<s>
The	O
wake-sleep	B-Algorithm
algorithm	I-Algorithm
is	O
an	O
unsupervised	B-General_Concept
learning	I-General_Concept
algorithm	O
for	O
a	O
stochastic	O
multilayer	O
neural	B-Architecture
network	I-Architecture
.	O
</s>
<s>
It	O
can	O
be	O
viewed	O
as	O
a	O
way	O
to	O
train	O
a	O
Helmholtz	B-Algorithm
Machine	I-Algorithm
.	O
</s>
<s>
It	O
can	O
also	O
be	O
used	O
in	O
Deep	B-Algorithm
Belief	I-Algorithm
Networks	I-Algorithm
(	O
DBN	O
)	O
.	O
</s>
<s>
The	O
wake-sleep	B-Algorithm
algorithm	I-Algorithm
is	O
visualized	O
as	O
a	O
stack	O
of	O
layers	O
containing	O
representations	O
of	O
data	O
.	O
</s>
<s>
The	O
wake-sleep	B-Algorithm
algorithm	I-Algorithm
is	O
convergent	B-Algorithm
and	O
can	O
be	O
stochastic	O
if	O
alternated	O
appropriately	O
.	O
</s>
<s>
The	O
wake-sleep	B-Algorithm
algorithm	I-Algorithm
has	O
been	O
suggested	O
not	O
to	O
be	O
powerful	O
enough	O
for	O
the	O
layers	O
of	O
the	O
inference	O
network	O
in	O
order	O
to	O
recover	O
a	O
good	O
estimator	O
of	O
the	O
posterior	O
distribution	O
of	O
latent	O
variables	O
.	O
</s>
