<s>
In	O
machine	O
learning	O
,	O
a	O
learning	B-General_Concept
curve	I-General_Concept
(	O
or	O
training	O
curve	O
)	O
plots	B-Application
the	O
optimal	O
value	O
of	O
a	O
model	O
's	O
loss	O
function	O
for	O
a	O
training	O
set	O
against	O
this	O
loss	O
function	O
evaluated	O
on	O
a	O
validation	O
data	O
set	O
with	O
same	O
parameters	O
as	O
produced	O
the	O
optimal	O
function	O
.	O
</s>
<s>
The	O
machine	O
learning	B-General_Concept
curve	I-General_Concept
is	O
useful	O
for	O
many	O
purposes	O
including	O
comparing	O
different	O
algorithms	O
,	O
choosing	O
model	O
parameters	O
during	O
design	O
,	O
adjusting	O
optimization	O
to	O
improve	O
convergence	B-Algorithm
,	O
and	O
determining	O
the	O
amount	O
of	O
data	O
used	O
for	O
training	O
.	O
</s>
<s>
In	O
the	O
machine	O
learning	O
domain	O
,	O
there	O
are	O
two	O
implications	O
of	O
learning	B-General_Concept
curves	I-General_Concept
differing	O
in	O
the	O
x-axis	O
of	O
the	O
curves	O
,	O
with	O
experience	O
of	O
the	O
model	O
graphed	O
either	O
as	O
the	O
number	O
of	O
training	O
examples	O
used	O
for	O
learning	O
or	O
the	O
number	O
of	O
iterations	O
used	O
in	O
training	O
the	O
model	O
.	O
</s>
<s>
Many	O
optimization	O
processes	O
are	O
iterative	O
,	O
repeating	O
the	O
same	O
step	O
until	O
the	O
process	O
converges	B-Algorithm
to	O
an	O
optimal	O
value	O
.	O
</s>
<s>
Gradient	B-Algorithm
descent	I-Algorithm
is	O
one	O
such	O
algorithm	O
.	O
</s>
