I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
p
u
t
er
Science
Vo
l.
21
,
No
.
2
,
Feb
r
u
ar
y
2
0
2
1
,
p
p
.
1
0
5
7
~1
0
64
I
SS
N:
2
5
02
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/i
j
ee
cs.v
2
1
.i
2
.
p
p
1
0
57
-
10
64
1057
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
Dia
lo
g
ue state tra
ck
i
ng
accura
cy
i
m
pro
v
e
m
en
t
by
disting
uishi
ng
slo
t
-
v
a
lue pairs
and
dia
lo
g
ue behav
io
ur
K
ha
ldo
o
n H
.
Alhu
s
s
a
y
ni
1
,
A
lex
a
nd
er
Z
a
m
y
a
t
in
2
,
S.
E
m
a
n Alsh
a
m
er
y
3
1
Co
m
p
u
ter
Ce
n
ter,
Ba
b
y
lo
n
Un
iv
e
rsit
y
,
Ba
b
y
lo
n
,
Ira
q
.
2
Co
m
p
u
terScie
n
c
e
De
p
a
rtme
n
t,
T
o
m
s
k
S
tate
Un
iv
e
rsit
y
,
T
o
m
s
k
,
Ru
ss
ia
3
S
o
f
tw
a
r
e
De
p
a
rt
m
e
n
t,
Ba
b
y
lo
n
Un
iv
e
rsit
y
,
Ba
b
y
lo
n
,
Ira
q
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
A
p
r
2
3
,
2
0
2
0
R
ev
i
s
ed
J
u
n
1
5
,
2
0
2
0
A
cc
ep
ted
J
u
l 8
,
2
0
2
0
Dia
lo
g
sta
te
trac
k
in
g
(DST
)
p
la
y
s
a
c
rit
ica
l
ro
le
in
c
y
c
le
li
f
e
o
f
a
tas
k
-
o
rien
ted
d
ialo
g
u
e
sy
ste
m
.
DS
T
r
e
p
re
se
n
ts
th
e
g
o
a
ls
o
f
th
e
c
o
n
su
m
e
r
a
t
e
a
c
h
ste
p
b
y
d
ialo
g
u
e
a
n
d
d
e
sc
rib
e
s
su
c
h
o
b
jec
ti
v
e
s
a
s
a
c
o
n
c
e
p
tu
a
l
stru
c
tu
re
c
o
m
p
risin
g
slo
t
-
v
a
lu
e
p
a
irs
a
n
d
d
ialo
g
u
e
a
c
ti
o
n
s th
a
t
sp
e
c
if
ica
ll
y
i
m
p
ro
v
e
th
e
p
e
rf
o
r
m
a
n
c
e
a
n
d
e
f
fe
c
ti
v
e
n
e
ss
o
f
d
ialo
g
u
e
sy
ste
m
s.
DS
T
f
a
c
e
s
se
v
e
ra
l
c
h
a
ll
e
n
g
e
s:
d
iv
e
rsit
y
o
f
li
n
g
u
isti
c
s,
d
y
n
a
m
ic
so
c
ial
c
o
n
tex
t
a
n
d
th
e
d
isse
m
in
a
ti
o
n
o
f
th
e
sta
te
o
f
d
i
a
lo
g
u
e
o
v
e
r
c
a
n
d
i
d
a
te
v
a
lu
e
s
b
o
th
in
slo
t
v
a
lu
e
s
a
n
d
in
d
ialo
g
u
e
a
c
ts
d
e
term
in
e
d
in
o
n
to
l
o
g
y
.
In
m
a
n
y
tu
rn
s
d
u
ri
n
g
th
e
d
ialo
g
u
e
,
u
se
rs
in
d
irec
tl
y
r
e
f
e
r
to
th
e
p
re
v
io
u
s
u
tt
e
ra
n
c
e
s,
a
n
d
th
a
t
p
ro
d
u
c
e
a
c
h
a
ll
e
n
g
e
to
d
isti
n
g
u
is
h
in
g
a
n
d
u
se
o
f
re
late
d
d
ialo
g
u
e
h
isto
r
y
,
Re
c
e
n
t
met
h
o
d
s u
se
d
a
n
d
p
o
p
u
lar
f
o
r
th
a
t
a
re
in
e
ff
e
c
ti
v
e
.
In
th
is
p
a
p
e
r,
w
e
p
ro
p
o
se
a
d
ialo
g
u
e
h
isto
rica
l
c
o
n
tex
t
se
lf
-
Atten
ti
o
n
f
ra
m
e
w
o
rk
f
o
r
DST
th
a
t
re
c
o
g
n
ize
s
re
lev
a
n
t
h
isto
rica
l
c
o
n
tex
t
b
y
in
c
l
u
d
i
n
g
p
re
v
io
u
s
u
se
r
u
tt
e
ra
n
c
e
b
e
s
id
e
c
u
rre
n
t
u
se
r
u
tt
e
ra
n
c
e
s
a
n
d
p
re
v
io
u
s
sy
s
tem
a
c
ti
o
n
s
w
h
e
re
sp
e
c
if
ic
slo
t
-
v
a
lu
e
p
iers
v
a
riatio
n
s
a
n
d
u
se
s
t
h
a
t
t
o
g
e
th
e
r
w
it
h
w
e
ig
h
ted
sy
st
e
m
u
tt
e
ra
n
c
e
to
o
u
t
p
e
rf
o
rm
e
x
isti
n
g
m
o
d
e
ls
b
y
re
c
o
g
n
izin
g
th
e
re
late
d
c
o
n
tex
t
a
n
d
th
e
re
lev
a
n
c
e
o
f
a
s
y
ste
m
u
tt
e
ra
n
c
e
.
F
o
r
th
e
e
v
a
lu
a
ti
o
n
o
f
th
e
p
ro
p
o
s
e
d
m
o
d
e
l
th
e
W
o
Z
d
a
tas
e
t
w
a
s
u
se
d
.
T
h
e
imp
lem
e
n
tatio
n
w
a
s
a
tt
e
m
p
ted
w
it
h
t
h
e
p
ri
o
r
u
se
r
u
tt
e
ra
n
c
e
a
s
a
d
ialo
g
u
e
e
n
c
o
d
e
r
a
n
d
se
c
o
n
d
b
y
th
e
a
d
d
it
i
o
n
a
l
sc
o
re
c
o
m
b
in
e
d
w
it
h
a
ll
th
e
c
a
n
d
i
d
a
te
slo
t
-
v
a
lu
e
p
a
irs
i
n
t
h
e
c
o
n
tex
t
o
f
p
re
v
io
u
s
u
se
r
u
tt
e
ra
n
c
e
s
a
n
d
c
u
rre
n
t
u
tt
e
ra
n
c
e
s.
T
h
e
p
ro
p
o
se
d
m
o
d
e
l
o
b
tain
e
d
0
.
8
p
e
r
c
e
n
t
b
e
tt
e
r
re
su
l
ts
th
a
n
a
ll
sta
te
-
of
-
th
e
-
a
rt
m
e
th
o
d
s
i
n
t
h
e
c
o
m
b
in
e
d
p
re
c
isio
n
o
f
th
e
targ
e
t,
b
u
t
th
is
is
n
o
t
t
h
e
tu
r
n
a
ro
u
n
d
c
h
a
ll
e
n
g
e
f
o
r
th
e
su
b
m
is
sio
n
.
K
ey
w
o
r
d
s
:
Dialo
g
s
tate
tr
ac
k
i
n
g
E
n
co
d
er
-
d
ec
o
d
er
Nau
tr
al
la
n
g
u
a
g
e
p
r
o
ce
s
s
in
g
T
ask
-
o
r
ien
ted
d
ialo
g
u
e
s
y
s
te
m
T
h
is
is
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Kh
ald
o
o
n
H.
A
lh
u
s
s
a
y
n
i
C
o
m
p
u
ter
C
en
ter
B
ab
y
lo
n
U
n
i
v
er
s
it
y
,
B
ab
y
lo
n
,
I
r
aq
E
m
ail: a
l
h
u
s
s
a
y
n
i8
3
@
g
m
a
il.c
o
m
1.
I
NT
RO
D
UCT
I
O
N
Dialo
g
s
tate
tr
ac
k
in
g
(
DST
)
is
a
cr
u
c
ial
p
ar
t
o
f
ta
s
k
-
o
r
i
en
ted
d
ialo
g
u
e
s
y
s
te
m
s
t
h
at
tr
ac
k
u
s
er
o
b
j
ec
tiv
es
at
ea
ch
tu
r
n
b
ased
o
n
d
ialo
g
u
e
h
i
s
to
r
y
.
T
ask
-
o
r
ie
n
ted
d
ialo
g
u
e
s
y
s
te
m
s
co
m
m
u
n
icate
w
it
h
n
at
u
r
al
-
lan
g
u
a
g
e
u
s
er
s
b
y
ta
lk
i
n
g
to
a
v
o
ice
i
n
ter
f
ac
e
o
r
w
r
i
tin
g
t
ex
t
to
p
er
f
o
r
m
tas
k
s
t
h
e
y
h
a
v
e
in
m
i
n
d
.
Mo
d
er
n
DST
m
et
h
o
d
s
r
el
y
o
n
m
o
d
els
f
o
r
d
ee
p
lear
n
in
g
.
P
r
o
p
o
s
als
th
at
ar
e
m
o
r
e
r
ec
en
t
h
av
e
b
ee
n
m
ad
e
f
o
r
s
e
v
er
al
n
eu
r
al
-
b
ased
D
ST
s
tr
u
ct
u
r
es.
I
n
ad
d
itio
n
to
h
an
d
cr
af
t
s
f
ea
t
u
r
es,
w
h
er
e
a
r
ep
r
esen
tatio
n
b
ased
o
n
p
r
e
-
tr
ain
ed
w
o
r
d
in
cr
u
s
tatio
n
s
is
co
m
p
u
ted
to
ag
r
ee
o
n
s
u
c
h
v
ec
to
r
s
,
th
e
n
e
u
r
al
f
aith
tr
ac
k
er
(
NB
T
)
m
o
d
el
p
r
o
p
o
s
ed
to
u
s
e
d
ep
th
lear
n
in
g
(
DNN)
an
d
co
n
v
o
lu
tio
n
ar
y
n
et
w
o
r
k
(
C
N
N)
to
ca
lcu
late
t
h
ese
r
ep
r
esen
tatio
n
v
ec
to
r
s
in
o
r
d
er
to
lear
n
f
ea
tu
r
e
ap
p
r
o
p
r
iate
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
21
,
No
.
2
,
Feb
r
u
ar
y
2
0
2
1
:
10
57
-
10
64
1058
f
o
r
ea
ch
s
tate
b
y
[
1
]
.
Nev
er
t
h
eless
,
m
u
lti
-
v
alu
ed
s
lo
ts
d
o
n
o
t
tak
e
i
n
to
ac
co
u
n
t
all
v
al
u
es
,
alth
o
u
g
h
t
h
i
s
w
o
r
k
m
a
y
p
r
ed
ict
m
u
lt
ip
le
p
o
ten
tial
v
alu
e
s
[
2
]
.
Z
h
o
n
g
et
al.
(
2
0
1
8
)
p
r
o
p
o
s
ed
th
e
GL
A
D
m
o
d
el
w
it
h
s
e
lf
-
atte
n
tio
n
-
b
ased
r
ec
u
r
r
en
t
n
et
w
o
r
k
s
an
d
s
i
m
u
lated
r
ep
r
esen
tatio
n
s
f
o
r
ea
ch
u
ter
u
s
a
n
d
p
r
ev
io
u
s
s
y
s
te
m
b
eh
a
v
io
u
r
,
ca
lcu
lati
n
g
a
s
i
m
ilar
it
y
to
ea
ch
s
lo
t
-
v
a
lu
e
[
3
]
.
T
h
is
h
as
i
m
p
r
o
v
ed
GL
A
D
ar
ch
itect
u
r
e
w
it
h
th
e
el
i
m
in
at
io
n
o
f
s
lo
t
-
d
ep
en
d
en
t
r
ec
u
r
r
en
t
u
tter
an
ce
n
e
t
w
o
r
k
s
a
n
d
d
ev
ice
i
n
t
er
v
en
tio
n
en
co
d
er
an
d
t
h
e
u
s
e
o
f
an
ap
p
licatio
n
v
ec
to
r
g
lo
b
all
y
co
n
d
itio
n
ed
en
co
d
er
(
GC
E
)
.
I
n
th
e
last
t
w
o
ap
p
r
o
ac
h
es,
h
o
w
e
v
er
,
d
is
tr
ib
u
to
r
s
in
th
e
p
r
o
d
u
ctio
n
s
y
s
te
m
w
er
e
s
ti
ll
i
n
ef
f
ec
ti
v
e
d
u
e
to
th
eir
f
a
ilu
r
e
to
r
ec
o
g
n
is
e
a
n
d
ta
k
e
th
e
ap
p
licab
le
co
n
tex
t
o
n
b
o
ar
d
,
w
h
ile
t
h
e
w
o
r
k
co
u
ld
id
en
ti
f
y
t
h
e
as
s
o
ciate
d
co
n
te
x
t.
T
h
e
p
u
r
p
o
s
e
o
f
DST
is
t
o
p
r
ed
ict
th
e
s
et
o
f
o
b
j
ec
tiv
es
t
h
at
co
r
r
esp
o
n
d
to
o
n
to
lo
g
ical
p
r
o
d
u
cts
in
u
s
er
s
t
ate
m
e
n
ts
th
a
t
ar
e
r
ep
r
esen
ted
as
s
lo
t
-
v
alu
e
p
air
s
.
T
h
is
w
il
l
b
ec
o
m
e
a
co
m
p
licated
task
w
h
e
n
t
h
e
lex
ica
l
d
is
cr
ep
an
c
y
is
c
h
allen
g
ed
d
y
n
a
m
ics
o
f
th
e
c
o
n
te
x
t
p
r
o
d
u
ctio
n
,
th
e
d
is
tr
ib
u
tio
n
o
f
th
e
s
tate
o
f
t
h
e
d
ialo
g
u
e,
to
tal
p
air
s
o
f
ca
n
d
id
ates
an
d
d
ialo
g
u
e
ac
tio
n
s
d
escr
ib
ed
in
co
n
ce
p
tu
aliza
tio
n
.
W
e
p
r
esen
t
h
er
e
an
i
m
p
r
o
v
e
m
en
t
m
o
d
el
u
s
ed
b
y
[
3
]
,
u
s
in
g
th
e
p
r
io
r
v
alu
es
in
ea
ch
s
p
i
n
,
w
h
ich
f
o
r
m
th
e
s
tat
e.
T
o
th
is
e
n
d
,
th
e
c
u
r
r
en
t
a
n
d
p
r
ev
io
u
s
u
s
er
s
tate
m
en
t
s
h
av
e
b
ee
n
m
er
g
ed
an
d
v
ar
iatio
n
s
b
et
w
ee
n
t
h
e
m
a
n
d
o
n
to
lo
g
y
h
a
v
e
b
ee
n
d
eter
m
i
n
ed
.
2.
RE
L
AT
E
D
WO
RK
T
h
e
m
e
t
h
o
d
s
o
f
d
ialo
g
u
e
s
tate
m
o
n
ito
r
in
g
ca
n
b
e
d
iv
id
ed
in
to
a
r
u
les
-
b
ased
ap
p
r
o
ac
h
,
s
tatis
tics
a
n
d
a
d
ee
p
lea
r
n
in
g
s
y
s
te
m
.
T
h
e
u
s
e
r
u
le
-
b
ased
h
eu
r
i
s
tic
s
an
d
ca
lcu
late
th
e
co
n
f
id
en
ce
s
c
o
r
es
o
f
th
e
N
-
b
est
ca
n
d
id
ates
p
r
o
d
u
ce
d
to
d
eter
m
i
n
e
t
h
e
co
r
r
ec
t
d
ialo
g
u
e
s
tat
es
f
r
o
m
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
t
h
e
n
atu
r
al
lan
g
u
ag
e
co
m
p
r
e
h
en
s
io
n
m
o
d
u
le
[4
-
7]
.
T
h
e
p
u
r
p
o
s
e
is
m
o
n
ito
r
th
e
d
etails
n
ee
d
ed
to
tr
ac
k
th
e
s
tat
u
s
o
f
t
h
e
d
ialo
g
u
e.
Nev
er
th
e
less
,
th
e
s
e
p
r
in
cip
le
s
ar
e
n
o
t
co
llected
a
u
to
m
at
i
ca
ll
y
f
r
o
m
ac
t
u
al
i
n
f
o
r
m
atio
n
o
n
a
d
ialo
g
u
e
to
in
v
o
l
v
e
ca
r
e
f
u
l
tu
n
i
n
g
an
d
r
e
s
p
o
n
s
i
v
e
d
esi
g
n
atte
m
p
t
s
.
D
u
e
to
th
ese
tech
n
iq
u
e
s
,
t
h
e
d
ef
in
itio
n
o
f
d
ialo
g
u
e
s
tates o
f
ten
lead
s
to
i
n
ac
cu
r
ac
y
.
Statis
t
ical
m
et
h
o
d
s
f
o
r
d
ialo
g
u
e
b
et
w
ee
n
t
h
e
Sta
te
T
r
ac
k
in
g
[8
-
10]
h
a
v
e
b
ee
n
u
s
ed
t
o
p
r
o
v
id
e
alter
n
ati
v
es
to
cr
af
ted
r
u
les.
S
tatis
tical
ap
p
r
o
ac
h
es
s
u
ch
as
l
o
g
is
tic
r
eg
r
es
s
io
n
a
n
d
th
e
B
ay
esia
n
n
et
w
o
r
k
f
o
r
th
e
ac
co
m
p
lis
h
m
en
t
o
f
h
i
g
h
-
p
er
f
o
r
m
an
ce
m
o
n
ito
r
i
n
g
a
n
d
th
e
i
m
p
le
m
e
n
tatio
n
o
f
a
co
n
f
id
en
ce
ev
al
u
atio
n
o
f
u
s
er
i
n
f
o
r
m
atio
n
.
T
h
ese
r
ep
o
r
ts
,
h
o
w
e
v
er
,
h
a
v
e
a
co
m
m
o
n
p
r
o
b
lem
t
h
at
e
v
er
y
co
n
v
er
s
atio
n
,
r
ath
er
e
x
p
en
s
i
v
e
co
m
p
u
tat
io
n
,
s
h
o
u
ld
b
e
in
clu
d
ed
in
ev
er
y
.
R
ec
en
tl
y
,
th
e
u
s
e
o
f
d
ee
p
lear
n
in
g
tech
n
iq
u
e
s
f
o
r
d
ialo
g
u
e
tr
ac
k
in
g
th
e
s
tate
[
1
,
1
1
-
19]
.
I
n
ad
d
iti
o
n
,
o
th
er
s
ar
e
ab
le
to
u
n
d
er
s
t
an
d
s
p
ec
i
f
ic
u
s
er
an
d
s
y
s
te
m
u
tter
an
ce
s
a
n
d
p
r
io
r
s
y
s
te
m
ac
t
io
n
s
to
p
r
ed
ict
tu
r
n
a
r
o
u
n
d
.
First
u
s
ed
f
o
r
s
tate
tr
ac
k
in
g
d
ialo
g
u
es
w
as
t
h
e
n
e
u
r
al
n
et
w
o
r
k
[
2
0
]
.
Ou
r
r
esear
ch
is
i
m
p
o
r
tan
t
b
ec
au
s
e,
i
n
a
p
ip
elin
e
ap
p
r
o
ac
h
,
t
h
e
f
ir
s
t
atte
m
p
t
is
to
u
s
e
a
n
eu
r
al
n
et
w
o
r
k
to
d
ialo
g
u
e
s
tate
tr
ac
k
i
n
g
to
o
b
tain
ap
p
r
o
p
r
iate
d
ata
f
r
o
m
t
h
e
u
s
er
u
tter
an
ce
s
.
I
n
th
e
ab
s
e
n
ce
o
f
th
e
n
ec
e
s
s
ar
y
u
s
er
i
n
te
r
p
r
etatio
n
d
ialo
g
u
e
f
r
a
m
e
w
o
r
k
,
th
e
s
e
s
ch
e
m
e
s
ca
n
b
e
u
s
ed
to
ac
cu
m
u
late
er
r
o
r
s
s
ep
ar
atel
y
w
it
h
i
n
th
e
la
n
g
u
a
g
e
m
o
d
u
le.
State
T
r
ac
k
in
g
Dia
lo
g
r
es
u
lt
s
d
em
o
n
s
tr
ate
th
e
u
tili
t
y
o
f
lear
n
in
g
to
co
llecti
v
el
y
i
n
ter
p
r
et
s
p
ee
ch
an
d
m
o
n
ito
r
d
ialo
g
u
e
[
1
4
,
1
8
,
1
9
]
.
T
h
ese
s
o
l
u
tio
n
s
ar
e
ex
tr
ac
ted
f
r
o
m
th
e
N
-
B
est
l
is
t
d
ev
elo
p
ed
b
y
t
h
e
a
u
to
m
ated
s
p
ee
ch
r
ec
o
g
n
i
tio
n
p
r
o
g
r
a
m
m
e.
B
y
av
o
id
i
n
g
er
r
o
r
b
u
ild
-
u
p
f
r
o
m
t
h
e
co
m
p
r
eh
e
n
s
io
n
d
im
e
n
s
io
n
o
f
t
h
e
o
r
ig
in
al
la
n
g
u
a
g
e.
S
u
c
h
f
r
a
m
e
w
o
r
k
s
i
n
clu
d
e
t
h
e
u
s
e
o
f
co
m
m
o
n
ta
g
i
n
g
s
to
o
v
er
r
id
e
u
n
i
q
u
e
s
lo
t
f
o
r
m
s
an
d
v
alu
e
s
,
as
w
ell
a
s
h
a
n
d
cr
af
ted
p
r
o
ce
d
u
r
al
d
ictio
n
ar
ies.
Nev
e
r
th
eles
s
,
s
u
ch
m
o
d
els r
el
y
o
n
h
an
d
cr
af
ted
f
ea
t
u
r
e
s
an
d
co
m
p
licated
d
o
m
ain
-
s
p
ec
if
ic
lex
ico
n
s
,
w
h
ic
h
ar
e
d
if
f
ic
u
lt
to
s
ca
le
f
o
r
ea
c
h
f
o
r
m
o
f
s
lo
t
an
d
th
er
e
f
o
r
e
d
if
f
ic
u
lt to
ap
p
l
y
to
n
e
w
d
o
m
a
in
s
.
R
ec
en
t
s
tate
-
o
f
-
t
h
e
-
ar
t
p
r
o
j
ec
tio
n
s
f
o
r
DST
p
r
ed
icted
th
e
co
n
d
itio
n
o
f
ev
er
y
tr
an
s
itio
n
b
y
o
b
s
er
v
in
g
u
n
i
f
o
r
m
co
n
s
u
m
er
a
n
d
m
ac
h
i
n
e
u
tter
an
ce
s
r
ep
r
esen
ta
tio
n
s
.
Nev
er
t
h
eles
s
,
t
h
e
e
f
f
ic
ien
c
y
o
f
th
e
s
e
s
ch
e
m
es
i
s
lo
w
i
n
th
e
u
n
u
s
u
a
l
an
d
u
n
f
a
m
i
liar
s
lo
t
v
al
u
es
t
h
at
h
av
e
r
ec
e
n
tl
y
b
ee
n
ad
d
r
ess
ed
b
y
lo
ca
l
s
l
o
t
en
co
d
er
s
[
3
]
an
d
th
e
p
o
in
ter
n
et
w
o
r
k
[
2
1
]
.
T
h
e
Glo
b
al
-
L
o
ca
l
Sel
f
-
A
tte
n
t
io
n
E
n
co
d
er
m
o
d
el
as
[
2
]
s
u
g
g
ested
r
ec
u
r
r
en
t
s
el
f
-
atten
tio
n
n
et
w
o
r
k
s
w
it
h
a
co
m
p
u
ted
r
ep
r
esen
tat
i
o
n
b
y
co
m
p
ar
i
n
g
t
h
e
s
i
m
ilar
it
ies
o
f
ev
er
y
s
lo
t
v
al
u
e
to
ea
ch
u
s
er
u
tter
an
ce
a
n
d
p
r
io
r
d
ev
ice
b
eh
av
io
u
r
.
[
3
]
i
m
p
r
o
v
ed
g
lo
b
al
-
lo
ca
l
s
e
lf
-
c
ar
e
en
co
d
er
s
tr
u
ctu
r
e
b
y
eli
m
i
n
ati
n
g
s
lo
t
-
b
ased
r
ec
u
r
r
en
t
v
o
ice
n
e
t
w
o
r
k
s
an
d
s
y
s
te
m
en
co
d
er
s
a
n
d
u
s
i
n
g
a
g
lo
b
al
-
co
n
d
it
io
n
ed
e
m
b
ed
d
ed
s
lo
t
s
t
y
le
en
co
d
er
.
Nev
er
th
e
less
,
b
ec
a
u
s
e
o
f
t
h
eir
lack
o
f
ac
ti
v
it
y
i
n
u
n
d
er
s
ta
n
d
in
g
an
d
i
n
co
r
p
o
r
atin
g
th
e
r
ele
v
an
t
co
n
te
x
t,
t
h
ese
ap
p
r
o
ac
h
es
w
er
e
n
o
t
s
u
cc
e
s
s
f
u
l
i
n
t
h
e
p
r
o
d
u
ctio
n
s
y
s
te
m
,
w
h
ile
t
h
i
s
r
esear
ch
m
a
y
u
n
d
er
s
tan
d
t
h
e
co
n
n
ec
ti
n
g
s
en
s
e.
3.
P
RO
P
O
SE
D
M
O
DE
L
W
e
p
r
e
s
e
n
t
th
e
p
r
o
p
o
s
e
d
m
o
d
e
l
i
n
th
i
s
s
e
ct
i
o
n
.
S
e
ct
i
o
n
3
.
1
f
i
r
s
t
ex
p
l
a
in
s
th
e
r
e
ce
n
t
ly
p
r
o
p
o
s
e
d
a
r
c
h
it
e
c
tu
r
e
o
f
GC
E
[
3
]
,
f
o
l
l
o
w
e
d
b
y
th
e
p
r
o
p
o
s
e
d
en
c
o
d
e
r
in
S
e
c
t
i
o
n
3
.
2
th
en
th
e
m
o
d
e
l
s
c
o
u
r
in
Se
c
t
i
o
n
3
.
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4752
Dia
lo
g
u
e
s
ta
te
tr
a
ck
in
g
a
cc
u
r
a
cy
imp
r
o
ve
men
t b
y
d
is
tin
g
u
is
h
in
g
…
(
K
h
a
ld
o
o
n
H.
A
lh
u
s
s
a
yn
i
)
1059
3
.
1
.
G
lo
ba
lly
co
nd
it
io
ned e
nco
de
r
(
G
CE
)
E
ac
h
en
co
d
er
i
n
p
u
t i
s
r
ep
r
ese
n
ted
as a
r
ep
r
esen
tatio
n
o
f
th
e
v
ec
to
r
(
C
)
.
I
n
t
h
e
G
C
E
m
o
d
el
t
h
er
e
is
th
e
t
w
o
-
w
a
y
L
ST
M
s
y
s
te
m
[
2
2
]
an
d
all
s
lo
ts
w
h
ich
ar
e
s
h
a
r
ed
to
ac
h
iev
e
a
s
eq
u
en
ce
o
f
h
id
d
en
s
tate
s
b
y
en
cr
y
p
ti
n
g
t
h
e
i
n
p
u
t
s
eq
u
en
ce
,
s
h
ad
ed
b
y
a
s
el
f
atte
n
tio
n
la
y
er
[
2
3
]
.
T
h
e
GC
E
ap
p
r
o
ac
h
tak
es
in
to
ac
co
u
n
t
u
s
er
ex
p
r
es
s
io
n
a
n
d
th
e
p
r
ev
i
o
u
s
s
y
s
te
m
ac
tio
n
f
o
r
m
o
d
el
lear
n
in
g
.
Ho
w
ev
er
,
i
n
o
u
r
w
o
r
k
,
w
e
u
s
ed
a
GC
E
ap
p
r
o
ac
h
to
lea
r
n
d
is
tr
ib
u
tio
n
o
f
Slo
t
Valu
e
P
air
s
an
d
co
n
tex
t,
n
o
t
o
n
l
y
c
u
r
r
en
t
u
s
er
u
tte
r
an
ce
s
an
d
p
r
ev
io
u
s
s
y
s
te
m
ac
t
io
n
s
b
u
t a
ls
o
p
r
ev
io
u
s
u
s
er
u
t
ter
an
ce
.
3
.
2
.
E
nco
der
m
o
del
W
e
f
o
llo
w
ed
th
e
p
r
o
p
o
s
ed
ar
ch
itect
u
r
e
f
o
r
ca
lc
u
latio
n
o
f
ea
ch
s
lo
t
v
al
u
e
p
air
'
s
en
co
d
er
,
u
s
er
u
tter
an
ce
a
n
d
p
r
ev
io
u
s
s
y
s
te
m
ac
tio
n
s
at
G
C
E
.
Ho
w
e
v
er
,
we
u
s
e
a
n
ad
d
itio
n
al
en
co
d
er
to
ex
tr
ac
t
th
e
h
i
s
to
r
y
an
d
th
e
co
n
te
x
t
f
o
r
p
r
ev
io
u
s
u
s
er
u
tter
a
n
ce
s
.
T
h
e
en
co
d
er
m
o
d
el
is
u
s
ed
to
e
n
co
d
e
th
e
p
r
ev
io
u
s
u
s
er
u
tter
an
ce
s
(
Hp
er
,
C
p
er
)
,
th
e
c
u
r
r
en
t
u
s
er
u
tter
a
n
ce
s
(
Hcu
r
,
C
cu
r
)
,
th
e
p
r
ev
io
u
s
s
y
s
te
m
ac
tio
n
s
(
He,
C
a)
an
d
s
lo
t
v
al
u
e
(
H
v
,
C
v
)
f
o
r
ea
ch
s
y
s
te
m
ac
t.
As
s
h
o
w
n
i
n
F
ig
u
r
e
1
,
th
e
s
lo
t
-
e
m
b
ed
d
in
g
v
ec
to
r
f
o
r
th
e
k
th
s
lo
t
i
s
u
s
ed
f
o
r
co
n
t
e
x
t e
x
tr
ac
tio
n
.
Fig
u
r
e
1
.
Dialo
g
u
e
h
is
to
r
ical
c
o
n
tex
t
s
elf
-
at
ten
tio
n
m
o
d
el
f
o
r
d
ialo
g
u
e
s
tate
tr
ac
k
er
T
o
ca
lcu
late
r
ep
r
esen
tatio
n
H
k
f
o
r
ea
ch
s
lo
t
k
th
a
s
s
h
o
w
n
i
n
(
1
)
,
w
e
co
n
ca
ten
ated
t
h
e
s
lo
t
e
m
b
ed
d
i
n
g
s
k
w
it
h
in
p
u
t
s
eq
u
e
n
ce
X,
i.e
.
,
cu
r
r
en
t
u
s
er
u
tter
an
ce
,
p
r
ev
io
u
s
u
s
er
u
tter
an
ce
,
o
r
p
r
ev
io
u
s
s
y
s
te
m
ac
tio
n
s
,
a
s
in
p
u
t to
t
h
e
en
co
d
er
,
w
h
er
e
co
n
ca
ten
a
tio
n
i
s
d
en
o
ted
as
(
)
.
(
(
)
)
(
1
)
W
h
er
e
d
r
is
th
e
d
im
e
n
s
io
n
o
f
t
h
e
L
ST
M
s
tate.
T
h
en
w
e
ca
lc
u
late
th
e
atte
n
tio
n
s
co
r
e
o
f
th
e
s
lo
t
f
o
r
ea
ch
to
k
e
n
h
id
d
en
r
ep
r
esen
tat
io
n
as
s
h
o
w
n
in
(
2
)
,
b
y
co
n
c
aten
ati
n
g
th
e
m
to
t
h
e
s
lo
t
e
m
b
ed
d
i
n
g
an
d
tr
an
s
ito
r
y
to
a
lin
ea
r
lay
er
,
th
en
ap
p
ly
i
n
g
a
s
o
f
t
m
a
x
in
(
3
)
to
n
o
r
m
alize
t
h
e
d
is
tr
ib
u
tio
n
.
I
n
(
4
)
co
m
p
u
t
e
s
i
m
ilar
l
y
o
f
th
e
co
n
tex
t
.
(
)
(
2
)
(
)
(
3
)
∑
(
4
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
21
,
No
.
2
,
Feb
r
u
ar
y
2
0
2
1
:
10
57
-
10
64
1060
E
ac
h
o
f
th
e
f
o
u
r
e
n
co
d
er
s
in
t
h
e
en
co
d
er
m
o
d
el,
as sh
o
w
n
i
n
Fi
g
u
r
e
1
,
ca
n
b
e
r
ep
r
esen
ted
as f
o
llo
w
s
:
U,
s
k
a
s
in
p
u
ts
a
n
d
as o
u
tp
u
t
s
,
w
h
er
e
U
d
en
o
tes
w
o
r
d
e
m
b
ed
d
in
g
s
o
f
t
h
e
u
s
er
u
tter
an
ce
.
P
,
s
k
as in
p
u
ts
a
n
d
as o
u
tp
u
t
s
,
w
h
er
e
P
d
en
o
tes
w
o
r
d
e
m
b
ed
d
in
g
s
o
f
t
h
e
p
r
ev
io
u
s
u
s
er
u
tter
an
ce
.
A
j
, s
k
a
s
in
p
u
ts
a
n
d
as o
u
tp
u
t
s
,
w
h
er
e
A
j
j
th
is
t
h
e
p
r
ev
io
u
s
s
y
s
te
m
ac
tio
n
.
V,
s
k
a
s
in
p
u
ts
a
n
d
as o
u
tp
u
t
s
,
w
h
er
e
V
d
en
o
tes cu
r
r
e
n
t slo
t
-
v
alu
e
p
air
.
3
.
3
.
Sco
ring
m
o
de
l
I
n
GL
A
D
[
2
]
,
w
e
ad
o
p
ted
th
e
s
u
g
g
e
s
ted
m
et
h
o
d
to
ca
lcu
late
th
e
p
er
f
o
r
m
a
n
ce
o
f
ea
ch
s
lo
t
-
v
alu
e
p
air
in
cu
r
r
e
n
t
an
d
p
r
ev
io
u
s
co
n
s
u
m
er
u
tter
a
n
ce
s
a
n
d
p
r
ev
io
u
s
s
y
s
te
m
s
tep
s
.
T
h
e
y
h
o
w
ev
er
u
s
e
t
h
e
ad
d
itio
n
al
p
o
in
t to
i
m
p
r
o
v
e
th
e
m
ea
n
in
g
an
d
d
eliv
er
y
o
v
er
t
h
e
h
i
s
to
r
y
o
f
d
ialo
g
u
e.
T
h
e
s
co
r
es
m
o
d
el
is
u
s
ed
to
m
ea
s
u
r
e
s
lo
t
k
f
o
r
its
s
lo
t
v
al
u
es
i
n
o
r
d
er
to
ev
al
u
ate
t
h
e
s
l
o
t
v
alu
e
th
a
t
th
e
u
s
e
r
s
n
a
m
e.
T
h
is
w
as
also
d
o
n
e
u
s
in
g
f
iv
e
e
x
a
m
p
le
s
.
T
h
e
f
ir
s
t
s
co
r
e
as
s
h
o
w
n
in
(
5
)
is
th
e
cu
r
r
en
t
u
s
er
u
tter
an
ce
H
cur
,
ta
k
i
n
g
i
n
to
ac
co
u
n
t
t
h
e
s
lo
t
-
v
al
u
e
p
ai
r
b
ein
g
co
n
s
id
er
ed
c
v
an
d
u
s
i
n
g
th
e
r
es
u
lti
n
g
atten
tio
n
co
n
te
x
t q
cur
as sh
o
w
n
in
(
6
)
to
s
co
r
e
th
e
s
lo
t
-
v
al
u
e
p
air
.
(
(
)
)
(
5
)
∑
)
(
6
)
(
7
)
w
h
er
e
m
i
n
d
icate
s
a
n
u
m
b
er
o
f
w
o
r
d
s
i
n
t
h
e
in
p
u
t
s
eq
u
e
n
ce
.
T
h
e
s
co
r
e
as
s
h
o
w
n
i
n
(
7
)
d
en
o
tes
th
e
p
r
ed
icted
v
alu
es o
f
th
e
u
s
e
r
u
tter
an
ce
.
T
h
e
s
ec
o
n
d
s
co
r
e
as
s
h
o
w
n
in
(
8
)
is
s
i
m
ilar
to
th
e
f
ir
s
t
s
co
r
e,
b
u
t
u
s
es
p
r
ev
io
u
s
u
s
er
u
tter
a
n
ce
H
pre
in
s
tead
o
f
th
e
c
u
r
r
en
t
u
s
er
u
tter
a
n
ce
,
ta
k
i
n
g
in
to
ac
co
u
n
t
th
e
s
lo
t
-
v
a
lu
e
p
air
b
ein
g
co
n
s
id
er
ed
c
v
a
n
d
u
s
i
n
g
th
e
r
es
u
lti
n
g
a
tten
t
io
n
co
n
te
x
t
as
s
h
o
w
n
i
n
(
9
)
to
s
co
r
e
th
e
s
lo
t
-
v
al
u
e
p
air
.
T
h
e
s
co
r
e
as
s
h
o
w
n
in
(
1
0
)
d
en
o
tes th
e
p
r
ed
icted
v
alu
eso
f
th
e
p
r
ev
io
u
s
u
tter
an
ce
.
(
(
)
)
(
8
)
∑
)
(
9
)
(
1
0
)
T
h
en
th
e
p
r
ed
icted
v
alu
es
o
f
b
o
th
cu
r
r
en
t
an
d
p
r
ev
io
u
s
u
s
e
r
u
tter
an
ce
s
ar
e
ad
d
ed
as
s
h
o
w
n
in
t
h
e
f
o
llo
w
in
g
as
s
h
o
w
n
in
(
1
1
)
:
(
1
1
)
Si
m
i
lar
l
y
,
th
i
s
is
u
s
ed
to
d
eter
m
i
n
e
t
h
e
m
e
n
tio
n
ed
p
r
ev
io
u
s
s
y
s
te
m
ac
tio
n
s
in
t
h
e
c
u
r
r
en
t
o
r
p
r
ev
io
u
s
u
s
er
u
tter
a
n
ce
s
ep
ar
atel
y
to
r
ea
ch
s
u
f
f
icie
n
t
i
n
f
o
r
m
atio
n
a
b
o
u
t
u
s
er
u
t
ter
an
ce
w
h
e
n
t
h
i
s
is
n
o
t
i
n
f
o
r
m
ati
v
e.
T
h
e
th
ir
d
s
co
r
e
as
s
h
o
w
n
in
(
1
2
)
,
th
e
co
n
tex
t
o
f
cu
r
r
en
t
u
s
er
u
tter
an
ce
C
cur
o
v
er
th
e
p
r
ev
io
u
s
ac
tio
n
r
ep
r
esen
tatio
n
s
C
a
=
[
C
a1
·
·
·
C
al
]
.
Her
e,
l
is
t
h
e
n
u
m
b
e
r
o
f
p
r
ev
io
u
s
s
y
s
te
m
ac
tio
n
s
.
T
h
en
w
e
u
s
e
th
e
s
i
m
ilar
it
y
b
et
w
ee
n
t
h
e
atten
t
io
n
co
n
te
x
t
q
acur
as
s
h
o
w
n
in
(
1
3
)
an
d
th
e
s
l
ot
-
v
al
u
e
p
air
c
v
to
s
co
r
e
th
e
s
lo
t
-
v
alu
e
p
air
.
T
h
e
s
co
r
e
as
s
h
o
w
n
i
n
(
1
4
)
d
en
o
tes
th
e
p
r
ed
icted
v
alu
e
s
o
f
th
ep
r
ev
io
u
s
s
y
s
te
m
ac
tio
n
s
i
n
th
e
cu
r
r
en
t u
s
er
u
tter
a
n
ce
s
ep
ar
ate
l
y
.
(
(
)
)
(
1
2
)
∑
)
(
1
3
)
(
1
4
)
T
h
e
f
o
u
r
th
s
o
u
r
ce
as
s
h
o
w
n
i
n
(
1
5
)
,
s
i
m
ilar
to
th
e
th
ir
d
s
co
r
e,
b
u
t
u
s
es
th
e
co
n
tex
t
o
f
p
r
ev
io
u
s
u
s
er
u
tter
an
ce
C
pre
in
s
er
ted
in
t
o
th
e
co
n
tex
t o
f
a
cu
r
r
en
t
u
s
er
u
tter
an
ce
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4752
Dia
lo
g
u
e
s
ta
te
tr
a
ck
in
g
a
cc
u
r
a
cy
imp
r
o
ve
men
t b
y
d
is
tin
g
u
is
h
in
g
…
(
K
h
a
ld
o
o
n
H.
A
lh
u
s
s
a
yn
i
)
1061
(
(
)
)
(
1
5
)
∑
)
(
1
6
)
(
1
7
)
T
h
en
th
e
p
r
ed
icted
v
alu
e
s
o
f
b
o
th
th
e
co
n
te
x
t o
f
th
e
c
u
r
r
en
t a
n
d
p
r
ev
io
u
s
u
s
er
u
tter
an
ce
s
ar
e
ad
d
ed
:
(
1
8
)
I
n
t
h
e
las
t
s
co
r
e,
w
e
f
o
llo
w
ed
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
in
[
2
4
]
to
d
eter
m
in
e
th
e
r
elev
a
n
ce
o
f
t
h
e
s
lo
t
-
v
alu
e
p
air
in
t
h
e
c
u
r
r
en
t
t
u
r
n
.
W
e
co
n
ca
ten
a
te
t
h
e
co
n
te
x
t
o
f
b
o
th
c
u
r
r
en
t
u
s
er
u
t
ter
an
ce
an
d
p
r
ev
io
u
s
u
s
er
u
tter
an
ce
an
d
u
s
e
a
s
i
g
m
o
id
ac
tiv
atio
n
o
f
t
h
e
li
n
ea
r
la
y
er
to
co
m
p
u
te
t
h
e
s
co
r
e.
(
)
(
1
9
)
(
(
)
)
(
2
0
)
T
h
en
w
e
co
m
p
u
te
co
n
te
x
t s
u
m
m
ar
ies l
u
of
atte
n
tio
n
f
r
o
m
C
v
o
v
er
H
cur
an
d
l
s
o
f
atten
t
io
n
f
r
o
m
C
v
o
v
er
H
per
.
(
)
(
2
1
)
(
)
(
2
2
)
T
o
c
o
m
p
u
te
th
e
ad
d
itio
n
al
s
c
o
r
e
y
f
t
h
at
es
tab
lis
h
es
t
h
e
p
r
o
b
ab
ilit
y
o
f
t
h
e
ca
n
d
id
ate
s
lo
t
-
v
alu
e
b
a
s
ed
o
n
b
o
th
th
e
cu
r
r
e
n
t a
n
d
p
r
ev
io
u
s
u
s
er
u
t
ter
an
ce
s
a
n
d
th
e
p
r
ev
io
u
s
s
y
s
te
m
u
tter
an
ce
,
w
e
u
s
e:
(
)
(
2
3
)
Fin
all
y
,
w
e
ad
d
th
e
w
ei
g
h
t
o
f
all
s
co
r
es
o
f
s
lo
t
k
,
i.e
.
,
,
,
an
d
,
w
h
ich
ar
e
n
o
r
m
a
lized
b
y
t
h
e
s
ig
m
o
id
f
u
n
ctio
n
:
(
+
+
)
(
2
4
)
w
h
er
e
w
is
a
lear
n
ed
p
ar
a
m
ete
r
.
4.
E
XP
E
R
I
M
E
NT
T
h
e
W
izar
d
o
f
Oz
(
W
o
Z
)
is
a
s
in
g
le
d
o
m
ai
n
f
o
r
a
m
o
n
ito
r
in
g
p
latf
o
r
m
d
ialo
g
u
e
b
o
o
k
in
g
(
W
en
et
al.
2
0
1
6
)
.
T
h
e
e
m
o
tio
n
d
ataset
co
n
s
i
s
ts
o
f
6
0
0
in
s
tr
u
c
tio
n
d
ialo
g
u
e
s
,
2
0
0
ap
p
r
aisal
d
ialo
g
u
e
s
a
n
d
4
0
0
ass
es
s
m
en
t
s
.
-
d
ialo
g
u
e
h
as
a
t
o
tal
o
f
ei
g
h
t
t
u
r
n
s
,
w
it
h
a
m
ac
h
in
e
tr
an
s
cr
ip
t,
a
u
s
er
u
tter
an
c
e
tr
an
s
cr
ip
t,
a
tu
r
n
m
ar
k
an
d
d
ialo
g
u
e
s
tatu
s
f
o
r
ea
ch
s
w
itc
h
.
T
h
e
o
n
to
lo
g
y
co
n
s
i
s
ts
o
f
th
r
ee
s
ep
a
r
ate
s
lo
t
f
o
r
m
s
:
f
o
o
d
w
it
h
7
2
v
alu
e
s
,
7
v
al
u
es,
4
v
al
u
e
s
,
an
d
7
d
if
f
er
e
n
t
k
in
d
o
f
s
lo
ts
,
in
cl
u
d
in
g
telep
h
o
n
e
n
u
m
b
er
an
d
ad
d
r
ess
.
T
h
e
o
n
to
lo
g
y
co
n
s
is
t
s
o
f
th
r
ee
s
p
ec
if
ic
s
lo
t
s
t
y
le
s
.
I
n
o
r
d
er
t
o
ass
ess
t
h
e
m
etr
ic
o
f
m
u
tu
al
tar
g
et
m
o
n
ito
r
i
n
g
p
r
ec
is
io
n
,
w
e
o
b
s
er
v
ed
t
h
e
ag
g
r
eg
atio
n
o
f
t
u
r
n
s
tar
g
ets
[
2
]
.
4
.
1
.
I
m
ple
m
ent
a
t
io
n
det
a
ils
T
h
e
p
r
e
-
tr
ain
ed
em
b
ed
d
in
g
o
f
GL
o
Ve
w
o
r
d
[
2
5
]
is
u
s
ed
,
w
h
ic
h
is
p
air
ed
w
it
h
n
-
g
r
a
m
e
m
b
ed
d
in
g
ch
ar
ac
ter
is
tic
s
,
an
d
is
r
etai
n
ed
th
r
o
u
g
h
o
u
t
th
e
tr
ain
i
n
g
.
T
h
e
u
n
i
ts
f
o
r
B
I
-
L
ST
Ms
ar
e
d
escr
ib
ed
in
2
0
0
h
id
d
en
d
i
m
en
s
io
n
s
an
d
th
e
i
n
itia
l
le
a
r
n
in
g
r
ate
is
tr
ain
ed
u
s
i
n
g
an
A
D
AM
o
p
ti
m
izer
[
2
6
]
.
Fo
r
th
e
e
m
b
ed
d
in
g
la
y
er
,
w
e
d
ef
in
ed
t
h
e
d
r
o
p
-
o
u
t r
ate.
T
h
e
av
er
ag
e
n
u
m
b
er
o
f
ti
m
e
s
i
s
5
0
to
tr
ain
th
e
5
0
b
atch
m
o
d
els.
4
.
2
.
Resul
t
s
a
n
d
a
na
ly
s
is
T
h
e
p
r
ev
io
u
s
DST
co
n
s
u
m
er
d
ec
lar
atio
n
is
u
s
ed
to
g
ain
t
h
e
n
ec
ess
ar
y
in
f
o
r
m
a
tio
n
to
i
n
cr
ea
s
e
th
e
lik
eli
h
o
o
d
o
f
allo
ca
tio
n
a
m
o
n
g
all
ap
p
lican
t
s
lo
t
-
v
al
u
e
p
air
s
an
d
ac
tio
n
b
eh
a
v
io
u
r
.
Usi
n
g
a
m
o
d
el
to
p
r
ac
tis
e
an
d
ch
ec
k
t
h
e
w
h
o
le
c
y
cle
i
s
s
tr
ea
m
li
n
ed
an
d
p
ac
e
b
en
ef
its
.
T
h
e
en
co
d
er
s
id
e
o
f
th
e
m
o
d
el
is
d
eter
m
i
n
ed
f
o
r
b
o
th
ac
tiv
ities
o
n
l
y
o
n
ce
.
I
n
a
s
eg
m
e
n
t
s
co
r
in
g
m
o
d
el,
w
e
d
eter
m
i
n
e
t
h
e
o
u
tco
m
e
s
i
n
d
ep
en
d
en
tl
y
f
r
o
m
th
e
p
r
ev
io
u
s
u
s
er
s
tate
m
e
n
t
a
n
d
c
u
r
r
en
t
u
s
er
s
tate
m
en
t
o
v
er
a
ll
ap
p
lican
t
s
lo
t
p
air
s
a
n
d
in
t
er
v
en
tio
n
s
tep
s
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
21
,
No
.
2
,
Feb
r
u
ar
y
2
0
2
1
:
10
57
-
10
64
1062
in
cr
ea
s
e
co
v
er
a
g
e
an
d
d
is
tr
ib
u
tio
n
o
v
er
d
ialo
g
u
e
h
is
to
r
y
,
a
s
s
h
o
w
n
th
r
o
u
g
h
Fi
g
u
r
e
1
.
I
n
f
a
ct,
b
o
th
th
e
c
u
r
r
en
t
an
d
p
r
ev
io
u
s
co
n
s
u
m
er
u
tter
a
n
ce
s
ar
e
f
u
s
ed
in
t
h
e
s
en
s
e
o
f
ea
ch
ap
p
lican
t
s
lo
t
m
ea
n
i
n
g
p
air
,
as
s
h
o
w
n
i
n
(
th
e
ad
d
itio
n
al
s
co
r
e
o
f
)
Fi
g
u
r
e
1
.
C
o
m
p
ar
ed
w
it
h
all
s
tate
-
o
f
-
th
e
-
ar
t
s
tr
ate
g
ie
s
i
n
t
h
e
s
h
ar
ed
g
o
al
p
h
ase,
t
h
e
p
r
o
p
o
s
ed
m
o
d
el
o
b
tain
ed
s
u
p
e
r
io
r
r
esu
lts
,
w
h
ic
h
d
id
n
o
t in
c
l
u
d
e
th
e
r
eq
u
est t
u
r
n
in
g
c
h
alle
n
g
e.
T
ab
le
1
p
r
o
v
id
es
a
s
u
m
m
ar
y
o
f
th
e
r
esu
lt
s
o
f
o
u
r
n
e
w
m
o
d
els
w
i
th
t
h
o
s
e
o
f
p
r
ev
io
u
s
s
ta
t
e
-
o
f
-
t
h
e
-
ar
t
p
r
o
j
ec
ts
.
T
h
e
NB
T
[
1
]
,
w
h
ich
u
t
ilis
e
s
co
m
m
o
n
m
ar
k
s
i
n
s
tead
o
f
s
lo
ts
a
n
d
m
ea
n
i
n
g
s
in
th
e
u
t
ter
an
ce
,
u
s
es
C
NN
w
it
h
p
r
e
-
tr
ai
n
i
n
g
ter
m
e
m
b
ed
d
in
g
s
f
o
r
r
ep
r
esen
tatio
n
al
lear
n
i
n
g
i
n
s
tead
o
f
d
ele
x
i
ca
lis
ed
[
1
5
]
.
I
n
th
e
d
esig
n
o
f
G
L
A
D
[
2
]
,
th
e
u
tter
an
ce
,
p
r
ev
io
u
s
s
y
s
te
m
o
p
er
atio
n
,
an
d
all
s
lo
t
v
alu
e
s
ar
e
s
eg
r
eg
ated
in
e
n
co
d
er
s
.
Fo
r
all
en
co
d
er
s
,
a
co
m
m
o
n
d
esig
n
i
s
u
s
ed
.
T
h
e
w
r
iter
s
u
s
ed
t
w
o
r
atin
g
s
,
ea
c
h
m
ea
s
u
r
i
n
g
th
e
r
el
atio
n
b
et
w
ee
n
th
e
s
lo
t
m
ea
n
i
n
g
p
ai
r
s
an
d
th
e
co
n
s
u
m
er
u
tter
a
n
c
e
o
r
p
r
ev
io
u
s
s
y
s
te
m
o
p
er
atio
n
.
T
h
e
r
esear
ch
er
s
f
o
llo
w
ed
a
s
i
m
ilar
m
e
th
o
d
to
GL
A
D
in
t
h
e
G
C
E
m
o
d
el.
T
h
e
E
n
co
d
er
h
as
h
o
w
e
v
er
b
ee
n
u
p
d
ated
to
in
cr
ea
s
e
laten
c
y
a
n
d
s
p
ee
d
o
f
in
f
er
en
ce
b
y
r
e
m
o
v
i
n
g
i
n
ad
eq
u
ate
r
ec
u
r
r
en
t la
y
er
s
a
n
d
s
el
f
-
a
tten
tio
n
l
a
y
er
s
.
T
h
e
W
o
Z
d
ataset
h
a
s
b
ee
n
u
s
ed
to
tes
t
o
u
r
p
r
o
p
o
s
ed
m
o
d
el
an
d
in
tr
o
d
u
ce
t
w
o
m
o
d
els.
I
n
t
h
e
d
ialo
g
u
e
o
f
id
en
tical
d
esig
n
[
3
]
,
p
r
ev
io
u
s
u
s
er
d
e
f
i
n
itio
n
s
w
er
e
u
s
ed
as
ad
d
itio
n
a
l
en
co
d
er
s
w
it
h
t
h
e
s
a
m
e
ac
cu
r
ac
ies as
GC
E
la
y
o
u
t.
T
h
e
s
ec
o
n
d
w
as b
y
in
te
g
r
ati
n
g
th
e
m
ea
n
in
g
o
f
p
r
e
v
io
u
s
u
s
er
u
tt
er
an
ce
s
a
n
d
cu
r
r
en
t
u
s
er
u
t
ter
an
ce
s
w
i
th
all
ca
n
d
id
ate
s
lo
t
v
alu
e
p
air
s
.
I
n
cr
ea
s
i
n
g
e
x
is
t
in
g
u
s
er
u
tter
a
n
ce
w
i
th
th
e
co
r
r
ec
t
u
s
er
in
f
o
r
m
atio
n
a
n
d
th
e
p
r
ev
io
u
s
s
y
s
te
m
u
tter
a
n
ce
also
i
m
p
r
o
v
e
s
th
e
s
tan
d
ar
d
tar
g
et
ac
cu
r
ac
y
b
y
0
.
8
%,
as
s
h
o
w
n
b
y
t
h
e
f
in
d
i
n
g
s
i
n
T
ab
le
1
an
d
r
eq
u
est ac
cu
r
ac
y
.
T
ab
le
1
.
C
o
m
p
ar
is
o
n
o
f
o
u
r
m
o
d
el
to
p
r
e
v
io
u
s
l
y
p
u
b
lis
h
ed
W
o
Z
r
estau
r
an
t r
eser
v
atio
n
d
ataset
T
u
r
n
R
e
q
u
e
st
Jo
i
n
t
G
o
a
l
M
o
d
e
l
9
1
.
2
%
8
4
.
4
%
N
e
u
r
a
l
B
e
l
i
e
f
Tr
a
c
k
e
r
(
N
B
T
)
—
DNN
9
1
.
6
%
8
4
.
2
%
N
e
u
r
a
l
B
e
l
i
e
f
Tr
a
c
k
e
r
(
N
B
T
)
—
C
N
N
9
7
.
1
%
8
8
.
1
%
G
l
o
b
a
l
-
L
o
c
a
l
l
y
S
e
l
f
-
A
t
t
e
n
t
i
v
e
(
G
L
D
A
)
9
7
.
3
8
%
8
8
.
5
%
G
l
o
b
a
l
l
y
C
o
n
d
i
t
i
o
n
e
d
E
n
c
o
d
e
r
(
G
C
E)
9
6
.
9
8
%
9
6
.
9
8
%
8
8
.
5
%
8
9
.
3
%
G
C
E
+
p
r
e
v
i
o
u
s
u
t
t
e
r
a
n
c
e
G
C
E
+
p
r
e
v
i
o
u
s
u
t
t
e
r
a
n
c
e
+
a
d
d
i
t
i
o
n
a
l
s
c
o
r
e
5.
CO
NCLU
SI
O
N
DST
'
s
o
b
j
ec
tiv
e
is
to
d
ef
i
n
e
t
w
o
g
o
als:
t
h
e
cu
r
r
e
n
t
s
tate
o
f
ea
ch
r
o
tatio
n
o
f
d
ialo
g
u
e
s
lo
t
v
alu
e
p
air
s
an
d
th
e
u
s
er
d
ialo
g
u
e
ac
t
to
s
u
m
u
p
all
t
h
e
u
s
er
's
o
b
j
ec
tiv
es.
I
n
th
is
p
ap
er
,
w
e
p
r
o
p
o
s
e
a
d
if
f
er
en
t
f
r
a
m
e
w
o
r
k
f
o
r
DST
as
a
d
ialo
g
u
e
o
n
t
h
e
h
i
s
to
r
ical
c
o
n
te
x
t
s
el
f
-
Atte
n
tio
n
to
d
is
ti
n
g
u
i
s
h
a
n
d
u
s
e
t
h
e
r
elate
d
d
ialo
g
u
e
h
is
to
r
y
b
y
i
n
cl
u
d
in
g
th
e
p
r
ev
i
o
u
s
u
s
er
's
u
tter
a
n
ce
in
t
w
o
s
e
p
ar
ate
w
a
y
s
a
n
d
to
f
u
s
e
in
o
r
d
er
to
p
r
o
v
id
e
th
e
m
o
d
el
w
it
h
th
e
n
ec
es
s
ar
y
in
f
o
r
m
atio
n
.
R
ela
tiv
e
to
all
s
tate
-
o
f
-
th
e
-
ar
t
m
eth
o
d
s
,
th
e
p
r
o
p
o
s
ed
m
o
d
el
o
b
tain
ed
o
u
tco
m
es
o
f
0
.
8
%
in
th
e
s
h
ar
ed
aim
p
r
ec
is
io
n
w
it
h
th
e
W
o
Z
d
ata
s
et
u
s
ed
to
test
o
u
r
p
r
o
p
o
s
ed
m
o
d
el.
Ou
r
m
o
d
el
u
s
es
a
s
m
aller
n
u
m
b
er
o
f
lear
n
in
g
p
ar
a
m
eter
s
ap
p
lied
to
in
ter
p
r
et
cu
r
r
en
t
an
d
p
r
ev
io
u
s
u
s
er
s
ta
te
m
e
n
t
s
u
s
i
n
g
t
h
e
ap
p
r
o
p
r
iate
co
n
tex
t.
T
h
ey
also
h
av
e
a
h
ig
h
d
is
p
ar
it
y
i
n
t
h
e
p
r
ec
is
io
n
o
f
th
e
j
o
in
t
ai
m
b
ec
a
u
s
e
t
h
e
j
o
in
t
o
b
j
ec
tiv
e
is
d
ec
id
ed
b
y
p
ili
n
g
tu
r
n
tar
g
ets
a
n
d
er
r
o
r
s
w
h
en
es
ti
m
ati
n
g
a
tu
r
n
d
esti
n
at
io
n
ar
e
co
m
p
ar
ati
v
el
y
s
m
aller
.
RE
F
E
R
E
NC
E
S
[1
]
N.
M
rk
šić
,
D.
Ó.
S
é
a
g
h
d
h
a
,
T
.
-
H.
W
e
n
,
e
t
a
l.
,
“
Ne
u
ra
l
b
e
li
e
f
trac
k
e
r
:
Da
t
a
-
d
riv
e
n
d
ialo
g
u
e
sta
te
trac
k
in
g
,
”
Pro
c
e
e
d
in
g
s
o
f
t
h
e
5
5
t
h
An
n
u
a
l
M
e
e
ti
n
g
o
f
th
e
Asso
c
i
a
ti
o
n
f
o
r Co
mp
u
ta
ti
o
n
a
l
L
i
n
g
u
isti
c
s
,
p
p
.
1
7
7
7
-
1
7
8
8
,
2
0
1
6
.
[2
]
V
.
Z
h
o
n
g
,
C.
X
i
o
n
g
,
a
n
d
R
.
S
o
c
h
e
r,
“
G
lo
b
a
l
-
L
o
c
a
ll
y
S
e
l
f
-
A
tt
e
n
ti
v
e
Dia
lo
g
u
e
S
tate
T
ra
c
k
e
r,
”
Pro
c
e
e
d
in
g
s
o
f
t
h
e
5
6
t
h
A
n
n
u
a
l
M
e
e
ti
n
g
o
f
t
h
e
Asso
c
ia
ti
o
n
f
o
r Co
m
p
u
t
a
ti
o
n
a
l
L
in
g
u
ist
ic
,
p
p
.
1
4
5
8
-
1
4
6
7
,
2
0
1
8
.
[3
]
E.
No
u
ri
a
n
d
E.
Ho
ss
e
in
i
-
A
sl,
“
To
w
a
rd
sc
a
lab
le
n
e
u
ra
l
d
ialo
g
u
e
sta
te
trac
k
in
g
m
o
d
e
l,
”
3
2
n
d
Co
n
fer
e
n
c
e
o
n
Ne
u
ra
l
In
fo
rm
a
t
io
n
Pro
c
e
ss
in
g
S
y
ste
ms
(
Ne
u
rIPS
2
0
1
8
),
2
n
d
C
o
n
v
e
rs
a
ti
o
n
a
l
AI
wo
rk
sh
o
p
M
o
n
tré
a
l
,
Ca
n
a
d
a
.
2
0
1
8
.
[4
]
R.
Hig
a
sh
in
a
k
a
,
M
.
N
a
k
a
n
o
,
a
n
d
K.
A
ik
a
wa
,
“
Co
rp
u
s
-
b
a
se
d
d
isc
o
u
rse
u
n
d
e
rsta
n
d
in
g
in
sp
o
k
e
n
d
ial
o
g
u
e
s
y
ste
m
s,”
Pro
c
.
4
1
st A
n
n
u
.
M
e
e
t.
Asso
c
.
C
o
mp
u
t.
L
in
g
u
ist.
,
p
p
.
2
4
0
-
2
4
7
,
2
0
0
3
.
[5
]
J.
W
il
li
a
m
s,
A
.
Ra
u
x
,
D.
Ra
m
a
c
h
a
n
d
ra
n
,
A
.
Blac
k
,
“
T
h
e
Dia
lo
g
S
tate
T
ra
c
k
in
g
Ch
a
ll
e
n
g
e
w
it
h
Ba
y
e
sia
n
A
p
p
ro
a
c
h
,
”
Pro
c
.
o
f1
4
th
An
n
u
.
M
e
e
t.
S
p
e
c
.
I
n
ter
e
s.
Gr
.
Disc
o
u
rs
e
Dia
lo
g
u
e
,
n
o
.
A
u
g
u
st,
p
p
.
4
0
4
-
4
1
3
,
2
0
1
7
.
[6
]
K.
S
u
n
,
L
.
Ch
e
n
,
S
.
Z
h
u
,
K.
Y
u
.
“
A
g
e
n
e
ra
li
z
e
d
ru
le
b
a
se
d
trac
k
e
r
f
o
r
d
ialo
g
u
e
sta
te
trac
k
in
g
,
”
Pro
c
.
S
p
o
k
.
L
a
n
g
.
T
e
c
h
n
o
l
.
W
o
rk
.
,
p
p
.
3
3
0
-
3
3
5
,
2
0
1
4
.
[7
]
S
.
L
.
a
n
d
D.
R.
T
ra
u
m
,
“
In
f
o
r
m
a
ti
o
n
sta
te
a
n
d
d
ialo
g
u
e
m
a
n
a
g
e
m
e
n
t
in
th
e
tri
n
d
i
d
ial
o
g
u
e
m
o
v
e
e
n
g
in
e
to
o
lk
it
,
”
v
o
l.
6
,
n
o
.
3
,
p
p
.
3
2
3
-
3
4
0
,
2
0
0
0
.
[8
]
D.
Bo
h
u
s
a
n
d
A
.
Ru
d
n
ick
y
,
“
K
h
y
p
o
th
e
se
s
+
o
th
e
r”
b
e
li
e
f
u
p
d
a
ti
n
g
m
o
d
e
l,
”
Pro
c
.
AA
AI
W
o
rk
.
S
t
a
t.
Emp
ir
.
Ap
p
ro
a
c
h
e
s t
o
S
p
o
k
.
Dia
l
o
g
u
e
S
y
st.
,
p
p
.
1
3
-
1
8
,
2
0
0
6
.
[9
]
Y.
M
a
,
A
.
Ra
u
x
,
D.
Ra
m
a
c
h
a
n
d
ra
n
,
a
n
d
R.
G
u
p
ta,
“
L
a
n
d
m
a
r
k
-
b
a
se
d
lo
c
a
ti
o
n
b
e
li
e
f
trac
k
in
g
in
a
sp
o
k
e
n
d
ialo
g
s
y
ste
m
,
”
S
IGD
IAL
2
0
1
2
-
1
3
t
h
A
n
n
u
.
M
e
e
t.
S
p
e
c
.
In
ter
e
s.
Gr
.
Disc
o
u
rs
e
Dia
l
o
g
u
e
,
Pro
c
.
Co
n
f.
,
p
p
.
1
6
9
-
1
7
8
,
2
0
1
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4752
Dia
lo
g
u
e
s
ta
te
tr
a
ck
in
g
a
cc
u
r
a
cy
imp
r
o
ve
men
t b
y
d
is
tin
g
u
is
h
in
g
…
(
K
h
a
ld
o
o
n
H.
A
lh
u
s
s
a
yn
i
)
1063
[1
0
]
B.
T
h
o
m
so
n
a
n
d
S
.
Y
o
u
n
g
,
“
Ba
y
e
sia
n
u
p
d
a
te
o
f
d
ialo
g
u
e
sta
t
e
:
A
P
OMDP
f
ra
m
e
w
o
rk
f
o
r
s
p
o
k
e
n
d
ial
o
g
u
e
s
y
ste
m
s,”
Co
mp
u
t.
S
p
e
e
c
h
L
a
n
g
.
,
v
o
l.
2
4
,
n
o
.
4
,
p
p
.
5
6
2
-
5
8
8
,
2
0
1
0
.
[1
1
]
A
b
h
in
a
v
Ra
sto
g
i,
Dilek
Ha
k
k
a
n
i
-
T
u
r
,
L
.
H
eck
,
“
S
c
a
lab
le
m
u
lt
i
-
d
o
m
a
in
d
ial
o
g
u
e
sta
te
trac
k
in
g
,
”
2
0
1
7
IEE
E
Au
to
m
a
ti
c
S
p
e
e
c
h
Rec
o
g
n
it
i
o
n
a
n
d
Un
d
e
rs
ta
n
d
in
g
W
o
rk
sh
o
p
(
AS
R
U),
p
p
.
5
6
1
-
5
6
8
,
2
0
1
7
.
[1
2
]
I.
Ca
sa
n
u
e
v
a
e
t
a
l.
,
“
A
b
e
n
c
h
m
a
rk
in
g
e
n
v
iro
n
m
e
n
t
f
o
r
re
in
f
o
rc
e
m
e
n
t
lea
rn
in
g
b
a
se
d
tas
k
o
rien
ted
d
ial
o
g
u
e
m
a
n
a
g
e
m
e
n
t,
”
Co
RR
,
v
o
l
.
a
b
s/
1
7
1
1
.
1
,
2
0
1
7
.
[1
3
]
M
.
He
n
d
e
rso
n
,
B.
T
h
o
m
so
n
,
a
n
d
S
.
Yo
u
n
g
,
“
De
e
p
n
e
u
ra
l
n
e
tw
o
rk
a
p
p
ro
a
c
h
f
o
r
th
e
d
ialo
g
sta
te t
ra
c
k
in
g
c
h
a
ll
e
n
g
e
,
”
S
IGD
IAL
2
0
1
3
-
1
4
th
An
n
u
.
M
e
e
t.
S
p
e
c
.
I
n
ter
e
s.
Gr
.
Disc
o
u
rs
e
Dia
lo
g
u
e
,
Pro
c
.
Co
n
f.
,
p
p
.
4
6
7
-
4
7
1
,
2
0
1
3
.
[1
4
]
M
.
He
n
d
e
rso
n
,
e
t
a
l.
,
“
W
o
rd
-
b
a
se
d
d
i
a
lo
g
sta
te
trac
k
in
g
w
it
h
re
c
u
rre
n
t
n
e
u
ra
l
n
e
tw
o
rk
s
,
”
in
Pro
c
e
e
d
in
g
s
o
f
th
e
1
5
t
h
A
n
n
u
a
l
M
e
e
ti
n
g
o
f
t
h
e
S
p
e
c
i
a
l
In
ter
e
st Gro
u
p
o
n
Disc
o
u
rs
e
a
n
d
Di
a
lo
g
u
e
(
S
IGD
IAL
)
,
2
0
1
4
,
p
p
.
2
9
2
-
2
9
9
.
[1
5
]
M
.
He
n
d
e
rso
n
,
B.
T
h
o
m
so
n
,
a
n
d
S
.
Yo
u
n
g
,
“
Ro
b
u
st
d
ial
o
g
sta
te
trac
k
in
g
u
sin
g
d
e
lex
ic
a
li
se
d
re
c
u
rre
n
t
n
e
u
ra
l
n
e
tw
o
rk
s an
d
u
n
s
u
p
e
rv
ise
d
a
d
a
p
t
a
ti
o
n
,
”
2
0
1
4
I
EE
E
S
p
o
k
e
n
L
a
n
g
u
a
g
e
T
e
c
h
n
o
l
o
g
y
W
o
rk
sh
o
p
,
p
p
.
3
6
0
-
3
6
5
,
2
0
1
4
.
[1
6
]
N.
M
rk
šić
,
e
t
a
l.
,
“
M
u
lt
i
-
d
o
m
a
in
d
ialo
g
sta
te
trac
k
in
g
u
sin
g
re
c
u
rre
n
t
n
e
u
ra
l
n
e
tw
o
rk
s,”
ACL
-
IJ
CN
L
P
2
0
1
5
-
5
3
rd
An
n
u
.
M
e
e
t.
Asso
c
.
Co
m
p
u
t
.
L
i
n
g
u
ist.
7
t
h
In
t.
J
t.
C
o
n
f
.
Na
t
.
L
a
n
g
.
Pro
c
e
ss
.
Asi
a
n
Fe
d
.
Na
t.
L
a
n
g
.
Pro
c
e
ss
.
Pro
c
.
Co
n
f.
,
v
o
l.
2
,
p
p
.
7
9
4
-
7
9
9
,
2
0
1
5
.
[1
7
]
J.
P
e
re
z
a
n
d
F
.
L
iu
,
“
Dia
lo
g
st
a
te
trac
k
in
g
,
a
m
a
c
h
in
e
re
a
d
in
g
a
p
p
r
o
a
c
h
u
sin
g
M
e
m
o
r
y
Ne
t
w
o
rk
,
”
Co
RR
,
v
o
l.
a
b
s/1
6
0
6
.
0
4
0
5
2
,
2
0
1
6
.
[1
8
]
T
.
W
e
n
,
e
t
a
l.
,
“
A
n
e
t
w
o
rk
-
b
a
se
d
e
n
d
-
to
-
e
n
d
train
a
b
le
tas
k
-
o
rien
ted
d
ialo
g
u
e
sy
st
e
m
,”
Pro
c
e
e
d
in
g
s
o
f
th
e
1
5
t
h
Co
n
fer
e
n
c
e
o
f
th
e
E
u
ro
p
e
a
n
C
h
a
p
ter
o
f
t
h
e
Asso
c
i
a
ti
o
n
f
o
r Co
m
p
u
ta
ti
o
n
a
l
L
i
n
g
u
isti
c
s,
v
o
l
.
1
,
p
p
.
4
3
8
-
4
4
9
,
2
0
1
7
.
[1
9
]
L
.
Zi
lk
a
a
n
d
F
.
Ju
rc
ice
k
,
“
In
c
re
m
e
n
tal
L
S
T
M
-
b
a
se
d
d
ialo
g
sta
te
trac
k
e
r,
”
2
0
1
5
IEE
E
W
o
rk
.
Au
to
m.
S
p
e
e
c
h
Rec
o
g
n
it
.
Un
d
e
rs
ta
n
d
in
g
,
AS
RU
2
0
1
5
-
Pro
c
.
,
v
o
l.
1
,
n
o
.
L
,
p
p
.
7
5
7
-
7
6
2
,
2
0
1
6
.
[2
0
]
M
.
He
n
d
e
rso
n
,
B.
T
h
o
m
so
n
,
a
n
d
J.
W
il
li
a
m
s,
“
Dia
lo
g
S
tate
T
r
a
c
k
in
g
Ch
a
ll
e
n
g
e
2
&
3
,
”
Di
a
l
o
g
u
e
s
wit
h
S
o
c
.
Ro
b
o
t.
,
n
o
.
S
e
p
tem
b
e
r,
p
p
.
1
-
2
2
,
2
0
1
3
.
[2
1
]
P
.
X
u
a
n
d
Q.
Hu
,
“
A
n
e
n
d
-
to
-
e
n
d
a
p
p
ro
a
c
h
f
o
r
h
a
n
d
li
n
g
u
n
k
n
o
w
n
slo
t
v
a
lu
e
s
in
d
ialo
g
u
e
s
tate
trac
k
in
g
,
”
Pro
c
e
e
d
in
g
s
o
f
t
h
e
5
6
t
h
An
n
u
a
l
M
e
e
ti
n
g
o
f
th
e
Asso
c
i
a
ti
o
n
f
o
r Co
mp
u
ta
ti
o
n
a
l
L
i
n
g
u
isti
c
s
,
2
0
1
8
.
[2
2
]
H.
S
e
p
p
a
n
d
S
.
J
u
rg
e
n
,
“
L
o
n
g
sh
o
rt
-
term
m
e
m
o
r
y
,
”
Ne
u
ra
l
Co
mp
u
t
.
,
v
o
l.
9
,
n
o
.
8
,
p
p
.
1
7
3
5
-
1
7
8
0
,
1
9
9
7
.
[2
3
]
Z.
L
in
e
t
a
l.
,
“
A
stru
c
tu
re
d
s
e
lf
-
a
tt
e
n
ti
v
e
s
e
n
ten
c
e
e
m
b
e
d
d
i
n
g
,
”
5
th
I
n
ter
n
a
t
io
n
a
l
Co
n
fer
e
n
c
e
o
n
L
e
a
rn
i
n
g
Rep
re
se
n
ta
ti
o
n
s (
ICL
R
2
0
1
7
),
p
p
.
1
-
1
5
,
2
0
1
7
.
[2
4
]
S
.
S
h
a
rm
a
,
P
.
K.
Ch
o
u
b
e
y
,
a
n
d
R.
Hu
a
n
g
,
“
Im
p
ro
v
in
g
Dia
lo
g
u
e
S
tate
T
ra
c
k
in
g
b
y
Disc
e
rn
in
g
th
e
Re
lev
a
n
t
Co
n
tex
t,
”
Pro
c
e
e
d
in
g
s
o
f
NAA
C
L
-
HLT
2
0
1
9
,
p
p.
5
7
6
-
5
8
1
,
2
0
1
9
.
[2
5
]
C.
D.
M
.
Je
f
f
re
y
P
e
n
n
in
g
to
n
,
Rich
a
rd
S
o
c
h
e
r,
“
G
lo
V
e
:
G
lo
b
a
l
V
e
c
to
rs
f
o
r
W
o
rd
Re
p
re
se
n
tatio
n
,
”
P
ro
c
.
2
0
1
4
C
o
n
-
fer
e
n
c
e
Emp
ir.
M
e
th
o
d
s N
a
t.
L
a
n
g
.
Pro
c
e
ss
.
,
p
p
.
1
5
3
2
-
1
5
4
3
,
2
0
1
4
.
[2
6
]
D.
P
.
Kin
g
m
a
a
n
d
J.
Ba
,
“
A
d
a
m
:
A
M
e
th
o
d
f
o
r
S
to
c
h
a
stic
Op
ti
m
iz
a
ti
o
n
,
”
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
L
e
a
r
n
in
g
Rep
re
se
n
ta
ti
o
n
s
,
p
p
.
1
-
1
5
,
2
0
1
4
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
K
h
a
ld
o
o
n
H
a
sa
n
Al
h
u
ss
a
y
n
i
re
c
e
iv
e
d
th
e
BS
c
d
e
g
re
e
s
in
Co
m
p
u
ter
S
c
ien
c
e
f
ro
m
th
e
Un
iv
e
rsit
y
o
f
Ba
b
y
lo
n
,
Ira
q
,
in
2
0
0
8
.
A
f
ter
c
o
m
p
letin
g
h
is
BS
c
,
h
e
w
o
rk
e
d
a
s
a
p
ro
g
ra
m
m
e
r
a
t
th
e
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
Ce
n
ter,
th
e
Un
iv
e
rsity
o
f
B
a
b
y
lo
n
.
In
2
0
0
9
,
re
c
e
iv
e
d
th
e
M
S
c
d
e
g
re
e
s
in
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
Co
m
p
u
ter
En
g
in
e
e
rin
g
f
ro
m
th
e
T
u
la
S
tate
Un
iv
e
rsit
y
,
Ru
ss
ian
,
in
2
0
1
4
.
S
e
c
in
c
e
No
v
2
0
1
6
h
e
e
n
tere
d
th
e
T
o
m
sk
sta
te u
n
iv
e
rsit
y
(
T
S
U),
De
p
a
rt
m
e
n
t
o
f
c
o
m
p
u
ter sc
ien
c
e
th
e
o
ry
a
s a
P
h
.
D.
stu
d
e
n
t
in
sp
e
c
ialty
M
a
th
e
m
a
ti
c
a
l
a
n
d
so
f
tw
a
r
e
o
f
c
o
m
p
u
ters
,
co
m
p
lex
e
s
a
n
d
c
o
m
p
u
ter
n
e
tw
o
r
k
s.
His
m
a
in
re
s
e
a
rc
h
in
tere
sts
a
re
m
a
c
h
in
e
lea
rn
in
g
,
n
e
u
ra
l
n
e
tw
o
rk
s,
d
e
e
p
lea
rn
in
g
,
d
ialo
g
u
e
sy
ste
m
s an
d
n
a
tu
ra
l
lan
g
u
a
g
e
p
ro
c
e
ss
in
g
.
Z
a
m
y
a
t
i
n
A
l
e
x
a
n
d
e
r
V
l
a
d
im
i
r
o
v
i
c
h
D
.
S
c
.
,
a
s
s
o
c
i
a
t
e
p
r
o
f
e
s
s
o
r
,
H
e
a
d
o
f
D
e
p
a
r
t
m
e
n
t
o
f
c
o
m
p
u
t
e
r
s
c
i
e
n
c
e
t
h
e
o
r
y
i
n
t
h
e
T
o
m
s
k
S
t
a
t
e
U
n
s
i
v
e
r
s
i
t
y
(
T
S
U
)
.
I
n
D
e
c
e
m
b
e
r
2
0
0
2
h
e
e
n
t
e
r
e
d
t
h
e
f
u
l
l
-
t
i
m
e
p
o
s
t
g
r
a
d
u
a
t
e
s
t
u
d
y
o
f
T
o
m
s
k
P
o
l
y
t
e
c
h
n
i
c
U
n
i
v
e
r
s
i
t
y
(
T
P
U
)
i
n
s
p
e
c
i
a
l
t
y
0
5
.
1
3
.
1
1
-
M
a
t
h
e
m
a
t
i
c
a
l
a
n
d
s
o
f
t
w
a
r
e
o
f
c
o
m
p
u
t
e
r
s
,
c
o
m
p
l
e
x
e
s
a
n
d
c
o
m
p
u
t
e
r
n
e
t
w
o
r
k
s
,
w
h
i
c
h
h
e
g
r
a
d
u
a
t
e
d
a
h
e
a
d
o
f
s
c
h
e
d
u
l
e
d
u
e
t
o
w
i
t
h
t
h
e
d
e
f
e
n
s
e
o
f
h
i
s
t
h
e
s
i
s
.
A
s
a
n
a
s
s
o
c
i
a
t
e
P
r
o
f
e
s
s
o
r
o
f
t
h
e
D
e
p
a
r
t
m
e
n
t
o
f
c
o
m
p
u
t
e
r
s
c
i
e
n
c
e
(
p
a
r
t
-
t
i
m
e
)
c
o
n
t
i
n
u
e
d
t
o
c
a
r
r
y
o
u
t
e
d
u
c
a
t
i
o
n
a
l
a
n
d
r
e
s
e
a
r
c
h
a
c
t
i
v
i
t
i
e
s
,
F
r
o
m
J
a
n
u
a
r
y
2
0
1
9
h
e
w
a
s
a
p
p
o
i
n
t
e
d
a
c
t
i
n
g
D
i
r
e
c
t
o
r
o
f
t
h
e
I
n
s
t
i
t
u
t
e
o
f
a
p
p
l
i
e
d
m
a
t
h
e
m
a
t
i
c
s
a
n
d
c
o
m
p
u
t
e
r
s
c
i
e
n
c
e
.
Z
a
m
y
a
t
i
n
A
.
V
.
c
o
m
p
l
e
t
e
d
a
s
t
u
d
y
a
n
d
w
a
s
e
n
t
i
t
l
e
d
t
h
e
t
e
a
c
h
e
r
o
f
t
h
e
h
i
g
h
s
c
h
o
o
l
I
G
I
P
I
n
t
e
r
n
a
t
i
o
n
a
l
E
n
g
i
n
e
e
r
i
n
g
E
d
u
c
a
t
o
r
“
I
N
G
-
P
A
E
D
I
G
I
P
”
,
N
o
E
N
G
-
2
5
5
,
i
s
a
c
t
i
v
e
l
y
i
n
v
o
l
v
e
d
i
n
s
c
i
e
n
t
i
f
i
c
r
e
s
e
a
r
c
h
,
p
a
r
t
i
c
i
p
a
n
t
r
e
s
e
a
r
c
h
a
s
t
h
e
h
e
a
d
(
R
u
s
s
i
a
n
f
o
u
n
d
a
t
i
o
n
f
o
r
b
a
s
i
c
r
e
s
e
a
r
c
h
g
r
a
n
t
s
N
o
1
1
-
07
-
0
0
0
2
7
,
0
3
-
07
-
0
6
0
2
4
,
1
4
-
07
-
0
0
1
2
7
,
Г
К
N
o
1
4
.
5
1
5
.
1
1
.
0
0
4
7
o
n
t
h
e
i
n
s
t
r
u
c
t
i
o
n
s
o
f
M
i
n
i
s
t
r
y
o
f
e
d
u
c
a
t
i
o
n
a
n
d
s
c
i
e
n
c
e
o
f
t
h
e
R
u
s
s
i
a
n
F
e
d
e
r
a
t
i
o
n
)
a
n
d
r
e
s
p
o
n
s
i
b
l
e
o
f
f
i
c
e
r
(
g
r
a
n
t
s
R
F
B
R
№
№
0
0
-
07
-
9
0
1
2
4
,
0
3
-
07
-
9
0
1
2
4
,
0
6
-
05
-
78056
)
.
E
x
e
c
u
t
o
r
f
o
r
t
h
e
T
S
U
f
o
r
t
h
e
m
e
g
a
-
p
r
o
j
e
c
t
"
C
r
e
a
t
i
n
g
o
f
l
o
c
a
l
h
i
g
h
-
t
e
c
h
s
o
f
t
w
a
r
e
a
n
d
i
n
s
t
r
u
m
e
n
t
a
l
c
o
m
p
l
e
x
f
o
r
t
h
e
i
m
p
l
e
m
e
n
t
a
t
i
o
n
o
f
p
r
o
c
e
s
s
c
o
n
t
r
o
l
s
y
s
t
e
m
s
b
a
s
e
d
o
n
f
r
e
e
s
o
f
t
w
a
r
e
”
c
a
r
r
i
e
d
o
u
t
i
n
c
o
l
l
a
b
o
r
a
t
i
o
n
w
i
t
h
"
E
l
e
S
y
"
c
o
m
p
a
n
y
a
n
d
T
o
m
s
k
s
t
a
t
e
u
n
i
v
e
r
s
i
t
y
.
A
c
c
o
r
d
i
n
g
t
o
t
h
e
r
e
s
u
l
t
s
o
f
s
c
i
e
n
t
i
f
i
c
a
c
t
i
v
i
t
y
Z
a
m
y
a
t
i
n
p
u
b
l
i
s
h
e
d
m
o
r
e
t
h
a
n
7
0
s
c
i
e
n
t
i
f
i
c
p
a
p
e
r
s
,
i
n
c
l
u
d
i
n
g
2
m
o
n
o
g
r
a
p
h
s
.
Z
a
m
y
a
t
i
n
A
.
V
.
i
s
a
d
e
v
e
l
o
p
e
r
a
n
d
h
e
a
d
o
f
t
h
e
m
a
s
t
e
r
'
s
e
d
u
c
a
t
i
o
n
a
l
p
r
o
g
r
a
m
"
D
a
t
a
s
c
i
e
n
c
e
a
n
d
b
i
o
i
n
f
o
r
m
a
t
i
c
s
"
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
21
,
No
.
2
,
Feb
r
u
ar
y
2
0
2
1
:
10
57
-
10
64
1064
E
m
a
n
Al
-
S
h
a
m
e
r
y
r
e
c
e
iv
e
d
th
e
BS
c
a
n
d
M
S
c
d
e
g
re
e
s in
Co
m
p
u
ter S
c
ien
c
e
f
ro
m
th
e
Un
iv
e
rsit
y
o
f
Ba
b
y
lo
n
,
Ira
q
,
i
n
1
9
9
8
a
n
d
2
0
0
1
,
re
sp
e
c
ti
v
e
l
y
.
Af
ter
c
o
m
p
letin
g
h
e
r
M
S
c
,
s
h
e
w
o
rk
e
d
a
s
a
n
a
ss
istan
t
lec
tu
re
r
a
t
th
e
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
,
th
e
U
n
iv
e
rsit
y
o
f
Ba
b
y
lo
n
.
In
2
0
1
3
,
sh
e
re
c
e
iv
e
d
h
e
r
P
h
D
in
Co
m
p
u
te
r
S
c
ien
c
e
f
ro
m
th
e
Un
iv
e
rsit
y
o
f
Ba
b
y
lo
n
.
Cu
rre
n
t
ly
,
sh
e
h
o
l
d
s
a
p
ro
f
e
ss
o
r
p
o
siti
o
n
a
t
S
o
f
twa
r
e
De
p
a
rt
m
e
n
t,
Un
iv
e
rsit
y
o
f
B
a
b
y
lo
n
.
He
r
c
u
rre
n
t
re
se
a
r
c
h
in
tere
sts
in
c
lu
d
e
a
rti
f
icia
l
in
tell
ig
e
n
c
e
,
b
io
in
f
o
rm
a
ti
c
s,
m
a
c
h
in
e
lea
rn
in
g
,
n
e
u
ra
l
n
e
tw
o
rk
s,
d
e
e
p
lea
rn
in
g
a
n
d
d
a
ta m
in
in
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.