1.1. Why is it worth to learn about neural networks?

rumblecleverΤεχνίτη Νοημοσύνη και Ρομποτική

1 Δεκ 2013 (πριν από 3 χρόνια και 8 μήνες)

87 εμφανίσεις

1.1. Why is it worth to learn about neural networks?


(translation by Paweł Olaszek; pawelolaszek@tlen.pl)

This book is for you, dear reader, to learn in a nice and easy way what neural networks are,
how and why they work and how and where to use them. If

you hold this book in your hands,
I believe that it is what you want to find out. This book is thick so it needs a lot of effort to
read it
. You can ask yourself a question: Is it worthy? If yes, then why? Maybe it is better to
put it away and play some c
omputer games?

The simplest answer is:
Yes, It is worthy
to study neural networks because they currently are
in interest of many researchers and practitioners. With their help, people made many
interesting discoveries and for sure they will lead to further

achievements.

If those reasons still did not convince you need to spend some time studying neural networks
then you can recognize them worth knowing, because you can hear a lot about them and we
are observing a trend for them.

However are those reasons en
ough to convince you?

After all, we are observing various trends in computer science for years and tied with their
“comes and goes” of interest in specific problems influencing work of scientists, shaping
computer market and focusing effort of software
developers on specific tasks. A lifetime of a
trend varies from couple of months to few years. Usually a trend ends when something new
comes up and takes over all passionate of the previous one. We can recall many big trends
but some are really worth rem
inding: recently we went thru massive internet fascination (we
still witness faze of intelligent web browsers). Now, very popular is grid computing. We still
witness wave of popularity for cellular automaton and agent technique. Also fractals and
chaos hav
e their faithful fans. From time to time comes back (like a Bubonic plague)
fascination of genetic algorithms and a theory of fuzzy sets.

We can say that in computer science changeability of trends is very trendy.

From the beginning of 90’s neural networks

started becoming popular and they are still today.
A little more about this boom I wrote in preface therefore if you haven’t read it yet (I know
some people whose principle is to omit prefaces) I advice you to step back a few pages and
read
this

one. It c
ontains many interesting and important information that I am not going to
repeat further in this book. You will find them useful if you are planning to become interested
in neural networks and even use them as a useful tool in your work.

Now I am going to

tell you something important about this chapter.

Well if you are not interested in biological basics of neural networks then you can skip
this chapter.

Really!

In chapters ached I am going to show you how to build artificial neural networks and how you
c
an use them. Those chapters you
must
read one after the other because if you omit
something you will have problems understanding further solutions presented in this book.

But this one is different. It says the story how mankind discovered neural networks e
xamining
its own brain. Those examinations were carried out for years to find out secrets of human
intelligence when suddenly research findings turned out to be useful in computer science.
This chapter will tell how those borrowed from biologists artificia
l neural networks help today
discover another secrets of human brain. I find this incredibly interesting therefore I wrote so
much on this subject. If you are eager to explore secrets of neural networks as fast as it is
possible you really can skip this wh
ole chapter. I hope that when you see how cool do they
work you will come back here to learn their genesis but remember you
do not have to do it.

If you are still reading this means that you are truly interested how did it happen to discover
neural network
s and I will try to satisfy your curiosity. As you already know, neural networks
are simplified (therefore easier to understand and use in computer software) but surprisingly
complex and interesting model of biological nervous system. Shortly we could say
that neural
networks are simplified model of some functions of our own brain (Fig. 1).



Fig. 1.1. Human brain


source of inspiration for neural networks researchers