Workload - Private Study - Assessment - Description - Aims - Learning Outcomes - Content - Teaching Materials - Recommended Books
Module Code | COM00076M |
---|---|
Lecturers | Simon O'Keefe |
Taken By | LSCITS 1, LSCITS 2, LSCITS 3, LSCITS 4, MEng CSAI 4, MEng CSBES 4, MEng CSESE 4, MEng CSSE 4, MMath 4, NC |
Number of Credits | 10 |
Teaching | Autumn 2-5 |
Open Assessment |
Open Assessment [100%]
7th Nov → 9th Jan Feedback: 6th Feb |
Exercises for this module will require the use of MATLAB. At the start of the module there will be a very short introduction to MATLAB. Some of the training methods for neural systems are described using differential calculus. Some of the behaviours of neurons are described using differential equations.
You may not opt for this module if you have already taken either PAT or INCA in Stage 3.
Exercises will be provided for you to work on in your own time. Practical sessions will be available to provide assistance with the exercises. Model solutions to exercises will be provided after you have had a chance to work on the exercises.
The lectures will describe architectures and algorithms for neural computing. The exercises will reinforce some aspects of this. You will be expected to read around the lecture material to gain a clearer understanding.
Open assessment may include some or all of literature review, construction of a neural network using appropriate software, and design of experiments to evaluate a neural network in the context of a specific problem.
Feedback will be given on solutions submitted to exercises distributed during the module.
This module provides a foundation of theoretical and practical knowledge in the subject of neural computing systems; that is, systems that embody algorithms inspired by natural neural systems. After a brief introduction to biological (natural) neural systems, the main emphasis is on the principal artificial neural architectures derived from those natural systems. The emphasis will be on the characterisation of the artificial systems, rather than the analysis of their properties in statistical terms.
• Identify which neural system is suitable for a particular task.
• Design, implement and experiment with neural architectures for a particular task.
• Design appropriate encodings of data.
• Evaluate the application of a particular architecture to a given problem.
The topics to be covered in this module will include:
* Natural neural systems
* Learning in neural systems
* The perceptron
* The multilayer perceptron
* Radial basis function networks
* Associative memories
* Self-organising systems
* Spiking neurons
Copies of slides used in lectures and brief summaries of lecture content will be made available via the module web page. These are not substitutes for taking notes and further reading.
Rating | Author | Title | Publisher | Year |
---|---|---|---|---|
+++ | Haykin, S | Neural Networks and Learning Machines | Pearson | 2009 |
++ | Bishop, C | Neural Networks for Pattern Recognition | OUP | 1995 |
++ | Callan | The Essence of Neural Networks | Prentice Hall | 1999 |
++ | Beale and Jackson | Neural Computing: an introduction | Institute of Physics | 1990 |
Last updated: 27^{th} February 2013