<s>
Gated	B-Algorithm
recurrent	I-Algorithm
units	I-Algorithm
(	O
GRUs	O
)	O
are	O
a	O
gating	O
mechanism	O
in	O
recurrent	B-Algorithm
neural	I-Algorithm
networks	I-Algorithm
,	O
introduced	O
in	O
2014	O
by	O
Kyunghyun	O
Cho	O
et	O
al	O
.	O
</s>
<s>
The	O
GRU	O
is	O
like	O
a	O
long	B-Algorithm
short-term	I-Algorithm
memory	I-Algorithm
(	O
LSTM	B-Algorithm
)	O
with	O
a	O
forget	O
gate	O
,	O
but	O
has	O
fewer	O
parameters	O
than	O
LSTM	B-Algorithm
,	O
as	O
it	O
lacks	O
an	O
output	O
gate	O
.	O
</s>
<s>
GRU	O
's	O
performance	O
on	O
certain	O
tasks	O
of	O
polyphonic	O
music	O
modeling	O
,	O
speech	O
signal	O
modeling	O
and	O
natural	O
language	O
processing	O
was	O
found	O
to	O
be	O
similar	O
to	O
that	O
of	O
LSTM	B-Algorithm
.	O
</s>
<s>
:	O
The	O
original	O
is	O
a	O
sigmoid	B-Algorithm
function	I-Algorithm
.	O
</s>
<s>
Alternative	O
activation	B-Algorithm
functions	I-Algorithm
are	O
possible	O
,	O
provided	O
that	O
.	O
</s>
<s>
The	O
framework	O
has	O
the	O
advantage	O
of	O
having	O
been	O
generated	O
from	O
an	O
extensive	O
analysis	O
of	O
the	O
literature	O
and	O
dedicated	O
to	O
recurrent	B-Algorithm
neural	I-Algorithm
networks	I-Algorithm
and	O
their	O
variations	O
.	O
</s>
