<s>
Fast	B-Library
Artificial	I-Library
Neural	I-Library
Network	I-Library
(	O
FANN	O
)	O
is	O
cross-platform	B-Operating_System
open-source	B-Application
programming	B-Library
library	I-Library
for	O
developing	O
multilayer	O
feedforward	B-Algorithm
Artificial	B-Architecture
Neural	I-Architecture
Networks	I-Architecture
.	O
</s>
<s>
FAN	O
supports	O
cross-platform	B-Operating_System
execution	O
of	O
single	O
and	O
multilayer	O
networks	O
.	O
</s>
<s>
It	O
also	O
supports	O
fixed	O
point	O
and	O
floating	B-Algorithm
point	I-Algorithm
arithmetic	I-Algorithm
.	O
</s>
<s>
It	O
includes	O
functions	O
that	O
simplify	O
the	O
creating	O
,	O
training	O
and	O
testing	O
of	O
neural	B-Architecture
networks	I-Architecture
.	O
</s>
<s>
It	O
has	O
bindings	B-Application
for	O
over	O
20	O
programming	O
languages	O
,	O
including	O
commonly	O
used	O
languages	O
such	O
as	O
PHP	B-Application
,	O
C#	B-Application
and	O
python	B-Language
.	O
</s>
<s>
On	O
the	O
FANN	O
website	O
multiple	O
graphical	B-Application
user	I-Application
interfaces	I-Application
are	O
available	O
for	O
use	O
with	O
the	O
library	O
such	O
as	O
FANNTool	O
,	O
Agiel	O
Neural	B-Architecture
Network	I-Architecture
,	O
Neural	O
View	O
,	O
FannExeplorer	O
,	O
and	O
others	O
.	O
</s>
<s>
These	O
graphical	B-Application
interface	I-Application
facilitate	O
the	O
use	O
of	O
FANN	O
for	O
users	O
that	O
are	O
not	O
very	O
familiar	O
with	O
programming	O
or	O
for	O
users	O
who	O
are	O
seeking	O
for	O
a	O
simple	O
out-of-the	O
box	O
solution	O
.	O
</s>
<s>
Training	O
for	O
FANN	O
is	O
carried	O
out	O
through	O
backpropagation	B-Algorithm
.	O
</s>
<s>
Trained	O
Artificial	B-Architecture
Neural	I-Architecture
Networks	I-Architecture
can	O
be	O
stored	O
as	O
files	O
to	O
quickly	O
saved	O
and	O
load	O
ANNs	O
for	O
future	O
use	O
or	O
future	O
training	O
.	O
</s>
<s>
This	O
allows	O
the	O
user	O
to	O
partition	O
the	O
training	O
in	O
multiple	O
steps	O
which	O
can	O
be	O
useful	O
when	O
dealing	O
with	O
large	O
training	O
datasets	O
or	O
sizable	O
neural	B-Architecture
networks	I-Architecture
.	O
</s>
<s>
Its	O
original	O
implementation	O
is	O
described	O
in	O
Nissen	O
's	O
2003	O
report	O
Implementation	O
of	O
a	O
Fast	B-Library
Artificial	I-Library
Neural	I-Library
Network	I-Library
Library	O
(	O
FANN	O
)	O
.	O
</s>
<s>
This	O
report	O
was	O
submitted	O
to	O
the	O
computer	B-General_Concept
science	I-General_Concept
department	O
at	O
the	O
University	O
of	O
Copenhagen	O
(	O
DIKU	O
)	O
.	O
</s>
<s>
In	O
his	O
original	O
report	O
Nissen	O
stated	O
that	O
one	O
of	O
his	O
primary	O
motivations	O
in	O
writing	O
FANN	O
was	O
to	O
develop	O
a	O
neural	B-Architecture
network	I-Architecture
library	O
that	O
was	O
friendly	O
to	O
both	O
fixed	O
point	O
and	O
floating	B-Algorithm
point	I-Algorithm
arithmetic	I-Algorithm
.	O
</s>
<s>
Nissen	O
wanted	O
to	O
develop	O
an	O
autonomous	B-General_Concept
agent	I-General_Concept
that	O
can	O
learn	O
from	O
experience	O
.	O
</s>
<s>
His	O
goal	O
was	O
to	O
use	O
this	O
autonomous	B-General_Concept
agent	I-General_Concept
to	O
create	O
a	O
virtual	O
player	O
in	O
Quake	B-Application
III	I-Application
Arena	I-Application
that	O
can	O
learn	O
from	O
gameplay	O
.	O
</s>
<s>
Since	O
its	O
original	O
1.0.0	O
version	O
release	O
,	O
the	O
library	O
's	O
functionality	O
has	O
been	O
expanded	O
by	O
the	O
creator	O
and	O
its	O
many	O
contributors	O
to	O
include	O
more	O
practical	O
constructors	O
,	O
different	O
activation	B-Algorithm
functions	I-Algorithm
,	O
simpler	O
access	O
to	O
parameters	O
and	O
bindings	B-Application
to	O
multiple	O
programming	O
languages	O
.	O
</s>
<s>
It	O
has	O
been	O
downloaded	O
450,000	O
times	O
since	O
its	O
move	O
to	O
Source	B-Application
Forge	I-Application
in	O
2003	O
and	O
29,000	O
times	O
in	O
2016	O
alone	O
.	O
</s>
<s>
The	O
source	O
code	O
is	O
now	O
hosted	O
on	O
GitHub	B-Application
.	O
</s>
<s>
The	O
original	O
FANN	O
report	O
written	O
by	O
Steffen	O
Nissen	O
has	O
been	O
cited	O
526	O
times	O
per	O
Google	B-Library
Scholar	I-Library
.	O
</s>
<s>
The	O
library	O
has	O
been	O
used	O
for	O
research	O
in	O
image	O
recognition	O
,	O
machine	O
learning	O
,	O
biology	O
,	O
genetics	O
,	O
aerospace	O
engineering	O
,	O
environmental	O
sciences	O
and	O
artificial	B-Application
intelligence	I-Application
.	I-Application
</s>
<s>
While	O
FANN	O
was	O
originally	O
written	O
in	O
C	B-Language
,	O
the	O
following	O
language	B-Application
bindings	I-Application
have	O
been	O
created	O
by	O
FANN	O
contributors	O
:	O
</s>
