<s>
Automated	B-General_Concept
machine	I-General_Concept
learning	I-General_Concept
(	O
AutoML	B-General_Concept
)	O
is	O
the	O
process	O
of	O
automating	B-Application
the	O
tasks	O
of	O
applying	O
machine	O
learning	O
to	O
real-world	O
problems	O
.	O
</s>
<s>
AutoML	B-General_Concept
potentially	O
includes	O
every	O
stage	O
from	O
beginning	O
with	O
a	O
raw	O
dataset	O
to	O
building	O
a	O
machine	O
learning	O
model	O
ready	O
for	O
deployment	O
.	O
</s>
<s>
AutoML	B-General_Concept
was	O
proposed	O
as	O
an	O
artificial	O
intelligence-based	O
solution	O
to	O
the	O
growing	O
challenge	O
of	O
applying	O
machine	O
learning	O
.	O
</s>
<s>
The	O
high	O
degree	O
of	O
automation	B-Application
in	O
AutoML	B-General_Concept
aims	O
to	O
allow	O
non-experts	O
to	O
make	O
use	O
of	O
machine	O
learning	O
models	O
and	O
techniques	O
without	O
requiring	O
them	O
to	O
become	O
experts	O
in	O
machine	O
learning	O
.	O
</s>
<s>
Automating	B-Application
the	O
process	O
of	O
applying	O
machine	O
learning	O
end-to-end	O
additionally	O
offers	O
the	O
advantages	O
of	O
producing	O
simpler	O
solutions	O
,	O
faster	O
creation	O
of	O
those	O
solutions	O
,	O
and	O
models	O
that	O
often	O
outperform	O
hand-designed	O
models	O
.	O
</s>
<s>
Common	O
techniques	O
used	O
in	O
AutoML	B-General_Concept
include	O
hyperparameter	B-General_Concept
optimization	I-General_Concept
,	O
meta-learning	B-General_Concept
and	O
neural	B-General_Concept
architecture	I-General_Concept
search	I-General_Concept
.	O
</s>
<s>
To	O
make	O
the	O
data	O
amenable	O
for	O
machine	O
learning	O
,	O
an	O
expert	O
may	O
have	O
to	O
apply	O
appropriate	O
data	B-General_Concept
pre-processing	I-General_Concept
,	O
feature	B-General_Concept
engineering	I-General_Concept
,	O
feature	B-Algorithm
extraction	I-Algorithm
,	O
and	O
feature	B-General_Concept
selection	I-General_Concept
methods	O
.	O
</s>
<s>
After	O
these	O
steps	O
,	O
practitioners	O
must	O
then	O
perform	O
algorithm	B-Application
selection	I-Application
and	O
hyperparameter	B-General_Concept
optimization	I-General_Concept
to	O
maximize	O
the	O
predictive	O
performance	O
of	O
their	O
model	O
.	O
</s>
<s>
AutoML	B-General_Concept
automatically	O
streamlines	O
the	O
whole	O
machine	O
learning	O
process	O
from	O
data	O
loading	O
,	O
modelling	O
and	O
model	O
picking	O
.	O
</s>
<s>
AutoML	B-General_Concept
aims	O
to	O
simplify	O
these	O
steps	O
for	O
non-experts	O
,	O
and	O
to	O
make	O
it	O
easier	O
for	O
them	O
to	O
use	O
machine	O
learning	O
techniques	O
correctly	O
and	O
effectively	O
.	O
</s>
<s>
AutoML	B-General_Concept
plays	O
an	O
important	O
role	O
within	O
the	O
broader	O
approach	O
of	O
automating	B-Application
data	O
science	O
,	O
which	O
also	O
includes	O
challenging	O
tasks	O
such	O
as	O
data	O
engineering	O
,	O
data	O
exploration	O
and	O
model	O
interpretation	O
.	O
</s>
<s>
Automated	B-General_Concept
machine	I-General_Concept
learning	I-General_Concept
can	O
target	O
various	O
stages	O
of	O
the	O
machine	O
learning	O
process	O
.	O
</s>
