|Not a member yet? Register
for full benefits!
An Introduction to Neural Networks
- Dr. Leslie Smith
Centre for Cognitive and Computational
Department of Computing and Mathematics
University of Stirling.
last major update: 25 October 1996: last minor update 22 April 1998
and 12 Sept 2001: links updated (they were really out of date) 12 Sept
- This document is a roughly HTML-ised version of a talk given at the
NSYN meeting in Edinburgh, Scotland, on 28 February 1996, then updated
a few times in response to comments received. Please email me comments,
but remember that this is just the slides from an introductory talk!
Why would anyone want a `new' sort of computer?
What is a neural network?
Some algorithms and architectures.
Where have they been applied?
What new applications are likely?
Some useful sources of information.
Some comments added Sept 2001
Why would anyone want a `new' sort
What are (everyday) computer systems good at... .....and not so good
Not so good at
||Interacting with noisy data or data from the environment
|Doing precisely what the programmer programs them to do
|Adapting to circumstances
Where can neural network systems help?
we can't formulate an algorithmic solution.
we can get lots of examples of the behaviour we require.
we need to pick out the structure from existing data.
What is a neural network?
Neural Networks are a different paradigm for computing:
Neumann machines are based on the processing/memory abstraction of human
networks are based on the parallel architecture of animal brains.
Neural networks are a form of multiprocessor computer system, with
degree of interconnection
interaction between elements
A biological neuron may have as many as 10,000 different inputs, and
may send its output (the presence or absence of a short-duration spike)
to many other neurons. Neurons are wired up in a 3-dimensional pattern.
Real brains, however, are orders of magnitude more complex than any artificial
neural network so far considered.
Example: A simple single unit adaptive network:
||The network has 2 inputs, and one output. All are binary. The output
1 if W0 *I0 + W1 * I1 + Wb > 0
0 if W0 *I0 + W1 * I1 + Wb <= 0
We want it to learn simple OR: output a 1 if either I0 or I1 is
Algorithms and Architectures.
The simple Perceptron:
The network adapts as follows: change the weight by an amount proportional
to the difference between the desired output and the actual output.
As an equation:
&Delta Wi = &eta * (D-Y).Ii
where &eta is the learning rate, D is the desired output, and Y is
the actual output.
This is called the Perceptron Learning Rule, and goes back to
the early 1960's.
We expose the net to the patterns:
We train the network on these examples. Weights after each
epoch (exposure to complete set of patterns)
At this point (8) the network has finished learning. Since
(D-Y)=0 for all patterns, the weights cease adapting. Single perceptrons
are limited in what they can learn:
If we have two inputs, the decision surface is a line. ... and
its equation is
In general, they implement a simple hyperplane decision surface
This restricts the possible mappings available.
Developments from the simple perceptron:
Back-Propagated Delta Rule Networks (BP) (sometimes known and multi-layer
perceptrons (MLPs)) and Radial Basis Function Networks (RBF) are both
well-known developments of the Delta rule for single layer networks (itself
a development of the Perceptron Learning Rule). Both can learn arbitrary
mappings or classifications. Further, the inputs (and outputs) can have
Back-Propagated Delta Rule Networks (BP)
is a development from the simple Delta rule in which extra hidden layers
(layers additional to the input and output layers, not connected externally)
are added. The network topology is constrained to be feedforward:
i.e. loop-free - generally connections are allowed from the input layer
to the first (and possibly only) hidden layer; from the first hidden layer
to the second,..., and from the last hidden layer to the output layer.
Typical BP network architecture:
The hidden layer learns to recode (or to provide a representation
for) the inputs. More than one hidden layer can be used.
The architecture is more powerful than single-layer networks: it
can be shown that any mapping can be learned, given two hidden layers
The units are a little more complex than those in the original
perceptron: their input/output graph is
As a function:
The graph shows the output for k=0.5, 1, and 10, as the
activation varies from -10 to 10.
Training BP Networks
The weight change rule is a development of the perceptron learning rule.
Weights are changed by an amount proportional to the error at that unit
times the output of the unit feeding into the weight.
Running the network consists of
For each data pair to be learned a forward pass and backwards pass is performed.
This is repeated over and over again until the error is at a low enough
level (or we give up).
- Forward pass:
- the outputs are calculated and the error at the output units calculated.
- Backward pass:
- The output unit error is used to alter weights on the output units.
Then the error at the hidden nodes is calculated (by back-propagating
the error at the output units through the weights), and the weights
on the hidden nodes altered using these values.
Radial Basis function Networks
Radial basis function networks are also feedforward, but have only one
Typical RBF architecture:
Like BP, RBF nets can learn arbitrary mappings: the primary difference
is in the hidden layer.
RBF hidden layer units have a receptive field which has
a centre: that is, a particular input value at which they
have a maximal output.Their output tails off as the input moves
away from this point.
Generally, the hidden unit function is a Gaussian:
Gaussians with three different standard deviations.
Training RBF Networks. RBF networks are trained by
Generally, the centres and SDs are decided on first by examining the vectors
in the training data. The output layer weights are then trained using the
Delta rule. BP is the most widely applied neural network technique. RBFs
are gaining in popularity.
on how many hidden units there should be
on their centres and the sharpnesses (standard deviation) of their Gaussians
up the output layer.
Nets can be
RBFs have the advantage that one can add extra units with centres near parts
of the input which are difficult to classify. Both BP and RBFs can also
be used for processing time-varying data: one can consider a window
on the data:
on classification data (each output represents one class), and then
used directly as classifiers of new data.
trained on (x,f(x)) points of an unknown function f, and then used to
Networks of this form (finite-impulse response) have been used in many
There are also networks whose architectures are specialised for processing
Simple Perceptrons, BP, and RBF networks need a teacher to tell the network
what the desired output should be. These are supervised networks.
In an unsupervised net, the network adapts purely in response to its
inputs. Such networks can learn to pick out structure in their input.
Applications for unsupervised nets
Although learning in these nets can be slow, running the trained net is
very fast - even on a computer simulation of a neural net.
- clustering data:
- exactly one of a small number of output units comes on in response
to an input.
- reducing the dimensionality of data:
- data with high dimension (a large number of input units) is compressed
into a lower dimension (small number of output units).
Kohonen clustering Algorithm:
- takes a high-dimensional input, and clusters it, but retaining
some topological ordering of the output.
After training, an input will cause some the output units
in some area to become active.
Such clustering (and dimensionality reduction) is very useful as
a preprocessing stage, whether for further neural network data processing,
or for more traditional techniques.
Where are Neural Networks applicable?
..... or are they just a solution in search of a problem?
Neural networks cannot do anything that cannot be done using traditional
computing techniques, BUT they can do some things which would otherwise
be very difficult.
In particular, they can form a model from their training data (or possibly
input data) alone.
This is particularly useful with sensory data, or with data from a complex
(e.g. chemical, manufacturing, or commercial) process. There may be an
algorithm, but it is not known, or has too many variables. It is easier
to let the network learn from examples.
Neural networks are being used:
- in investment analysis:
- to attempt to predict the movement of stocks currencies etc., from
previous data. There, they are replacing earlier simpler linear models.
- in signature analysis:
- as a mechanism for comparing signatures made (e.g. in a bank) with
those stored. This is one of the first large-scale applications of neural
networks in the USA, and is also one of the first to use a neural network
- in process control:
- there are clearly applications to be made here: most processes cannot
be determined as computable algorithms. Newcastle University Chemical
Engineering Department is working with industrial partners (such as
Zeneca and BP) in this area.
- in monitoring:
- networks have been used to monitor
state of aircraft engines. By monitoring vibration levels and sound,
early warning of engine problems can be given.
Rail have also been testing a similar application monitoring diesel
- in marketing:
- networks have been used to improve marketing mailshots. One technique
is to run a test mailshot, and look at the pattern of returns from this.
The idea is to find a predictive mapping from the data known about the
clients to how they have responded. This mapping is then used to direct
To probe further:
longer introduction (which is more commercially oriented) is hosted
by StatSoft, Inc
The Neural Computing Applications Forum
runs meetings (with attendees from industry, commerce and academe) on
applications of Neural Networks. Contact NCAF through Dr. Tom Harris,
(44) 1784 477271.
Internet addresses: NeuroNet which was at Kings College, London, was
a European Network of Excellence in Neural Networks which finished in
March 2001. Howwever, their website
remains a very useful source of information
IEEE Neural Networks Councilhttp://www.ieee.org/nnc/index.html
CRC NCRC Institute for Information Technology Artificial Intelligence
subject index has a useful entry on Neural
Newscomp.ai.neural-nets has an very useful set
of frequently asked questions (FAQ's), available as a WWW document at:
Quite a few organisations run courses: we used to run a 1 year Masters course
in Neural Computation:
unfortunately, this course is in abeyance. We can even run courses to suit
you. We are about to start up a centre in Computational Intelligence, called
More Specialised Information
Some further information about applications can be found at the Stimulation
Initiative for European Neural Applications (SIENA) pages, and there
is also an interesting page
For more information on Neural Networks in the Process Industries, try
A. Bulsari's home page .
The company BrainMaker has a nice list of references
on applications of its software package that shows the breadth of
The best journal for application-oriented information is
Neural Computing and Applications, Springer-Verlag. (address:
Sweetapple Ho, Catteshall Rd., Godalming, GU7 3DJ)
There's a lot of books on Neural Computing. See the FAQ
above for a much longer list.
For a not-too-mathematical introduction, try
Fausett L., Fundamentals of Neural Networks, Prentice-Hall, 1994.
ISBN 0 13 042250 9 or
Gurney K., An Introduction to Neural Networks, UCL Press, 1997,
ISBN 1 85728 503 4
Haykin S., Neural Networks , 2nd Edition, Prentice Hall, 1999, ISBN 0
13 273350 1 is a more detailed book, with excellent coverage of the whole
Where are neural networks going?
A great deal of research is going on in neural networks worldwide.
This ranges from basic research into new and more efficient learning
algorithms, to networks which can respond to temporally varying patterns
(both ongoing at Stirling), to techniques for implementing neural networks
directly in silicon. Already one chip commercially available exists, but
it does not include adaptation. Edinburgh University have implemented
a neural network chip, and are working on the learning problem.
Production of a learning chip would allow the application of this technology
to a whole range of problems where the price of a PC and software cannot
There is particular interest in sensory and sensing applications: nets
which learn to interpret real-world sensors and learn about their environment.
New Application areas:
Reading this through, it is a bit outdated: not that there's anything incorrect
above, but the world has moved on. Neural Networks should be seen as part
of a larger field sometimes called Soft Computing or Natural
Computing. In the last few years, there has been a real movement of
the discipline in three different directions:
- Pen PC's
- PC's where one can write on a tablet, and the writing will be recognised
and translated into (ASCII) text.
- Speech and Vision recognition systems
- Not new, but Neural Networks are becoming increasingly part of such
systems. They are used as a system component, in conjunction with traditional
- White goods and toys
- As Neural Network chips become available, the possibility of simple
cheap systems which have learned to recognise simple entities (e.g.
walls looming, or simple commands like Go, or Stop), may lead to their
incorporation in toys and washing machines etc. Already the Japanese
are using a related technology, fuzzy logic, in this way. There is considerable
interest in the combination of fuzzy and neural technologies.
- Neural networks, statistics, generative models, Bayesian inference
- There is a sense in which these fields are coalescing. The real problem
is making conclusions from incomplete, noisy data, and all of these
fields offer something in this area. Developments in the mathematics
underlying these fileds have shown that there are real similarities
in the techniques used. Chris Bishop's book
Neural Networks for Pattern Recognition, Oxford University Press is
a good start on this area.
- Neuromorphic Systems
- Existing neural network (and indeed other soft computing) systems
are generally software models for solving static problems on PCs. But
why not free the concept from the workstation? The area of neuromorphic
systems is concerned with real-time implementations of neurally inspired
systems, generally implemented directly in silicon, for sensory and
motor tasks. Another aspect is direct implementation of detailed aspects
of neurons in silicon (see Biological Neural Networks below). The main
centres worldwide are at the Institute
for neuroinformatics at Zurich, and at the
Center for Neuromorphic Systems Engineering at Caltech. There are
also some useful links at this
page (from a UK EPSRC Network Project on Silicon
- Biological Neural Networks
- There is real interest in how neural network research and neurophysiology
can come together. The pattern recognition aspects of Artificial Neural
Networks don't really explain too much about how real brains actually
work. The field called Computational
Neuroscience has taken inspiration from both artificial neural networks
and neurophysiology, and attempts to put the two together.
- Back to top of page