficiently in the analysis and solution to problems.

On successful completion of this module you will be able to...

demonstrate the advance ability to specify, analyse, design and train feedforward neural networks, recurrent neural networks, and self-organising neural networks, fuzzy logic control systems, evolution computing for sensor-actuator mapping and pattern classification.
Critically analyse and deploy appropriate theory, practices and tools for the specification, design and implementation of neural networks and fuzzy systems.
Apply the principles of appropriate supporting engineering and scientific disciplines.

On successful completion of this module you will be able to...

demonstrate extended analytical and problem-solving skills.
Further advance your numeracy in both understanding and presenting cases involving a quantitative dimension.
Communicate effectively in electronic as well as written and oral form.

  Coursework   50%
 
  Neural Networks exercises and coursework
  Coursework   50%
 
  Fuzzy logic exercises and coursework
  Examination - closed book 3.00 100%
 
  Supplementary Examination for assessment of both content understanding and application capability

Outline Syllabus:
Typical neural networks, including M-P model, Perceptron, Multi-layered Perceptron, ADALINE, MADALINE, Radial Basis Function, Recurrent Neural Networks, Hopfield Networks, Adaptive Resonance Theory Neural Networks, Kohonen Neural Networks. Fuzzy set theory and fuzzy logic control. Typical learning and searching techniques, including supervised learning, unsupervised learning, competitive learning, hill climbing, and simulated annealing. A practical project, may involve both hardware and software development, to apply the taught neural networks and fuzzy logic for solving a real-life problem such as pattern recognition, data mining, or behaviour learning.

Version No:  4