Introduction to Neural Networks

Neural Networks

Hi and welcome to the introduction to Neural Networking; We will start, as all do. with an introduction to early computing and networks – if you wish to jump any of this please click on the next menu of the Neural Networks tab – alternately you can Click Here for the next section. This is a rather potted history and will probably not add much to your existing knowledge however, if you are interested then read on:

For many centuries ‘Humans’ have long been interested in computers as an aid to thinking. This goes right back many many centuries, to just a few lines drawn in the sand, in each line a pebble is placed – each pebble is a counter and when each line is filled, a pebble is moves over. This is both an aid to keeping tally, it is also a computer and, an Algorithm.

From a few lines in the sand came, around 3000 years ago : 1300BC in China – the abacus – a calculating device, used for counting which is still in popular use in China today.

Stonehenge

Then we can look at the Henge’s – such as Stonehenge; a monolithic device for counting the days of the year and keeping a tally of when it is best to celebrate the new year and consequently – when best to plant and reap the harvest

Next around 1000 years ago circa 100 BC came the Antikythera mechanism which is an ancient hand powered Greek analogue computer a device that kept track of the sun, moon and time. It is thought it possibly helped keep track of the time of year for the four year cycle of the Olympics.

Jumping now to the 19th century – we have the difference machine imagined by Charles Babbage- a device that could calculate numbers using gears and dials and a cranking handle which could produce results without the need for human mental calculation

Then in the early 20th century, based partly on the difference engine, innovations in electrical currents, relays and solenoids ( electrical switches ) and partly developing from ideas based on Binary logic came the early computers which were slow, large and electro-mechanical in their operation.

During the 1930’s and 1940’s came the development of the thermionic valve – a valve which could replicate electrically the mechanical relays and switches of the earlier computers . Computers were still large but less noisy – and still prone to breaking down.

It had been known, since Galvani in the 1780’s that the body had electricity within it. And that muscles worked through small electrical impulses.However, the electrical currents in the brain are so very small that it took the development of fine instruments to see the tiny currents at work in the brain.

This happened around 1924 when Hans Berger produced the first pattern of electrical activity – The Electroencephalograph – EEG of the human brain.

The Transistor.

Vintage Computers at Bletchley Park – Courtesy of Alex Motoc

It was in late 1947 that the transistor was invented – the transistor could replicate the switching mechanism of the earlier thermionic valves – It was developed further in the 1950’s and was more reliable and led to faster computations. This together with bigger storage capacity – During this time many innovations were made in the field of computing:

It was from the development of transistorized computers together with the development of the electron microscope that allowed the microscopic analysis of brain nerve cells in the 1950’s that the neurone theory of nerve cells in the brain was formulated – helped along with the Electroencephalograph which illustrated the electrical activity within the brain.

All these ideas coalesced in the concept of the artificial neural network. In essence, it brought forward the idea that the switching mechanism inherent in the transistor could replicate the switching mechanisms of the neurones in the brain.

To continue please go to the next part : The Neurone Theory of the Brain alternately you can Click Here.