Neural network

A Neural network (also called an ANN or an Articial

Neural network) is an articial system (made of articial
neuron cells). It is modeled after the way the human brain
works through imitating how the brains neurons are red
or activated. Several computing cells work in parallel to
produce a result. This is usually seen as one of the possible ways articial intelligence can work. Most neural
networks can tolerate if one or more of the processing
cells fail.

1.2 Unsupervised Learning

What is important in the idea of neural networks is that

they are able to learn by themselves, an ability which
makes them remarkably distinctive in comparison to normal computers, which cannot do anything for which they
are not programmed.

1.3 Reinforcement Learning

Unsupervised learning only trains using inputs, and the

network has to gure out how they relate to each other.
Clustering problems, estimation problems, and self organising maps work in this way. For example, a self organising map can be used to categorise iris owers by
stem size and colour.[2]

A reinforcement learning neural network uses an objects

or a teachers actions. It works out the smallest cost and
tries to use this to work out how to make the smallest
cost in future. It can be thought of as a Markov decision
process. Another simple way to think of this is as "carrot
and stick" learning (learning that rewards good behaviour
and punishes bad behaviour).

Learning methods

Recently, a research team from the University of Hertfordshire, UK used reinforcement learning to make an
There are three ways a neural network can learn: iCub humanoid robot learn to say simple words by
Supervised learning, Unsupervised learning and babbling.[3]
Reinforcement learning. All these methods work by
nding the biggest or smallest answer to a cost function.
Each method takes dierent inputs so each method is 2 Notes
better at certain tasks.
[1] http://www.ri.cmu.edu/research_project_detail.html?
project_id=160&menu_id=261

1.1

[2] http://www.mathworks.com/matlabcentral/
fileexchange/32664-iris-data-set-clustering

Supervised Learning

[3] http://www.newscientist.com/article/
dn21933-baby-robot-learns-first-words-from-human-teacher.
html

In Supervised learning, the neural network is trained using example inputs and the correct output. The network
can then work out the relationship between the input and
output. For example, a network could be trained by showing it details about houses and the sale price. Once it has
nished training it could estimate the sale price of another
house using the details like the number of bedrooms and
local crime rate.
Another example is the ALV (Autonomous Land Vehicle). DARPA funded this project in the 1980s. In a
demonstration in 1987 it travelled 600 metres at 3 km/h
over dicult land, with sharp rocks, vegetation and steep
ravines. This vehicle could drive itself as fast as 30 km/h.
This network watched a 'teacher' drive, and saw the road
using laser radar. The learning process was repeated for
dierent road types. ALV used a kind of neural network
called a multi-layer perceptron.[1]
1

3 TEXT AND IMAGE SOURCES, CONTRIBUTORS, AND LICENSES

Text and image sources, contributors, and licenses

3.1

Text

Neural network Source: http://simple.wikipedia.org/wiki/Neural%20network?oldid=4905983 Contributors: Eptalon, JAnDbot, SieBot,

DragonBot, BodhisattvaBot, PeterSymonds, Luckas-bot, Nallimbot, Chris G Bot, TinucherianBot II, Xqbot, GoblinBot4, FoxBot,
PiRSquared17Bot1, MerlIwBot, Auntof6Bot, Osiris, Addbot, Carriearchdale, YuyaJeremyOng and Anonymous: 9

3.2

Images

File:Ambox_content.png Source: http://upload.wikimedia.org/wikipedia/commons/f/f4/Ambox_content.png License: Public domain

Contributors: Derived from Image:Information icon.svg Original artist: El T (original icon); David Levy (modied design); Penubag (modied color)

3.3

Content license

Creative Commons Attribution-Share Alike 3.0