Neural networks

From Uncyclopedia, the content-free encyclopedia.

A neural network is a collection of interconnected neurons. Each neuron can send signals to other neurons via connections which cause those neurons to send signals to other neurons. Those neurons send signals to other neurons causing those neurons to send signals to other neurons. This process doesn't sound very interesting when written out in English, but in mathematics or on computers or even in reality, it's really interesting. Advanced biological life-forms on Earth all have neural networks usually centred around a brain and/or central nervous system.

Signals between neurons are mediated by electrical, chemical and magical means. The speed of these signals enables advanced life-forms to perform incredibly complex tasks by performing an enormous number of computations a second. Curiously, if the same number of computations was performed by a biological neural network in one minute, then less computations will have actually been performed. These computations are performed mostly in binary, but also in welsh or urdu, and have been utilised most recently in Intel's Strawberry Dodecacore 8.

[edit] History

The first reference to a neural network was in the bible. It described that the brain of Jesus had only 3 neurons, a possible origin of the holy trinity. Despite the fact that only three meaningful connections can be made between three neurons, the bible claimed that 8,000 connections existed in Jesus' brain. Biblical scholars have expended much fruitless time studying this apparently meaningless statement, but as Fr. E. Stone from Shannon, Ireland pointed out "we've spent considerably longer studying other parts of the bible and gained even less insight".

The first biological neural networks appeared on Earth not long after life appeared on Earth. The oldest fossil evidence of neural networks can be found in Spain, though the exact location has been kept a closely guarded secret as it's rumoured that contact with the fossils has an effect on the contactee much like that of the Krel machine in the Forbidden Planet.

The first artificial neural networks were developed by Alan Turing during World War II. On visiting Bletchley Park in 1942, Winston Churchill was extremely impressed by what he saw. Churchill was overheard saying "One day, men will fall at the feet of these machines, meek and hollow and in awe of their intelligence and wisdom". Turing, clearly moved to tears, replied "that's the wireless set Prime Minister".

[edit] 1960s

In the late 1960s, a neural network called the perceptron was proposed by Ron L. Hubbard. The perceptron was able to play chess, though always lost and rarely made valid moves. A perceptron was also used transplated into a cat. For a while the cat was able to twitch and bleed copiously but died within minutes. Perceptrons also ran several countries for a trial period in the late 1960s, including the UK, USA and Iran. No perceptible advantage was identified in these tests. In 1969 it was discovered that the humble perceptron was unable to solve the exclusive-OR (XOR) problem. This ended interest in the perceptron, but as Hubbard observed "most adult humans don't even know what the XOR problem is, let alone know how to solve it".

[edit] 1970s

There was no interest in neural networks in the 1970s.

[edit] 1980s and 1990s

By the mid 1980s, with the advent of cheaper computers and the invention of a learning algorithm called bottom-injection, neural networks enjoyed a resurgence of interest. Students and academics across the world spent countless hours trying to get neural networks to perform tasks. A notable success was the neuro-explosion project. A neural network was trained to overload itself, using a recursive algorithm called self-bottom-injection. The idea was that by creating neurons and connections at an ever increasing rate, it would expand at an exponential rate until the rapid pressure build-up caused a small thermonuclear detonation. Neuro-exploder 7 achieved this on 3 April 1986 and exploded. So successful was the explosion that it destroyed the laboratory and killed all the academics, students and technicians who worked on it and all record of how they did it.

Gradually it was realised the neural networks weren't really good at anything and as AIDS awareness grew in developed countries, the practice of bottom-injection became frowned upon.

[edit] Present day

Although no longer of any interest, neural networks have found plenty of niches in modern technology where they are not used, including:

• toasters
• aircraft
• cars
• marmite
• over 60s pornography
• welsh nationalism