C
omp
u
te
r
S
c
i
e
n
c
e
an
d
I
n
for
mati
on
T
e
c
h
n
ol
ogi
e
s
V
ol
.
6
,
N
o
.
3
,
N
ove
m
b
e
r
20
25
,
p
p.
253
~
261
IS
S
N
:
2722
-
3221
,
D
O
I
:
10.
1
1591
/
c
s
i
t
.
v
6
i
3
.
pp
25
3
-
261
253
Jou
r
n
al
h
o
m
e
pa
ge
:
ht
t
p:
/
/
i
ae
s
pr
i
m
e
.
c
om
/
i
nd
e
x
.
php
/
c
s
i
t
J
a
v
a
n
e
s
e
a
n
d
S
u
n
d
a
n
e
s
e
sp
e
e
c
h
r
e
c
o
g
n
i
t
i
o
n
u
si
n
g
W
h
i
sp
e
r
A
l
i
m
R
ah
ar
j
o
,
A
ma
l
i
a
Zah
r
a
D
e
p
a
rt
m
e
n
t
o
f
Co
m
p
u
t
e
r
S
c
i
e
n
c
e
,
BIN
U
S
G
ra
d
u
a
t
e
P
ro
g
ra
m
-
M
a
s
t
e
r
o
f
Co
m
p
u
t
e
r
S
c
i
e
n
c
e
,
Bi
n
a
N
u
s
a
n
t
a
ra
U
n
i
v
e
rs
i
t
y
,
J
a
k
a
rt
a
,
In
d
o
n
e
s
i
a
A
r
ti
c
l
e
I
n
fo
A
BS
TR
A
C
T
Ar
t
i
c
l
e
h
i
s
t
or
y
:
Re
c
e
i
v
e
d
N
ov
12
,
202
4
Re
vi
s
e
d
M
a
y
27
,
20
25
A
c
c
e
pt
e
d
J
un
13
,
2025
A
ut
o
m
a
t
i
c
s
pe
e
c
h
r
e
c
ogn
i
t
i
on
(
A
S
R
)
t
e
c
h
nol
ogy
is
e
s
s
e
n
t
i
a
l
f
o
r
a
dva
nc
i
n
g
hum
a
n
-
c
o
m
pu
t
e
r
i
n
t
e
r
a
c
t
i
on
,
p
a
r
t
i
c
ul
a
r
l
y
in
a
l
i
ng
ui
s
t
i
c
a
l
l
y
d
i
v
e
r
s
e
c
o
unt
r
y
l
i
k
e
I
nd
one
s
i
a
,
w
he
r
e
a
p
pr
o
xi
m
a
t
e
l
y
700
n
a
t
i
ve
l
a
ng
ua
g
e
s
a
r
e
s
po
ke
n
,
i
nc
l
ud
i
ng
w
i
d
e
l
y
us
e
d
l
a
ngu
a
ge
s
l
i
k
e
J
a
va
n
e
s
e
a
nd
S
und
a
ne
s
e
.
T
hi
s
s
t
udy
l
e
v
e
r
a
ge
s
t
he
pr
e
-
t
r
a
i
n
e
d
W
h
i
s
p
e
r
S
m
a
l
l
m
o
de
l
an
e
n
d‑
t
o
‑
e
nd
t
r
a
ns
f
o
r
m
e
r
pr
e
t
r
a
i
n
e
d
on
6
80
,
000
h
our
s
of
m
u
l
t
i
l
i
n
gua
l
s
p
e
e
c
h,
f
i
ne
t
uni
n
g
it
s
pe
c
i
f
i
c
a
l
l
y
to
i
m
p
r
ov
e
A
S
R
p
e
r
f
o
r
m
a
nc
e
f
or
t
h
e
s
e
l
ow
r
e
s
o
ur
c
e
l
a
ngua
ge
s
.
T
he
pr
i
m
a
r
y
goa
l
is
to
i
n
c
r
e
a
s
e
t
r
a
n
s
c
r
i
p
t
i
on
a
c
c
ur
a
c
y
a
nd
r
e
l
i
a
bi
l
i
t
y
f
o
r
J
a
va
ne
s
e
a
nd
S
und
a
ne
s
e
,
w
h
i
c
h
h
a
ve
h
i
s
t
or
i
c
a
l
l
y
ha
d
l
i
m
i
t
e
d
A
S
R
r
e
s
ou
r
c
e
s
.
A
ppr
ox
i
m
a
t
e
l
y
100
h
ou
r
s
of
s
pe
e
c
h
f
r
o
m
O
pe
nS
L
R
w
e
r
e
s
e
l
e
c
t
e
d,
c
o
ve
r
i
n
g
bot
h
r
e
a
di
ng
a
nd
c
onv
e
r
s
a
t
i
on
a
l
pr
om
pt
s
,
t
h
e
da
t
a
e
xh
i
b
i
t
e
d
d
i
a
l
e
c
t
a
l
va
r
i
a
t
i
o
n,
a
m
b
i
e
n
t
no
i
s
e
,
a
nd
i
n
c
o
m
p
l
e
t
e
d
e
m
og
r
a
ph
i
c
m
e
t
a
d
a
t
a
,
ne
c
e
s
s
i
t
a
t
i
ng
n
or
m
a
l
i
z
a
t
i
on
a
nd
f
i
x
e
d
‑
l
e
ng
t
h
pa
d
di
n
g.
w
i
t
h
m
od
e
l
e
va
l
u
a
t
i
on
ba
s
e
d
on
t
h
e
w
o
r
d
e
r
r
o
r
r
a
t
e
(
W
E
R
)
m
e
t
r
i
c
.
U
n
l
i
ke
a
pp
r
o
a
c
he
s
t
h
a
t
c
o
m
b
i
n
e
s
e
pa
r
a
t
e
a
c
ou
s
t
i
c
e
nc
od
e
r
s
w
i
t
h
e
x
t
e
r
na
l
l
a
ngu
a
ge
m
od
e
l
s
,
W
h
i
s
p
e
r
un
i
f
i
e
d
a
r
c
h
i
t
e
c
t
u
r
e
s
t
r
e
a
m
l
i
ne
s
a
d
a
p
t
a
t
i
o
n
f
or
l
ow
‑
r
e
s
ou
r
c
e
s
e
t
t
i
ngs
.
E
v
a
l
ua
t
e
d
on
he
l
d
‑
ou
t
t
e
s
t
s
e
t
s
,
t
he
f
i
ne
‑
t
une
d
m
od
e
l
s
a
c
h
i
e
v
e
d
W
o
r
d
E
r
r
or
R
a
t
e
s
of
14.
9
7%
f
o
r
J
a
va
ne
s
e
a
n
d
2
.
03
%
f
or
S
und
a
ne
s
e
,
s
ubs
t
a
n
t
i
a
l
l
y
ou
t
p
e
r
f
or
m
i
ng
ba
s
e
l
i
ne
s
y
s
t
e
m
s
.
T
he
s
e
r
e
s
ul
t
s
d
e
m
ons
t
r
a
t
e
W
h
i
s
pe
r
e
f
f
e
c
t
i
v
e
ne
s
s
in
l
ow
‑
r
e
s
ou
r
c
e
A
S
R
a
nd
hi
g
hl
i
gh
t
i
t
s
po
t
e
n
t
i
a
l
to
e
nha
nc
e
t
r
a
ns
c
r
i
p
t
i
on
a
c
c
u
r
a
c
y,
s
uppo
r
t
l
a
n
gua
ge
pr
e
s
e
r
v
a
t
i
on
,
a
nd
b
r
oa
de
n
d
i
g
i
t
a
l
a
c
c
e
s
s
f
o
r
unde
r
r
e
pr
e
s
e
nt
e
d
s
pe
e
c
h
c
o
m
m
uni
t
i
e
s
.
Ke
y
w
or
d
s
:
F
i
ne
t
un
e
J
a
v
a
n
e
s
e
S
pe
e
c
h
r
e
c
ogni
t
i
o
n
S
unda
n
e
s
e
W
hi
s
p
e
r
T
hi
s
i
s
an
op
e
n
ac
c
e
s
s
ar
t
i
c
l
e
u
nde
r
t
he
C
C
B
Y
-
SA
l
i
c
e
n
s
e
.
Cor
r
e
s
pon
di
n
g
Au
t
h
or
:
A
l
i
m
R
a
ha
r
j
o
D
e
pa
r
t
m
e
nt
of
Co
m
pu
t
e
r
S
c
i
e
n
c
e
,
BIN
U
S
G
r
a
dua
t
e
P
rogr
a
m
-
M
a
s
t
e
r
of
Co
m
pu
t
e
r
S
c
i
e
n
c
e
Bi
na
N
us
a
n
t
a
ra
U
ni
v
e
rs
i
t
y
D
K
I
J
a
k
a
rt
a
,
Indo
ne
s
i
a
,
114
80
E
m
a
i
l
:
a
l
i
m
.
r
a
ha
rj
o
@
bi
n
us
.
a
c
.
i
d
1.
I
N
TR
O
D
U
C
TI
O
N
A
ut
om
a
t
i
c
s
p
e
e
c
h
re
c
ogn
i
t
i
on
(A
S
R)
i
s
a
pi
vot
a
l
t
e
c
hn
ol
ogy
i
n
hu
m
a
n
-
c
o
m
put
e
r
i
nt
e
ra
c
t
i
on
,
a
l
l
ow
i
ng
s
e
a
m
l
e
s
s
c
o
m
m
uni
c
a
t
i
on
b
e
t
w
e
e
n
hu
m
a
ns
a
nd
m
a
c
hi
n
e
s
t
hrou
gh
s
p
e
e
c
h
[1]
.
O
ve
r
t
i
m
e
,
A
S
R
h
a
s
s
e
e
n
s
i
gn
i
fi
c
a
nt
a
dva
n
c
e
m
e
nt
s
,
l
e
a
di
ng
t
o
t
he
i
nt
e
gra
t
i
on
of
s
pe
e
c
h
i
n
t
e
r
fa
c
e
s
i
n
a
w
i
d
e
ra
ng
e
of
a
ppl
i
c
a
t
i
ons
,
s
uc
h
a
s
t
ra
ns
c
ri
p
t
i
on
s
e
r
vi
c
e
s
,
v
i
rt
u
a
l
a
s
s
i
s
t
a
n
t
s
,
a
nd
vo
i
c
e
-
c
o
nt
rol
l
e
d
d
e
vi
c
e
s
[2]
–
[
8]
.
T
he
gl
o
ba
l
A
S
R
m
a
rke
t
i
s
e
x
pe
c
t
e
d
t
o
grow
s
u
bs
t
a
n
t
i
a
l
l
y,
dr
i
ve
n
by
t
he
i
n
c
re
a
s
i
ng
d
e
m
a
nd
for
e
ff
i
c
i
e
n
t
a
nd
a
c
c
u
ra
t
e
s
pe
e
c
h
re
c
og
ni
t
i
on
t
e
c
hnol
ogi
e
s
[9
]
.
L
a
ngu
a
g
e
di
v
e
rs
i
t
y
i
s
o
ne
of
t
h
e
c
or
e
c
ha
ra
c
t
e
ri
s
t
i
c
s
of
hu
m
a
n
c
i
v
i
l
i
z
a
t
i
on,
w
i
t
h
ove
r
7,
0
00
l
a
ng
ua
g
e
s
s
pok
e
n
w
orl
dw
i
de
,
s
h
ow
c
a
s
i
ng
t
h
e
c
om
p
l
e
xi
t
y
of
hu
m
a
n
c
om
m
un
i
c
a
t
i
on
[1
0]
.
L
a
ng
ua
g
e
go
e
s
be
yond
s
i
m
p
l
e
c
o
m
m
un
i
c
a
t
i
on
;
i
t
e
n
c
a
ps
ul
a
t
e
s
w
orl
dv
i
e
w
s
,
c
u
l
t
u
ra
l
e
x
pre
s
s
i
o
ns
,
a
nd
un
i
qu
e
gra
m
m
a
t
i
c
a
l
a
nd
phon
e
t
i
c
s
ys
t
e
m
s
.
I
ndon
e
s
i
a
,
w
i
t
h
i
t
s
popu
l
a
t
i
on
of
a
pprox
i
m
a
t
e
l
y
280
m
i
l
l
i
on
,
i
s
ho
m
e
t
o
a
round
70
0
n
a
t
i
ve
l
a
ng
ua
g
e
s
,
i
n
a
dd
i
t
i
on
t
o
t
h
e
na
t
i
o
na
l
l
a
ngu
a
ge
,
Ba
ha
s
a
In
done
s
i
a
[10]
.
T
hi
s
m
a
ke
s
Indone
s
i
a
o
ne
of
t
h
e
m
os
t
l
i
ngu
i
s
t
i
c
a
l
l
y
di
v
e
rs
e
na
t
i
ons
i
n
t
he
w
o
rl
d
,
a
c
c
ount
i
ng
f
or
roug
hl
y
10%
of
t
h
e
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
6
,
N
o
.
3
,
N
ov
e
m
be
r
20
25
:
253
-
261
254
gl
oba
l
l
a
ng
ua
g
e
s
.
J
a
va
n
e
s
e
a
nd
S
unda
ne
s
e
,
s
po
ke
n
by
a
p
proxi
m
a
t
e
l
y
90
m
i
l
l
i
on
a
nd
40
m
i
l
l
i
o
n
p
e
opl
e
re
s
pe
c
t
i
ve
l
y,
a
re
t
h
e
s
e
c
on
d
a
nd
t
h
i
rd
m
os
t
w
i
de
l
y
s
po
ke
n
l
a
ngua
g
e
s
i
n
I
ndon
e
s
i
a
[11]
.
Re
c
e
nt
a
d
va
n
c
e
m
e
nt
s
i
n
A
S
R
for
In
done
s
i
a
n
l
a
ngu
a
ge
s
ha
v
e
b
e
e
n
dr
i
ve
n
by
s
e
l
f
-
s
u
pe
rv
i
s
e
d
l
e
a
rni
n
g
m
od
e
l
s
,
no
t
a
b
l
y
W
a
v2V
e
c
2
[12]
,
w
hi
c
h
ha
s
s
h
o
w
n
pro
m
i
s
e
i
n
l
ow
-
r
e
s
ourc
e
l
a
n
gua
ge
s
e
t
t
i
n
gs
.
S
t
udi
e
s
ha
v
e
r
e
por
t
e
d
w
ord
e
rro
r
r
a
t
e
s
(W
E
R)
a
s
l
ow
a
s
22%
i
n
s
o
m
e
a
pp
l
i
c
a
t
i
ons
[13]
–
[
22]
.
N
ot
a
bl
y,
re
s
e
a
rc
h
on
S
un
da
n
e
s
e
A
S
R
ut
i
l
i
z
i
ng
t
h
e
O
p
e
nS
L
R
da
t
a
s
e
t
a
c
hi
e
ve
d
a
W
E
R
of
23
.
5%
t
hrough
f
i
n
e
-
t
un
i
ng
of
t
he
W
a
v2V
e
c
2
m
od
e
l
[
23]
.
S
i
m
i
l
a
rl
y,
a
no
t
he
r
s
t
udy
on
J
a
va
ne
s
e
A
S
R,
l
e
ve
r
a
gi
ng
t
he
X
L
S
-
R
v
a
ri
a
n
t
of
W
a
v
2V
e
c
2,
a
c
hi
e
ve
d
a
W
E
R
of
17
.
95
%
[24
]
.
T
h
e
s
e
r
e
s
ul
t
s
hi
gh
l
i
g
ht
t
h
e
e
f
fe
c
t
i
ve
n
e
s
s
of
s
e
l
f
-
s
up
e
rv
i
s
e
d
m
ode
l
s
i
n
t
a
c
k
l
i
ng
t
he
c
ha
l
l
e
nge
s
pos
e
d
by
l
ow
-
r
e
s
ourc
e
l
a
ngu
a
ge
s
w
i
t
h
l
i
m
i
t
e
d
t
ra
i
ni
ng
d
a
t
a
.
T
hi
s
s
t
udy
foc
us
e
s
o
n
fi
n
e
-
t
uni
n
g
t
h
e
W
hi
s
p
e
r
m
od
e
l
[7
],
a
s
t
a
t
e
-
of
-
t
he
-
a
rt
A
S
R
m
od
e
l
,
t
o
e
va
l
ua
t
e
i
t
s
p
e
rfor
m
a
nc
e
for
l
ow
-
re
s
ou
rc
e
l
a
n
gua
g
e
s
,
s
p
e
c
i
fi
c
a
l
l
y
J
a
v
a
n
e
s
e
a
n
d
S
u
nda
n
e
s
e
.
T
he
d
a
t
a
s
e
t
s
us
e
d
a
re
s
o
u
rc
e
d
fro
m
O
p
e
nS
L
R
[11],
a
re
p
os
i
t
o
ry
of
fe
r
i
ng
va
l
ua
b
l
e
r
e
s
ourc
e
s
for
l
a
ngu
a
g
e
s
w
i
t
h
l
i
m
i
t
e
d
s
p
e
e
c
h
d
a
t
a
.
W
hi
s
p
e
r’s
e
xt
e
ns
i
v
e
pre
-
t
r
a
i
ni
ng
on
a
da
t
a
s
e
t
of
ove
r
680
,
000
h
ours
of
a
udi
o
-
t
r
a
ns
c
r
i
pt
i
on
d
a
t
a
,
i
nc
l
ud
i
ng
11
7,
0
00
m
ul
t
i
l
i
ngu
a
l
hours
c
ov
e
ri
ng
m
ore
t
h
a
n
96
l
a
ngu
a
g
e
s
[7]
,
pos
i
t
i
ons
i
t
a
s
a
s
t
rong
c
a
ndi
d
a
t
e
for
t
hi
s
t
a
s
k
.
A
r
e
c
e
nt
s
t
udy
[25]
–
[27]
c
o
m
pa
ri
ng
W
h
i
s
pe
r
t
o
W
a
v2V
e
c
2.
0
foun
d
t
h
a
t
W
hi
s
p
e
r
c
ons
i
s
t
e
nt
l
y
ou
t
p
e
rfor
m
s
W
a
v2V
e
c
2
.
0
a
c
ros
s
m
u
l
t
i
pl
e
l
a
ng
ua
g
e
s
,
a
c
hi
e
vi
ng
l
ow
e
r
W
E
Rs
,
e
s
pe
c
i
a
l
l
y
i
n
no
i
s
y
e
nvi
ron
m
e
n
t
s
s
uc
h
a
s
t
he
G
RA
C
E
[2
5]
a
nd
CO
RA
A
[28]
c
orpo
ra
.
F
or
i
ns
t
a
n
c
e
,
W
hi
s
pe
r
yi
e
l
d
e
d
a
n
a
ve
r
a
g
e
W
E
R
ra
nge
of
11.
3
%
t
o
2
4.
9
%,
no
t
a
bl
y
out
pe
rfor
m
i
ng
W
a
v2V
e
c
2
.
0,
w
hi
c
h
ra
nge
d
fro
m
13
.
1%
t
o
34.
8%
i
n
s
i
m
i
l
a
r
c
on
di
t
i
ons
[2
5].
T
h
e
s
t
a
t
i
s
t
i
c
a
l
s
i
gni
f
i
c
a
nc
e
of
W
h
i
s
pe
r’s
p
e
rfor
m
a
n
c
e
,
a
s
v
e
ri
f
i
e
d
by
a
M
a
nn
-
W
h
i
t
n
e
y
t
e
s
t
,
u
nde
r
l
i
n
e
s
i
t
s
a
d
a
p
t
a
b
i
l
i
t
y
a
n
d
e
ffi
c
i
e
nc
y
i
n
di
ve
rs
e
m
ul
t
i
l
i
ngu
a
l
s
c
e
na
r
i
os
.
U
nl
i
ke
W
a
v2V
e
c
2.
0
,
w
h
i
c
h
r
e
l
y
on
u
ns
upe
r
vi
s
e
d
pr
e
-
t
r
a
i
n
i
ng
on
unl
a
b
e
l
l
e
d
da
t
a
,
W
hi
s
p
e
r
b
e
ne
fi
t
s
from
i
t
s
pre
-
t
r
a
i
n
i
ng
di
r
e
c
t
l
y
on
s
pe
e
c
h
-
to
-
t
e
xt
t
a
s
ks
,
e
na
b
l
i
ng
i
t
t
o
ge
ne
r
a
l
i
z
e
w
e
l
l
a
c
r
os
s
di
ve
rs
e
l
a
ng
ua
g
e
s
,
do
m
a
i
ns
,
a
n
d
d
a
t
a
s
e
t
s
.
W
hi
s
p
e
r
us
e
of
l
a
b
e
l
e
d
da
t
a
a
l
l
ow
s
i
t
t
o
p
e
rfor
m
A
S
R
t
a
s
ks
w
i
t
h
m
i
n
i
m
a
l
fi
n
e
-
t
uni
n
g
c
om
pa
r
e
d
t
o
m
ode
l
s
t
h
a
t
r
e
qu
i
re
e
x
t
e
ns
i
ve
f
i
ne
-
t
un
i
ng
fo
r
opt
i
m
a
l
pe
rfo
r
m
a
n
c
e
.
It
s
pre
-
t
ra
i
ni
ng
on
a
di
ve
rs
e
m
u
l
t
i
l
i
ngua
l
da
t
a
s
e
t
m
a
k
e
s
i
t
pa
r
t
i
c
ul
a
rl
y
s
ui
t
e
d
fo
r
u
nde
rr
e
pr
e
s
e
n
t
e
d
l
a
n
gua
g
e
s
l
i
ke
J
a
v
a
n
e
s
e
a
nd
S
und
a
n
e
s
e
,
w
h
i
c
h
l
a
c
k
s
uffi
c
i
e
n
t
A
S
R
m
od
e
l
s
.
F
ur
t
he
rm
o
re
,
W
hi
s
p
e
r
a
r
c
hi
t
e
c
t
ur
e
a
t
ra
ns
fo
rm
e
r
b
a
s
e
d
e
nc
od
e
r
-
d
e
c
ode
r
m
ode
l
—
f
a
c
i
l
i
t
a
t
e
s
a
c
c
ur
a
t
e
t
r
a
ns
c
r
i
pt
i
ons
by
c
onv
e
rt
i
ng
ra
w
a
udi
o
i
nt
o
l
og
-
M
e
l
s
p
e
c
t
rogr
a
m
s
,
w
hi
c
h
a
r
e
e
nc
o
de
d
i
nt
o
hi
dd
e
n
s
t
a
t
e
s
a
nd
a
ut
o
re
gr
e
s
s
i
v
e
l
y
d
e
c
ode
d
i
nt
o
t
e
x
t
t
r
a
ns
c
r
i
pt
i
ons
[7]
.
T
h
e
m
ode
l
’s
de
e
p
fus
i
on
l
a
ngu
a
g
e
m
od
e
l
i
n
t
e
gr
a
t
i
on
prov
i
d
e
s
s
up
e
ri
or
p
e
rfor
m
a
n
c
e
c
om
pa
re
d
t
o
s
ha
l
l
ow
fus
i
on
a
ppr
oa
c
he
s
.
W
hi
s
p
e
r
e
x
c
e
pt
i
ona
l
pe
rfor
m
a
n
c
e
i
n
m
u
l
t
i
l
i
ngu
a
l
A
S
R
t
a
s
ks
,
w
i
t
h
W
E
Rs
of
3%
o
n
t
h
e
L
i
bri
S
p
e
e
c
h
t
e
s
t
-
c
l
e
a
n
s
ubs
e
t
a
nd
4.
7
%
on
t
he
T
E
D
-
L
IU
M
c
orp
us
[7]
,
und
e
rs
c
or
e
s
i
t
s
pot
e
n
t
i
a
l
t
o
ge
ne
r
a
l
i
z
e
a
c
ros
s
va
ri
o
us
l
a
n
gua
g
e
s
a
n
d
dom
a
i
ns
.
T
he
s
e
q
ua
l
i
t
i
e
s
m
a
k
e
i
t
a
n
i
de
a
l
c
a
ndi
da
t
e
for
i
m
p
rovi
ng
A
S
R
for
J
a
v
a
n
e
s
e
a
nd
S
u
nda
n
e
s
e
,
w
h
e
re
d
a
t
a
s
c
a
r
c
i
t
y
a
nd
l
i
n
gui
s
t
i
c
v
a
ri
a
bi
l
i
t
y
pre
s
e
nt
s
i
gn
i
fi
c
a
nt
c
h
a
l
l
e
ng
e
s
.
S
und
a
ne
s
e
a
nd
J
a
v
a
n
e
s
e
w
e
r
e
c
h
os
e
n
for
t
hi
s
s
t
u
dy
du
e
t
o
t
he
i
r
w
i
d
e
s
pre
a
d
us
e
i
n
Ind
one
s
i
a
,
de
s
pi
t
e
t
he
l
i
m
i
t
e
d
a
va
i
l
a
bi
l
i
t
y
of
hi
gh
-
qu
a
l
i
t
y
,
a
nno
t
a
t
e
d
s
pe
e
c
h
d
a
t
a
.
S
und
a
ne
s
e
,
s
pok
e
n
by
a
pprox
i
m
a
t
e
l
y
40
m
i
l
l
i
on
p
e
opl
e
,
r
e
pre
s
e
nt
s
one
o
f
In
done
s
i
a
’s
l
a
rge
s
t
l
i
ngu
i
s
t
i
c
groups
but
ha
s
re
c
e
i
ve
d
re
l
a
t
i
v
e
l
y
l
i
t
t
l
e
a
t
t
e
nt
i
on
i
n
A
S
R
r
e
s
e
a
rc
h
[11]
.
T
he
s
u
c
c
e
s
s
of
f
i
ne
-
t
un
i
ng
W
a
v2V
e
c
2
for
S
unda
ne
s
e
A
S
R,
a
s
s
how
n
by
[2
3],
s
ugge
s
t
s
t
h
a
t
f
urt
h
e
r
i
m
pro
ve
m
e
n
t
s
m
a
y
b
e
a
c
hi
e
va
b
l
e
,
pa
r
t
i
c
ul
a
rl
y
w
i
t
h
W
h
i
s
pe
r
’
s
c
o
m
pre
he
ns
i
ve
m
ul
t
i
l
i
ngu
a
l
c
a
pa
bi
l
i
t
i
e
s
.
In
I
ndon
e
s
i
a
’s
d
i
ve
rs
e
l
i
ngu
i
s
t
i
c
l
a
nds
c
a
pe
,
S
und
a
ne
s
e
,
o
ne
of
t
h
e
m
os
t
s
pok
e
n
r
e
gi
o
na
l
l
a
ngua
ge
s
,
i
l
l
us
t
ra
t
e
s
t
he
c
h
a
l
l
e
ng
e
s
of
l
a
n
gua
g
e
c
om
pl
e
xi
t
y
a
nd
v
a
ri
a
t
i
on.
A
s
a
pr
i
m
a
ry
l
a
ngua
ge
f
or
c
o
m
m
uni
t
i
e
s
i
n
W
e
s
t
J
a
va
,
S
unda
ne
s
e
s
e
rve
s
a
s
a
k
e
y
c
ul
t
ura
l
a
nd
c
om
m
un
i
c
a
t
i
v
e
t
oo
l
.
H
ow
e
ve
r
,
t
h
e
a
do
pt
i
on
of
Indon
e
s
i
a
n
a
s
t
he
na
t
i
o
na
l
l
a
ng
ua
g
e
a
nd
t
h
e
c
l
os
e
prox
i
m
i
t
y
of
S
und
a
n
e
s
e
-
s
pe
a
k
i
ng
a
r
e
a
s
t
o
J
a
v
a
ne
s
e
-
s
pe
a
ki
ng
r
e
g
i
ons
ha
ve
i
nt
rod
uc
e
d
a
dd
i
t
i
ona
l
l
a
y
e
rs
of
v
a
r
i
a
t
i
on
w
i
t
hi
n
S
und
a
ne
s
e
.
In
a
re
a
s
l
i
k
e
B
a
nj
a
r
Ci
t
y,
w
h
e
re
S
und
a
ne
s
e
a
nd
J
a
v
a
ne
s
e
c
o
m
m
uni
t
i
e
s
i
n
t
e
rs
e
c
t
,
d
i
s
t
i
n
c
t
s
ub
-
di
a
l
e
c
t
s
h
a
ve
e
m
e
rg
e
d,
i
n
c
l
ud
i
ng
J
a
va
-
i
nfl
u
e
n
c
e
d
a
nd
J
a
v
a
-
do
m
i
na
t
e
d
S
un
da
n
e
s
e
v
a
ri
e
t
i
e
s
[29]
.
A
de
t
a
i
l
e
d
s
t
ud
y
of
B
a
nj
a
r
S
und
a
n
e
s
e
i
d
e
n
t
i
fi
e
d
t
hre
e
pr
i
m
a
ry
s
ub
-
di
a
l
e
c
t
:
s
t
a
nda
r
d
S
und
a
ne
s
e
,
J
a
va
-
i
nfl
ue
n
c
e
d
S
unda
ne
s
e
,
a
nd
J
a
va
-
d
om
i
na
t
e
d
S
unda
ne
s
e
.
E
a
c
h
r
e
fl
e
c
t
di
ffe
r
e
n
t
l
e
v
e
l
s
of
J
a
va
ne
s
e
i
nt
e
gra
t
i
o
n
obs
e
rv
a
bl
e
i
n
l
e
xi
c
a
l
a
nd
phono
l
ogi
c
a
l
s
hi
ft
s
[29]
.
W
h
i
l
e
s
pe
e
c
h
r
e
c
o
g
ni
t
i
on
t
e
c
h
no
l
og
y
h
a
s
s
e
e
n
r
a
p
i
d
d
e
v
e
l
op
m
e
n
t
i
n
r
e
c
e
n
t
y
e
a
rs
,
i
t
s
a
pp
l
i
c
a
t
i
o
ns
re
m
a
i
n
l
a
rg
e
l
y
c
o
nc
e
n
t
r
a
t
e
d
on
m
a
j
o
r
n
a
t
i
o
n
a
l
a
n
d
i
n
t
e
r
n
a
t
i
on
a
l
l
a
ng
u
a
g
e
s
s
u
c
h
a
s
E
ng
l
i
s
h
,
M
a
n
d
a
r
i
n
C
hi
n
e
s
e
,
a
nd
I
nd
o
n
e
s
i
a
n
.
C
o
ns
e
q
u
e
n
t
l
y
,
m
a
ny
r
e
g
i
o
n
a
l
a
nd
i
nd
i
ge
n
o
us
l
a
ng
u
a
g
e
s
l
i
k
e
J
a
v
a
n
e
s
e
a
n
d
S
u
n
d
a
n
e
s
e
h
a
v
e
re
c
e
i
v
e
d
l
i
m
i
t
e
d
a
t
t
e
nt
i
on
i
n
t
e
c
h
n
ol
o
gi
c
a
l
r
e
s
e
a
r
c
h
a
nd
i
n
no
v
a
t
i
o
n
.
T
h
i
s
l
a
c
k
o
f
r
e
p
r
e
s
e
nt
a
t
i
on
n
o
t
o
n
l
y
r
e
f
l
e
c
t
s
a
d
i
g
i
t
a
l
d
i
v
i
d
e
bu
t
a
l
s
o
c
o
n
t
r
i
bu
t
e
s
t
o
t
h
e
o
ng
o
i
n
g
t
hr
e
a
t
of
l
a
ng
u
a
g
e
e
n
d
a
n
g
e
r
m
e
n
t
.
T
o
a
dd
r
e
s
s
t
h
i
s
,
i
n
t
e
g
r
a
t
i
n
g
s
p
e
e
c
h
t
e
c
hn
o
l
o
gy
i
n
t
o
l
a
ng
u
a
g
e
p
re
s
e
r
v
a
t
i
o
n
e
f
fo
r
t
s
h
a
s
b
e
c
o
m
e
i
n
c
r
e
a
s
i
ng
l
y
i
m
p
or
t
a
nt
.
F
o
r
i
n
s
t
a
n
c
e
,
i
m
p
l
e
m
e
n
t
i
n
g
A
S
R
s
ys
t
e
m
s
fo
r
u
n
d
e
r
-
r
e
pr
e
s
e
n
t
e
d
l
a
ng
u
a
g
e
s
c
a
n
e
n
c
ou
r
a
g
e
br
o
a
d
e
r
us
e
a
m
on
g
s
p
e
a
k
e
r
s
a
n
d
fos
t
e
r
l
i
n
g
ui
s
t
i
c
p
r
i
d
e
w
i
t
h
i
n
c
o
m
m
u
n
i
t
i
e
s
.
I
n
t
h
e
c
a
s
e
o
f
S
u
n
d
a
n
e
s
e
,
t
h
e
a
v
a
i
l
a
b
i
l
i
t
y
o
f
t
oo
l
s
s
u
c
h
a
s
s
p
e
e
c
h
-
to
-
t
e
x
t
a
p
p
l
i
c
a
t
i
on
s
c
a
n
r
a
i
s
e
a
w
a
r
e
n
e
s
s
a
b
ou
t
t
h
e
l
a
n
gu
a
g
e
d
e
c
l
i
n
i
ng
u
s
e
,
pr
o
m
ot
e
e
v
e
r
yd
a
y
c
o
m
m
u
n
i
c
a
t
i
o
n
i
n
S
un
d
a
ne
s
e
,
a
n
d
i
ns
p
i
r
e
y
o
un
g
e
r
g
e
n
e
r
a
t
i
on
s
t
o
l
e
a
r
n
a
n
d
e
n
ga
g
e
w
i
t
h
t
h
e
i
r
l
i
ng
u
i
s
t
i
c
h
e
r
i
t
a
g
e
.
F
u
r
t
h
e
r
m
o
r
e
,
s
u
c
h
s
ys
t
e
m
s
c
a
n
b
r
i
d
g
e
c
o
m
m
u
ni
c
a
t
i
on
g
a
p
s
a
n
d
m
a
k
e
d
i
g
i
t
a
l
c
o
n
t
e
n
t
m
o
re
a
c
c
e
s
s
i
b
l
e
i
n
l
o
c
a
l
l
a
n
gu
a
g
e
s
,
s
up
p
or
t
i
n
g
bo
t
h
c
u
l
t
u
r
a
l
p
r
e
s
e
r
v
a
t
i
o
n
a
n
d
i
n
c
l
us
i
v
i
t
y
i
n
di
g
i
t
a
l
e
nv
i
r
o
n
m
e
n
t
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
J
av
a
ne
s
e
and
s
unda
ne
s
e
s
p
e
e
c
h
r
e
c
og
ni
t
i
on
us
i
n
g
W
hi
s
pe
r
(
A
l
i
m
R
ahar
j
o
)
255
Be
yon
d
i
t
s
r
ol
e
i
n
l
a
ngu
a
g
e
pr
e
s
e
rv
a
t
i
on
,
s
p
e
e
c
h
r
e
c
og
ni
t
i
on
t
e
c
h
no
l
og
y
a
l
s
o
of
fe
rs
s
i
gn
i
f
i
c
a
nt
pot
e
n
t
i
a
l
i
n
e
du
c
a
t
i
o
na
l
c
on
t
e
x
t
s
,
pa
rt
i
c
u
l
a
rl
y
fo
r
s
up
por
t
i
ng
t
he
l
e
a
rn
i
ng
of
r
e
g
i
on
a
l
l
a
n
gu
a
g
e
s
s
u
c
h
a
s
S
und
a
n
e
s
e
a
nd
J
a
va
n
e
s
e
[3
0]
,
[
31]
.
By
e
na
bl
i
n
g
t
h
e
d
e
v
e
l
o
pm
e
n
t
of
l
a
ng
ua
ge
l
e
a
rn
i
n
g
s
ys
t
e
m
s
t
h
a
t
c
a
p
t
u
re
a
nd
t
r
a
ns
c
r
i
b
e
n
a
t
i
v
e
s
p
e
e
c
h
,
A
S
R
t
oo
l
s
c
a
n
h
e
l
p
e
ve
ry
da
y
us
e
rs
b
e
c
o
m
e
m
or
e
f
a
m
i
l
i
a
r
w
i
t
h
t
h
e
pr
on
un
c
i
a
t
i
on
,
s
t
ru
c
t
ur
e
,
a
nd
us
e
of
t
h
e
s
e
l
a
ng
ua
ge
s
i
n
r
e
a
l
-
l
i
f
e
s
c
e
n
a
ri
os
.
T
hi
s
i
s
e
s
p
e
c
i
a
l
l
y
v
a
l
ua
bl
e
f
or
l
a
n
gu
a
g
e
s
l
i
k
e
S
und
a
n
e
s
e
a
nd
J
a
va
ne
s
e
,
w
h
i
c
h
fe
a
t
ur
e
c
o
m
pl
e
x
s
p
e
e
c
h
l
e
v
e
l
s
or
r
e
g
i
s
t
e
rs
r
a
ng
i
n
g
fr
o
m
i
nfo
r
m
a
l
f
or
m
s
us
e
d
a
m
on
g
p
e
e
rs
t
o
m
o
re
re
fi
ne
d
a
nd
p
ol
i
t
e
f
or
m
s
us
e
d
w
h
e
n
a
d
dr
e
s
s
i
ng
e
l
de
rs
or
i
n
fo
r
m
a
l
s
e
t
t
i
ngs
.
T
he
s
e
l
i
ng
ui
s
t
i
c
nu
a
n
c
e
s
a
r
e
oft
e
n
di
f
fi
c
ul
t
f
or
l
e
a
rn
e
rs
or
n
on
-
na
t
i
v
e
s
p
e
a
ke
rs
t
o
gr
a
s
p
,
e
s
pe
c
i
a
l
l
y
w
i
t
hou
t
gu
i
de
d
e
x
pos
ure
t
o
n
a
t
i
ve
us
a
g
e
.
A
s
pe
e
c
h
-
to
-
t
e
x
t
o
r
a
ud
i
o
-
b
a
s
e
d
l
e
a
rni
ng
s
ys
t
e
m
p
ow
e
r
e
d
by
A
S
R
c
a
n
provi
de
r
e
a
l
-
t
i
m
e
f
e
e
d
ba
c
k
a
nd
s
t
r
uc
t
ure
d
l
a
ngu
a
ge
m
a
t
e
ri
a
l
s
t
h
a
t
i
nc
o
rpor
a
t
e
t
h
e
s
e
v
a
ri
a
t
i
ons
,
m
a
ki
ng
t
he
l
e
a
rn
i
ng
p
roc
e
s
s
m
o
re
i
nt
e
r
a
c
t
i
v
e
a
nd
c
ul
t
ura
l
l
y
i
nf
orm
e
d.
T
hr
ough
s
u
c
h
a
pp
l
i
c
a
t
i
ons
,
A
S
R
t
e
c
hno
l
ogy
not
onl
y
h
e
l
ps
br
i
dg
e
t
he
a
c
c
e
s
s
i
b
i
l
i
t
y
g
a
p
i
n
l
a
ngu
a
g
e
l
e
a
rn
i
ng
but
a
l
s
o
s
t
r
e
ngt
h
e
ns
c
o
m
m
uni
t
y
e
ng
a
g
e
m
e
nt
w
i
t
h
l
oc
a
l
l
a
ngu
a
ge
s
,
e
ns
uri
ng
t
ha
t
t
he
s
e
c
u
l
t
ura
l
i
de
n
t
i
t
i
e
s
re
m
a
i
n
vi
b
ra
n
t
i
n
t
he
di
g
i
t
a
l
a
g
e
.
F
or
J
a
va
n
e
s
e
,
s
poke
n
by
a
ro
und
90
m
i
l
l
i
on
p
e
opl
e
,
a
ddi
t
i
ona
l
c
ha
l
l
e
ng
e
s
s
t
e
m
fro
m
i
t
s
l
i
ngu
i
s
t
i
c
c
o
m
pl
e
x
i
t
y
a
nd
di
a
l
e
c
t
a
l
va
ri
a
t
i
ons
.
P
re
v
i
ous
re
s
e
a
r
c
h
us
i
ng
t
h
e
X
L
S
-
R
va
r
i
a
n
t
of
W
a
v
2V
e
c
2
a
c
h
i
e
v
e
d
a
W
E
R
of
17
.
95%
[24]
,
but
t
he
r
e
re
m
a
i
ns
s
i
gn
i
fi
c
a
n
t
roo
m
for
i
m
prov
e
m
e
n
t
.
W
hi
s
p
e
r,
w
i
t
h
i
t
s
br
oa
d
e
r
pr
e
-
t
r
a
i
n
i
ng
a
nd
m
u
l
t
i
l
i
ngu
a
l
a
rc
h
i
t
e
c
t
ure
,
i
s
e
xpe
c
t
e
d
t
o
e
nh
a
nc
e
A
S
R
pe
r
form
a
n
c
e
for
J
a
va
ne
s
e
.
S
t
udi
e
s
ha
v
e
a
l
s
o
s
h
ow
n
t
h
a
t
fi
ne
-
t
uni
ng
W
hi
s
p
e
r
fo
r
l
ow
-
re
s
our
c
e
c
h
i
l
d
s
pe
e
c
h
yi
e
l
ds
p
rom
i
s
i
ng
re
s
u
l
t
s
c
om
p
a
re
d
t
o
non
-
fi
n
e
t
un
e
d
m
od
e
l
s
[
32]
–
[3
4]
,
furt
h
e
r
s
up
port
i
ng
i
t
s
po
t
e
n
t
i
a
l
t
o
i
m
prov
e
A
S
R
fo
r
un
de
rr
e
pr
e
s
e
nt
e
d
l
a
ngu
a
ge
s
l
i
ke
J
a
v
a
ne
s
e
a
nd
S
und
a
ne
s
e
.
2.
M
ET
H
O
D
2.
1
.
R
e
s
e
a
r
c
h
s
tag
e
s
T
he
fi
rs
t
s
t
a
ge
of
t
h
i
s
re
s
e
a
r
c
h
i
nvo
l
ve
d
g
a
t
h
e
ri
ng
t
he
re
l
e
va
n
t
d
a
t
a
s
e
t
s
fro
m
p
ubl
i
c
l
y
a
va
i
l
a
bl
e
s
ourc
e
s
,
a
s
i
l
l
us
t
r
a
t
e
d
i
n
F
i
gur
e
1.
W
e
u
t
i
l
i
z
e
d
t
w
o
da
t
a
s
e
t
s
:
O
pe
nS
L
R
(S
L
R35)
fo
r
J
a
va
ne
s
e
a
nd
O
p
e
nS
L
R
(S
L
R36)
for
S
und
a
n
e
s
e
[11]
.
T
h
e
s
e
da
t
a
s
e
t
s
w
e
r
e
s
e
l
e
c
t
e
d
du
e
t
o
t
h
e
i
r
c
om
pr
e
h
e
ns
i
v
e
ne
s
s
,
c
ons
i
s
t
i
ng
of
t
hous
a
n
ds
of
ho
urs
of
t
ra
ns
c
ri
b
e
d
s
pe
e
c
h
.
A
ft
e
r
da
t
a
s
e
t
c
o
l
l
e
c
t
i
on
,
w
e
proc
e
e
de
d
w
i
t
h
t
h
e
pre
-
proc
e
s
s
i
ng
s
t
a
ge
.
T
h
i
s
ph
a
s
e
i
nvo
l
ve
d
org
a
ni
z
i
ng
t
he
da
t
a
i
n
t
o
t
ra
i
ni
ng
a
nd
v
a
l
i
da
t
i
o
n
s
e
t
s
,
e
ns
ur
i
ng
i
t
w
a
s
a
ppr
opri
a
t
e
l
y
pre
pa
re
d
f
or
t
h
e
W
hi
s
pe
r
m
o
de
l
’s
r
e
qu
i
re
m
e
nt
s
.
P
a
rt
o
f
t
h
i
s
pre
pa
ra
t
i
on
i
nc
l
ud
e
d
e
x
t
r
a
c
t
i
ng
a
udi
o
fe
a
t
u
re
s
by
c
onv
e
rt
i
ng
t
h
e
r
a
w
s
ou
nd
f
i
l
e
s
i
nt
o
l
o
g
-
M
e
l
s
p
e
c
t
rogr
a
m
s
,
t
h
e
i
nput
for
m
a
t
W
h
i
s
pe
r
us
e
s
.
T
h
i
s
t
r
a
ns
for
m
a
t
i
o
n
c
a
p
t
ure
s
e
s
s
e
nt
i
a
l
fr
e
qu
e
nc
y
d
e
t
a
i
l
s
i
n
t
he
s
pe
e
c
h
,
m
a
ki
n
g
t
he
da
t
a
w
e
l
l
-
s
u
i
t
e
d
for
t
h
e
f
i
ne
-
t
un
i
ng
pro
c
e
s
s
.
N
ot
a
b
l
y,
t
he
t
e
s
t
i
n
g
d
a
t
a
s
e
t
w
a
s
k
e
pt
s
e
p
a
ra
t
e
t
o
e
ns
ure
a
n
un
bi
a
s
e
d
f
i
na
l
e
v
a
l
u
a
t
i
on.
W
i
t
h
p
re
-
pr
oc
e
s
s
i
ng
c
o
m
pl
e
t
e
,
t
h
e
ne
x
t
s
t
a
g
e
i
nvo
l
ve
d
fi
ne
-
t
uni
n
g
t
he
W
hi
s
p
e
r
m
od
e
l
,
a
s
t
a
t
e
-
of
-
t
he
-
a
rt
s
e
qu
e
n
c
e
-
to
-
s
e
q
ue
n
c
e
A
S
R
m
od
e
l
ori
g
i
n
a
l
l
y
t
ra
i
ne
d
on
o
ve
r
680
,
000
hours
of
m
u
l
t
i
l
i
ngua
l
s
p
e
e
c
h
da
t
a
[
7].
T
hi
s
e
x
t
e
ns
i
ve
pr
e
-
t
ra
i
ni
ng
pr
ovi
d
e
d
a
s
t
ro
ng
fou
nda
t
i
on
,
b
ut
du
e
t
o
t
h
e
uni
q
ue
c
h
a
r
a
c
t
e
r
i
s
t
i
c
s
o
f
J
a
v
a
n
e
s
e
a
nd
S
und
a
n
e
s
e
,
s
p
e
c
i
fi
c
a
d
a
pt
a
t
i
ons
w
e
re
n
e
c
e
s
s
a
ry.
W
e
fi
ne
-
t
une
d
W
h
i
s
pe
r
us
i
n
g
t
he
t
r
a
i
ni
ng
s
ubs
e
t
,
m
odi
fy
i
ng
m
od
e
l
w
e
i
ght
s
t
o
e
nh
a
nc
e
i
t
s
r
e
c
o
gni
t
i
o
n
a
c
c
ur
a
c
y
for
t
h
e
s
e
l
a
n
gua
g
e
s
.
T
hr
oughou
t
t
he
t
ra
i
ni
ng
,
w
e
pe
r
i
odi
c
a
l
l
y
a
s
s
e
s
s
e
d
m
od
e
l
p
e
rfo
rm
a
nc
e
w
i
t
h
t
he
v
a
l
i
da
t
i
o
n
da
t
a
s
e
t
t
o
t
r
a
c
k
i
t
s
pro
gre
s
s
a
nd
m
i
t
i
g
a
t
e
ov
e
rf
i
t
t
i
ng
.
K
e
y
pe
rfor
m
a
n
c
e
m
e
t
ri
c
s
,
i
n
c
l
ud
i
ng
W
E
R
,
w
e
r
e
c
a
l
c
ul
a
t
e
d
a
t
va
ri
ous
i
n
t
e
r
va
l
s
t
o
m
oni
t
or
t
r
a
ns
c
r
i
pt
i
on
a
c
c
u
ra
c
y,
a
nd
t
ra
i
ni
ng
c
on
t
i
nu
e
d
unt
i
l
pre
de
fi
n
e
d
s
t
opp
i
ng
c
r
i
t
e
ri
a
w
e
re
m
e
t
.
F
i
gure
1
.
R
e
s
e
a
rc
h
s
t
a
ge
s
In
t
h
e
f
i
na
l
e
va
l
ua
t
i
o
n
s
t
a
ge
,
i
l
l
us
t
r
a
t
e
d
i
n
F
i
gu
re
1,
t
h
e
m
ode
l
w
a
s
t
e
s
t
e
d
on
t
he
he
l
d
-
ou
t
d
a
t
a
s
e
t
s
for
bot
h
J
a
v
a
ne
s
e
(S
L
R3
5)
a
nd
S
und
a
n
e
s
e
(S
L
R
36)
.
T
hi
s
s
t
a
ge
w
a
s
e
s
s
e
nt
i
a
l
fo
r
a
s
s
e
s
s
i
ng
how
w
e
l
l
t
he
m
od
e
l
c
oul
d
g
e
ne
ra
l
i
z
e
t
o
n
e
w
,
uns
e
e
n
da
t
a
.
P
e
rfor
m
a
nc
e
w
a
s
m
e
a
s
ure
d
us
i
ng
W
E
R,
pr
ovi
d
i
ng
a
qu
a
nt
i
t
a
t
i
ve
a
s
s
e
s
s
m
e
n
t
of
t
r
a
ns
c
ri
p
t
i
on
a
c
c
ur
a
c
y
i
n
bot
h
l
a
n
gua
g
e
s
.
T
h
e
out
c
om
e
s
w
e
r
e
t
h
e
n
c
o
m
p
a
re
d
w
i
t
h
pr
e
vi
ous
s
t
udi
e
s
on
J
a
va
n
e
s
e
a
nd
S
und
a
ne
s
e
A
S
R
[2
3],
[24]
t
o
ga
ug
e
t
he
i
m
prov
e
m
e
nt
s
W
h
i
s
pe
r’s
fi
n
e
-
t
u
ni
ng
brou
ght
t
o
A
S
R
a
c
c
ur
a
c
y
for
l
ow
-
re
s
ou
rc
e
l
a
ngu
a
ge
s
.
T
h
e
s
e
f
i
ndi
ng
s
off
e
r
i
ns
i
ght
s
i
nt
o
t
h
e
pot
e
nt
i
a
l
of
W
hi
s
p
e
r
i
n
a
dva
nc
i
ng
A
S
R
p
e
rfor
m
a
nc
e
for
unde
rre
pr
e
s
e
n
t
e
d
l
a
ngu
a
ge
s
l
i
ke
J
a
va
n
e
s
e
a
nd
S
u
nda
n
e
s
e
,
s
uppo
rt
i
ng
broa
d
e
r
e
ffor
t
s
t
o
m
a
ke
A
S
R
m
or
e
i
nc
l
us
i
v
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
6
,
N
o
.
3
,
N
ov
e
m
be
r
20
25
:
253
-
261
256
2.
2
.
D
ata
c
o
l
l
e
c
t
i
on
T
he
d
a
t
a
s
e
t
s
e
m
p
l
o
ye
d
i
n
t
h
i
s
s
t
ud
y
w
e
r
e
s
o
ur
c
e
d
f
r
o
m
O
pe
n
S
L
R
,
s
p
e
c
i
f
i
c
a
l
l
y
t
h
e
S
L
R
35
c
or
p
us
f
or
J
a
v
a
n
e
s
e
a
nd
S
L
R
36
f
or
S
u
nd
a
n
e
s
e
[
1
1]
.
S
L
R
35
c
ons
i
s
t
s
o
f
18
5
,
0
7
6
u
t
t
e
r
a
n
c
e
s
a
m
o
u
nt
i
ng
t
o
a
p
p
ro
x
i
m
a
t
e
l
y
29
6
ho
u
rs
o
f
a
ud
i
o
,
c
on
t
ri
b
u
t
e
d
b
y
1
,
0
1
9
n
a
t
i
v
e
J
a
v
a
n
e
s
e
s
p
e
a
k
e
rs
.
S
L
R3
6
c
o
nt
a
i
ns
2
19
,
15
6
u
t
t
e
r
a
n
c
e
s
t
o
t
a
l
i
ng
a
pp
ro
x
i
m
a
t
e
l
y
3
33
h
o
urs
of
S
un
d
a
n
e
s
e
s
p
e
e
c
h
fr
o
m
5
42
s
p
e
a
k
e
rs
.
Bo
t
h
d
a
t
a
s
e
t
s
w
e
r
e
r
e
c
or
d
e
d
a
t
1
6
kH
z
,
a
n
d
a
l
l
s
p
e
a
k
e
rs
w
e
r
e
b
e
t
w
e
e
n
t
h
e
a
g
e
s
of
1
8
a
n
d
35
.
H
ow
e
ve
r
,
g
e
n
d
e
r
-
s
p
e
c
i
f
i
c
m
e
t
a
d
a
t
a
w
a
s
n
ot
p
r
ov
i
d
e
d
i
n
t
h
e
d
a
t
a
s
e
t
do
c
u
m
e
n
t
a
t
i
o
n
.
T
h
e
s
p
e
e
c
h
d
a
t
a
w
a
s
c
o
l
l
e
c
t
e
d
i
n
c
ol
l
a
bo
r
a
t
i
o
n
w
i
t
h
t
h
e
D
e
p
a
rt
m
e
n
t
o
f
J
a
v
a
n
e
s
e
L
i
t
e
r
a
t
u
r
e
a
t
U
n
i
v
e
rs
i
t
a
s
G
a
d
j
a
h
M
a
d
a
(U
G
M
)
i
n
Y
o
g
y
a
k
a
rt
a
a
nd
U
ni
v
e
rs
i
t
a
s
P
e
nd
i
di
k
a
n
I
n
do
n
e
s
i
a
(U
P
I)
i
n
B
a
n
du
ng
.
T
h
e
r
e
c
or
di
n
gs
c
a
p
t
ur
e
a
v
a
ri
e
t
y
o
f
s
p
e
a
ki
n
g
s
t
y
l
e
s
,
i
n
c
l
ud
i
ng
r
e
a
d
i
n
g
s
p
e
e
c
h
a
n
d
c
o
nv
e
rs
a
t
i
o
n
a
l
ph
r
a
s
e
s
,
a
nd
a
r
e
a
c
c
o
m
p
a
n
i
e
d
b
y
m
a
n
u
a
l
l
y
t
r
a
n
s
c
r
i
b
e
d
t
e
x
t
.
N
e
v
e
r
t
h
e
l
e
s
s
,
c
h
a
l
l
e
ng
e
s
s
u
c
h
a
s
i
n
c
on
s
i
s
t
e
n
t
s
p
e
l
l
i
ng
,
i
nf
or
m
a
l
pu
n
c
t
u
a
t
i
o
n
,
a
n
d
o
c
c
a
s
i
on
a
l
t
r
a
ns
c
r
i
p
t
i
o
n
e
rr
or
s
w
e
r
e
i
d
e
n
t
i
f
i
e
d
du
r
i
n
g
p
r
e
p
ro
c
e
s
s
i
n
g
.
T
h
e
s
e
w
e
r
e
a
d
dr
e
s
s
e
d
t
hr
ou
g
h
a
n
or
m
a
l
i
z
a
t
i
o
n
p
ro
c
e
s
s
,
i
n
c
l
u
d
i
n
g
l
o
w
e
r
c
a
s
i
n
g
a
n
d
re
m
o
v
a
l
o
f
s
p
e
c
i
a
l
c
h
a
r
a
c
t
e
rs
.
T
o
fa
c
i
l
i
t
a
t
e
c
on
t
rol
l
e
d
e
v
a
l
u
a
t
i
on
a
nd
t
o
re
pl
i
c
a
t
e
pri
o
r
e
x
p
e
ri
m
e
n
t
a
l
s
e
t
ups
[24]
,
on
l
y
a
por
t
i
on
of
t
he
t
o
t
a
l
da
t
a
w
a
s
ut
i
l
i
z
e
d.
A
s
s
how
n
i
n
T
a
bl
e
1,
t
r
a
i
n
i
ng
a
nd
va
l
i
d
a
t
i
on
us
e
d
c
ont
e
nt
fro
m
z
i
p
a
r
c
hi
v
e
s
0
t
hrough
2,
w
h
i
l
e
t
e
s
t
i
ng
w
a
s
c
ond
uc
t
e
d
s
o
l
e
l
y
on
a
r
c
hi
v
e
num
b
e
r
3
.
T
h
e
r
e
s
ul
t
i
ng
da
t
a
s
p
l
i
t
,
a
s
s
how
n
i
n
T
a
b
l
e
2
,
c
o
m
pri
s
e
d
44
hours
a
nd
3
3
m
i
nu
t
e
s
for
J
a
v
a
n
e
s
e
t
r
a
i
ni
ng
a
nd
11
hours
a
n
d
9
m
i
nu
t
e
s
for
v
a
l
i
da
t
i
on
.
F
or
S
und
a
ne
s
e
,
t
he
t
r
a
i
n
i
ng
da
t
a
s
e
t
t
o
t
a
l
e
d
4
9
h
ours
a
nd
52
m
i
nu
t
e
s
,
w
h
i
l
e
t
h
e
v
a
l
i
da
t
i
on
da
t
a
s
e
t
c
ov
e
re
d
1
2
hours
a
nd
28
m
i
n
ut
e
s
.
T
h
e
t
e
s
t
i
ng
da
t
a
s
e
t
i
nv
ol
v
e
d
18
h
our
s
a
nd
28
m
i
nut
e
s
o
f
J
a
va
n
e
s
e
da
t
a
a
n
d
2
0
ho
urs
a
nd
55
m
i
nu
t
e
s
of
S
und
a
n
e
s
e
.
A
l
t
ho
ugh
t
h
e
c
o
rpor
a
a
r
e
e
xt
e
ns
i
ve
,
s
e
v
e
ra
l
l
i
m
i
t
a
t
i
o
ns
w
e
re
no
t
e
d.
T
he
s
e
i
nc
l
ude
t
h
e
l
a
c
k
of
s
pe
a
ke
r
ge
n
de
r
m
e
t
a
da
t
a
,
va
ri
a
t
i
o
ns
i
n
a
u
di
o
q
ua
l
i
t
y
du
e
t
o
d
i
ff
e
ri
n
g
re
c
ord
i
ng
e
nvi
r
onm
e
nt
s
,
a
nd
t
h
e
w
i
de
r
a
ng
e
of
u
t
t
e
ra
n
c
e
dur
a
t
i
on
s
.
S
i
nc
e
W
hi
s
p
e
r
r
e
qui
r
e
s
fi
xe
d
-
l
e
ngt
h
i
npu
t
s
e
que
n
c
e
s
of
30
s
e
c
o
nds
,
a
ud
i
o
c
l
i
ps
w
e
r
e
e
i
t
he
r
t
run
c
a
t
e
d
or
z
e
ro
-
p
a
dd
e
d
a
s
ne
e
de
d.
T
h
e
s
e
pr
e
pro
c
e
s
s
i
ng
a
dj
us
t
m
e
nt
s
e
ns
ur
e
d
uni
f
orm
i
t
y
a
nd
m
od
e
l
c
om
p
a
t
i
bi
l
i
t
y
t
hro
ughout
t
h
e
fi
ne
-
t
uni
ng
pro
c
e
s
s
.
T
a
b
l
e
1
.
D
a
t
a
s
e
t
s
pl
i
t
t
i
ng
S
u
n
d
a
n
e
s
e
J
a
v
a
n
e
s
e
U
s
e
d
i
n
S
u
n
d
a
n
e
s
e
-
a
s
r
-
1
.
.
.
2
J
a
v
a
n
e
s
e
-
a
s
r
-
1
.
.
.
2
T
ra
i
n
i
n
g
&
v
a
l
i
d
a
t
i
o
n
S
u
n
d
a
n
e
s
e
-
a
s
r
-
3
J
a
v
a
n
e
s
e
-
a
s
r
-
3
T
e
s
t
i
n
g
S
u
n
d
a
n
e
s
e
-
a
s
r
-
4
…1
5
J
a
v
a
n
e
s
e
-
a
s
r
-
4
…1
5
N
o
t
u
s
e
d
T
a
b
l
e
2
.
D
a
t
a
s
e
t
dur
a
t
i
on
a
nd
p
a
rt
i
t
i
oni
n
g
L
a
n
g
u
a
g
e
T
ra
i
n
i
n
g
d
u
ra
t
i
o
n
V
a
l
i
d
a
t
i
o
n
d
u
ra
t
i
o
n
T
e
s
t
i
n
g
d
u
ra
t
i
o
n
T
o
t
a
l
u
s
e
d
J
a
v
a
n
e
s
e
44
h
3
3
m
11
h
0
9
m
18
h
2
8
m
74
h
1
0
m
S
u
n
d
a
n
e
s
e
49
h
5
2
m
12
h
2
8
m
20
h
5
5
m
83
h
1
5
m
2.
3
.
P
r
e
-
p
r
oc
e
s
s
s
an
d
f
e
atu
r
e
e
xt
r
a
c
ti
on
T
he
O
pe
nS
L
R
da
t
a
s
e
t
s
w
e
re
di
vi
d
e
d
i
nt
o
t
w
o
m
a
i
n
s
ubs
e
t
s
,
8
0%
a
l
l
o
c
a
t
e
d
t
o
t
r
a
i
n
i
ng
a
n
d
20
%
re
s
e
rv
e
d
f
or
v
a
l
i
da
t
i
o
n.
T
h
i
s
di
vi
s
i
o
n
m
a
x
i
m
i
z
e
d
t
he
a
m
oun
t
of
da
t
a
a
v
a
i
l
a
b
l
e
for
t
ra
i
n
i
ng
w
h
i
l
e
pr
ovi
d
i
ng
a
va
l
i
da
t
i
on
s
e
t
fo
r
t
un
i
ng
a
nd
m
oni
t
o
r
i
ng
m
ode
l
pe
rfor
m
a
n
c
e
dur
i
ng
t
ra
i
ni
ng.
E
a
c
h
a
udi
o
fi
l
e
w
a
s
p
a
i
re
d
w
i
t
h
i
t
s
re
s
p
e
c
t
i
v
e
t
r
a
ns
c
r
i
pt
i
ons
,
f
orm
i
ng
or
ga
n
i
z
e
d
a
n
d
c
ons
i
s
t
e
nt
s
ubs
e
t
s
.
T
o
e
ns
ure
u
ni
for
m
i
t
y
a
c
ros
s
t
h
e
t
ra
ns
c
ri
p
t
i
ons
,
a
t
e
xt
no
rm
a
l
i
z
a
t
i
on
proc
e
s
s
w
a
s
a
pp
l
i
e
d,
w
hi
c
h
i
nvo
l
ve
d
re
m
ovi
ng
a
ny
s
p
e
c
i
a
l
c
ha
r
a
c
t
e
rs
a
nd
re
t
a
i
n
i
ng
onl
y
l
ow
e
r
c
a
s
e
a
l
ph
a
b
e
t
i
c
c
h
a
ra
c
t
e
rs
.
S
i
m
pl
i
fyi
ng
t
he
t
e
xt
i
n
t
h
i
s
w
a
y
m
i
ni
m
i
z
e
d
po
t
e
n
t
i
a
l
i
nc
ons
i
s
t
e
nc
i
e
s
,
h
e
l
p
i
ng
t
o
a
l
i
gn
t
he
t
r
a
ns
c
r
i
pt
i
ons
w
i
t
h
t
he
m
ode
l
'
s
e
xpe
c
t
a
t
i
ons
.
T
o
fi
n
e
-
t
une
t
he
W
hi
s
p
e
r
m
ode
l
w
i
t
h
t
he
O
p
e
nS
L
R
d
a
t
a
,
a
n
i
m
port
a
nt
s
t
e
p
w
a
s
e
ns
uri
ng
c
om
p
a
t
i
bi
l
i
t
y
w
i
t
h
W
h
i
s
pe
r’s
i
n
put
for
m
a
t
re
q
ui
r
e
m
e
nt
s
.
T
h
e
m
od
e
l
do
e
s
n
ot
di
r
e
c
t
l
y
a
c
c
e
p
t
a
ud
i
o
i
n
“
.
f
l
a
c
”
form
a
t
.
T
he
r
e
for
e
,
f
i
l
e
s
ne
e
d
e
d
t
o
be
c
o
nve
r
t
e
d
i
nt
o
l
og
-
M
e
l
s
p
e
c
t
rogr
a
m
s
,
w
hi
c
h
i
s
t
h
e
for
m
a
t
W
hi
s
p
e
r
p
roc
e
s
s
e
s
.
Conv
e
rt
i
ng
t
h
e
a
udi
o
d
a
t
a
r
e
qu
i
re
s
a
n
und
e
rs
t
a
ndi
ng
of
how
di
g
i
t
a
l
s
ys
t
e
m
s
i
nt
e
rpre
t
s
ou
nd:
a
u
di
o
i
s
re
pr
e
s
e
n
t
e
d
a
s
a
o
ne
-
d
i
m
e
ns
i
ona
l
a
rra
y
of
a
m
pl
i
t
ud
e
v
a
l
ue
s
o
ve
r
t
i
m
e
.
S
i
nc
e
c
ont
i
nuous
a
ud
i
o
h
a
s
a
n
i
nfi
n
i
t
e
ra
nge
o
f
va
l
ue
s
,
i
t
m
us
t
be
di
s
c
r
e
t
i
z
e
d
by
s
a
m
pl
i
ng
a
m
p
l
i
t
ude
a
t
fi
x
e
d
i
n
t
e
rva
l
s
,
a
pro
c
e
s
s
know
n
a
s
s
a
m
pl
i
ng
.
S
a
m
pl
i
ng
r
a
t
e
,
t
ypi
c
a
l
l
y
m
e
a
s
u
re
d
i
n
H
e
rt
z
(H
z
),
d
e
fi
n
e
s
t
he
s
e
i
nt
e
rva
l
s
.
M
a
i
nt
a
i
ni
ng
c
ons
i
s
t
e
n
t
s
a
m
pl
i
ng
r
a
t
e
s
i
s
e
s
s
e
n
t
i
a
l
t
o
pr
e
ve
n
t
e
r
rors
dur
i
ng
A
S
R
pro
c
e
s
s
i
ng.
M
i
s
m
a
t
c
h
e
d
s
a
m
p
l
i
n
g
r
a
t
e
s
be
t
w
e
e
n
t
he
a
ud
i
o
i
npu
t
s
a
n
d
W
h
i
s
pe
r
m
o
de
l
r
e
qui
r
e
m
e
n
t
s
c
o
ul
d
re
s
u
l
t
i
n
une
xp
e
c
t
e
d
out
p
ut
s
.
F
o
r
i
ns
t
a
n
c
e
,
pl
a
y
i
ng
a
16
kH
z
a
u
di
o
f
i
l
e
a
t
a
n
8
kH
z
r
a
t
e
di
s
t
ort
s
p
l
a
yba
c
k
s
pe
e
d
a
nd
qua
l
i
t
y
.
F
or
t
hi
s
s
t
ud
y,
W
h
i
s
pe
r’s
f
e
a
t
ure
e
x
t
ra
c
t
or
re
qu
i
re
d
a
16
kH
z
s
a
m
pl
i
ng
ra
t
e
,
s
o
a
udi
o
da
t
a
w
a
s
up
-
s
a
m
pl
e
d
or
dow
n
-
s
a
m
p
l
e
d
a
s
n
e
c
e
s
s
a
ry
t
o
e
ns
u
re
c
o
m
p
a
t
i
bi
l
i
t
y
.
E
ns
uri
ng
c
orr
e
c
t
s
a
m
p
l
i
n
g
r
a
t
e
s
a
l
l
ow
e
d
t
h
e
m
ode
l
t
o
proc
e
s
s
t
he
da
t
a
a
c
c
ur
a
t
e
l
y
a
nd
r
e
du
c
e
d
t
he
ri
s
k
o
f
m
i
s
i
nt
e
rpr
e
t
a
t
i
ons
d
uri
n
g
A
S
R
t
a
s
ks
.
E
ns
uri
ng
co
rre
c
t
s
a
m
p
l
i
ng
r
a
t
e
s
a
l
l
ow
e
d
t
he
m
od
e
l
t
o
proc
e
s
s
t
he
da
t
a
a
c
c
ura
t
e
l
y
a
nd
r
e
du
c
e
d
t
he
r
i
s
k
of
m
i
s
i
n
t
e
r
pre
t
a
t
i
ons
dur
i
ng
A
S
R
t
a
s
ks
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
J
av
a
ne
s
e
and
s
unda
ne
s
e
s
p
e
e
c
h
r
e
c
og
ni
t
i
on
us
i
n
g
W
hi
s
pe
r
(
A
l
i
m
R
ahar
j
o
)
257
In
t
e
r
m
s
of
d
a
t
a
a
ug
m
e
nt
a
t
i
on,
no
a
d
va
n
c
e
d
t
e
c
hn
i
qu
e
s
s
u
c
h
a
s
no
i
s
e
i
nj
e
c
t
i
on
,
pi
t
c
h
s
hi
ft
i
ng
,
or
t
i
m
e
s
t
r
e
t
c
hi
n
g
w
e
r
e
a
pp
l
i
e
d
i
n
t
hi
s
s
t
udy.
W
h
i
l
e
t
h
e
s
e
m
e
t
h
ods
h
a
ve
be
e
n
s
how
n
t
o
e
nh
a
nc
e
g
e
ne
r
a
l
i
z
a
t
i
on
i
n
A
S
R
m
ode
l
s
b
y
s
i
m
u
l
a
t
i
ng
v
a
ri
o
us
a
c
ous
t
i
c
c
ond
i
t
i
ons
a
nd
s
pe
a
k
e
r
va
r
i
a
b
i
l
i
t
y
[5]
,
[35]
,
t
h
e
f
oc
us
of
t
h
i
s
re
s
e
a
rc
h
w
a
s
t
o
e
v
a
l
ua
t
e
t
he
e
ff
e
c
t
i
v
e
n
e
s
s
of
t
h
e
W
hi
s
p
e
r
m
ode
l
w
i
t
h
m
i
n
i
m
a
l
i
n
t
e
rv
e
n
t
i
on
.
N
on
e
t
he
l
e
s
s
,
s
i
m
pl
e
pa
dd
i
ng
a
n
d
t
run
c
a
t
i
on
t
e
c
h
ni
qu
e
s
w
e
re
us
e
d
t
o
s
t
a
nda
rd
i
z
e
a
udi
o
s
a
m
p
l
e
l
e
ngt
hs
t
o
30
s
e
c
onds
,
a
s
re
qui
re
d
by
t
he
W
hi
s
p
e
r
a
rc
h
i
t
e
c
t
ure
.
S
hort
e
r
c
l
i
ps
w
e
re
z
e
r
o
-
pa
dd
e
d
(r
e
pre
s
e
nt
i
ng
s
i
l
e
n
c
e
)
,
a
n
d
l
ong
e
r
c
l
i
ps
w
e
re
t
run
c
a
t
e
d
t
o
m
a
i
nt
a
i
n
uni
f
orm
i
np
ut
di
m
e
ns
i
ons
.
T
h
e
s
e
s
t
e
ps
e
ns
u
re
i
npu
t
c
o
m
pa
t
i
b
i
l
i
t
y
a
nd
s
t
a
bi
l
i
t
y
duri
ng
m
ode
l
t
r
a
i
n
i
ng
,
a
l
t
hou
gh
fut
ur
e
w
ork
m
a
y
e
xpl
o
re
m
ore
s
oph
i
s
t
i
c
a
t
e
d
a
ug
m
e
n
t
a
t
i
o
n
s
t
ra
t
e
g
i
e
s
t
o
furt
h
e
r
i
m
p
rove
rob
us
t
n
e
s
s
.
2.
4
.
F
i
n
e
tu
n
i
n
g
In
t
h
i
s
s
t
a
ge
,
t
h
e
f
i
ne
-
t
un
i
ng
proc
e
s
s
be
ga
n
by
l
oa
d
i
ng
t
he
pre
-
t
r
a
i
ne
d
W
h
i
s
pe
r
-
S
m
a
l
l
m
od
e
l
fro
m
t
he
H
ug
gi
ng
F
a
c
e
H
ub
,
a
popul
a
r
r
e
pos
i
t
ory
f
or
m
a
c
hi
n
e
l
e
a
rn
i
ng
m
od
e
l
s
.
W
h
i
s
pe
r
,
pr
e
-
t
r
a
i
ne
d
on
a
l
a
rg
e
s
c
a
l
e
m
ul
t
i
l
i
ng
ua
l
d
a
t
a
s
e
t
,
i
n
c
l
ude
s
a
u
t
o
m
a
t
i
c
l
a
n
gu
a
ge
d
e
t
e
c
t
i
on
.
H
ow
e
v
e
r,
du
e
t
o
t
he
l
i
ngu
i
s
t
i
c
s
i
m
i
l
a
r
i
t
i
e
s
be
t
w
e
e
n
r
e
gi
o
na
l
l
a
ngu
a
g
e
s
,
pa
r
t
i
c
ul
a
rl
y
i
n
In
d
one
s
i
a
,
i
t
w
a
s
ne
c
e
s
s
a
ry
t
o
e
xp
l
i
c
i
t
l
y
de
f
i
ne
t
h
e
t
a
rg
e
t
l
a
ngu
a
ge
.
T
o
e
ns
u
re
t
h
a
t
t
r
a
ns
c
r
i
pt
i
ons
w
e
r
e
g
e
n
e
ra
t
e
d
i
n
e
i
t
h
e
r
J
a
v
a
ne
s
e
or
S
und
a
n
e
s
e
,
t
he
l
a
ngu
a
ge
a
nd
t
a
s
k
a
rgum
e
n
t
s
w
e
re
s
pe
c
i
f
i
e
d
i
n
t
he
ge
ne
r
a
t
i
on
c
onf
i
gur
a
t
i
on
p
ri
or
t
o
t
ra
i
ni
ng
.
F
or
fi
n
e
-
t
un
i
ng
,
t
he
m
od
e
l
w
a
s
t
r
a
i
n
e
d
us
i
ng
l
a
ngu
a
g
e
-
s
pe
c
i
fi
c
da
t
a
s
e
t
s
w
i
t
h
i
n
put
fe
a
t
u
re
s
(l
og
-
M
e
l
s
pe
c
t
ro
gra
m
s
)
a
nd
c
orre
s
p
ondi
n
g
t
ra
ns
c
ri
pt
i
on
l
a
b
e
l
s
.
T
he
t
ra
i
ni
ng
w
a
s
p
e
rfor
m
e
d
us
i
ng
t
h
e
t
r
a
i
n
e
r
a
ppl
i
c
a
t
i
on
prog
ra
m
m
i
ng
i
nt
e
rfa
c
e
(
API
)
pr
ovi
d
e
d
by
H
u
g
gi
ng
F
a
c
e
,
w
hi
c
h
a
l
l
ow
e
d
c
o
nt
ro
l
ov
e
r
va
r
i
ous
hype
rp
a
ra
m
e
t
e
rs
.
T
he
m
od
e
l
w
a
s
t
r
a
i
ne
d
w
i
t
h
a
ba
t
c
h
s
i
z
e
o
f
16
,
a
l
e
a
rn
i
ng
ra
t
e
of
1e
-
5,
a
nd
a
t
ot
a
l
o
f
5
,
0
00
t
ra
i
ni
ng
s
t
e
ps
,
us
i
ng
t
h
e
A
d
a
m
W
op
t
i
m
i
z
e
r
und
e
r
t
h
e
ho
od.
A
w
a
rm
u
p
of
50
0
s
t
e
ps
w
a
s
a
ppl
i
e
d
t
o
g
ra
du
a
l
l
y
ra
m
p
up
t
he
l
e
a
rn
i
ng
ra
t
e
duri
n
g
t
h
e
e
a
r
l
y
p
ha
s
e
of
t
ra
i
ni
ng.
E
va
l
ua
t
i
o
n
w
a
s
p
e
rfo
rm
e
d
e
ve
r
y
1
,
000
s
t
e
ps
,
a
nd
c
he
c
kpo
i
nt
s
a
vi
n
g
o
c
c
urr
e
d
a
t
t
h
e
s
a
m
e
i
nt
e
rv
a
l
.
T
o
e
nha
n
c
e
m
e
m
ory
e
ff
i
c
i
e
n
c
y
,
gr
a
di
e
nt
c
h
e
c
k
poi
n
t
i
n
g
w
a
s
e
na
b
l
e
d
,
a
nd
t
he
m
o
de
l
w
a
s
t
r
a
i
n
e
d
us
i
ng
m
i
x
e
d
pre
c
i
s
i
on
(
f
p16)
fo
r
f
a
s
t
e
r
c
om
put
a
t
i
on
.
T
he
b
e
s
t
m
od
e
l
w
a
s
s
e
l
e
c
t
e
d
a
u
t
om
a
t
i
c
a
l
l
y
ba
s
e
d
on
t
he
l
ow
e
s
t
W
E
R
obs
e
rve
d
on
t
he
va
l
i
d
a
t
i
on
s
e
t
.
A
l
l
t
ra
i
ni
ng
pr
ogre
s
s
a
nd
e
va
l
ua
t
i
on
m
e
t
ri
c
s
w
e
re
l
og
ge
d
t
h
rough
T
e
ns
orB
oa
rd
.
A
l
t
houg
h
t
h
e
foc
us
of
e
va
l
u
a
t
i
on
w
a
s
on
W
E
R
,
w
hi
c
h
i
s
w
i
d
e
l
y
us
e
d
i
n
A
S
R
t
a
s
ks
,
w
e
a
l
s
o
c
ons
i
d
e
re
d
o
t
he
r
po
t
e
n
t
i
a
l
m
e
t
r
i
c
s
s
u
c
h
a
s
s
e
nt
e
nc
e
e
rro
r
r
a
t
e
(S
E
R)
a
nd
w
ord
re
c
ogn
i
t
i
on
a
c
c
ur
a
c
y
(
W
RA
).
H
ow
e
ve
r
,
du
e
t
o
t
h
e
l
i
m
i
t
e
d
a
v
a
i
l
a
b
i
l
i
t
y
of
s
e
n
t
e
n
c
e
-
s
e
g
m
e
n
t
e
d
l
a
be
l
s
i
n
t
h
e
da
t
a
s
e
t
,
S
E
R
w
a
s
not
c
a
l
c
u
l
a
t
e
d
.
W
RA
,
t
h
e
p
e
rc
e
n
t
a
g
e
of
c
or
re
c
t
l
y
r
e
c
og
ni
z
e
d
w
ords
ou
t
of
a
l
l
a
c
t
ua
l
w
ords
,
w
a
s
e
x
a
m
i
ne
d
dur
i
ng
i
nt
e
rn
a
l
e
va
l
ua
t
i
on
t
o
s
uppor
t
W
E
R
re
s
u
l
t
s
.
Inc
l
ud
i
ng
W
RA
,
i
t
h
e
l
p
e
d
pro
vi
d
e
a
ddi
t
i
on
a
l
i
ns
i
ght
i
nt
o
how
oft
e
n
t
h
e
m
ode
l
a
c
c
ur
a
t
e
l
y
r
e
c
ogni
z
e
d
fu
l
l
w
ords
,
pa
r
t
i
c
ul
a
rl
y
i
n
noi
s
y
or
di
a
l
e
c
t
-
i
n
fl
u
e
nc
e
d
s
a
m
pl
e
s
.
F
ut
u
re
s
t
ud
i
e
s
c
oul
d
b
e
n
e
fi
t
f
rom
m
ore
e
xt
e
ns
i
v
e
us
e
of
t
he
s
e
c
om
pl
e
m
e
nt
a
ry
m
e
t
r
i
c
s
t
o
of
fe
r
a
b
roa
d
e
r
e
v
a
l
ua
t
i
on
of
A
S
R
pe
rf
orm
a
n
c
e
.
2.
5
.
Eva
l
u
ati
on
T
he
e
v
a
l
u
a
t
i
on
proc
e
s
s
w
a
s
don
e
i
n
t
w
o
m
a
i
n
s
t
a
ge
s
:
a
n
i
ni
t
i
a
l
s
t
a
g
e
du
ri
ng
t
r
a
i
n
i
ng
t
o
fi
n
e
-
t
u
ne
t
h
e
m
ode
l
,
a
nd
a
f
i
na
l
e
va
l
ua
t
i
o
n
a
f
t
e
r
t
h
e
fi
n
e
-
t
uni
n
g
w
a
s
c
om
p
l
e
t
e
.
In
t
h
e
fi
rs
t
s
t
a
ge
,
e
va
l
ua
t
i
ons
w
e
re
pe
rfor
m
e
d
w
i
t
hi
n
t
h
e
t
ra
i
ni
ng
c
y
c
l
e
us
i
ng
t
he
va
l
i
d
a
t
i
on
d
a
t
a
s
pl
i
t
.
T
he
pri
m
a
r
y
m
e
t
ri
c
w
a
s
W
E
R,
w
he
r
e
l
ow
e
r
va
l
ue
s
i
n
di
c
a
t
e
d
be
t
t
e
r
m
od
e
l
a
c
c
u
ra
c
y.
T
he
s
e
e
va
l
ua
t
i
ons
w
e
r
e
p
e
rfor
m
e
d
a
t
e
a
c
h
c
he
c
kpo
i
nt
s
pe
c
i
f
i
e
d
i
n
t
he
t
r
a
i
ni
ng
c
onfi
g
ura
t
i
on
,
us
i
ng
t
h
e
va
l
i
d
a
t
i
on
da
t
a
s
e
t
t
o
a
s
s
e
s
s
m
ode
l
p
e
rfor
m
a
nc
e
.
T
hi
s
c
y
c
l
e
c
ont
i
nu
e
d
unt
i
l
t
he
s
t
op
pi
ng
c
ri
t
e
ri
a
w
e
r
e
m
e
t
,
w
i
t
h
t
h
e
m
od
e
l
s
a
vi
n
g
i
t
s
b
e
s
t
ve
rs
i
on
ba
s
e
d
on
t
he
l
ow
e
s
t
re
c
ord
e
d
W
E
R
.
In
t
he
s
e
c
o
nd
s
t
a
ge
,
t
h
e
be
s
t
-
pe
rfor
m
i
ng
m
od
e
l
fr
om
f
i
ne
-
t
uni
ng
w
a
s
r
e
l
o
a
de
d
for
t
e
s
t
i
ng
on
a
s
e
p
a
ra
t
e
s
e
t
of
a
ud
i
o
d
a
t
a
.
In
t
hi
s
e
va
l
ua
t
i
o
n,
t
h
e
m
ode
l
t
r
a
ns
c
ri
b
e
d
e
a
c
h
a
udi
o
s
a
m
pl
e
di
r
e
c
t
l
y,
a
nd
W
E
R
w
a
s
c
a
l
c
u
l
a
t
e
d
by
c
o
m
p
a
ri
n
g
t
he
pr
e
di
c
t
i
ons
t
o
t
h
e
pr
ovi
d
e
d
t
r
a
ns
c
ri
p
t
i
on
f
i
l
e
s
.
T
o
e
ns
ure
a
c
c
ur
a
t
e
W
E
R
c
a
l
c
u
l
a
t
i
on
,
a
nor
m
a
l
i
z
a
t
i
on
proc
e
s
s
w
a
s
a
pp
l
i
e
d
t
o
bot
h
t
h
e
pr
e
di
c
t
e
d
a
nd
t
rue
t
r
a
ns
c
r
i
p
t
i
ons
.
T
hi
s
pro
c
e
s
s
i
n
c
l
ud
e
d
c
on
ve
r
t
i
ng
a
l
l
t
e
x
t
t
o
l
ow
e
r
c
a
s
e
,
r
e
m
ov
i
ng
pun
c
t
u
a
t
i
on
,
t
ri
m
m
i
ng
s
pa
c
e
s
,
e
l
i
m
i
n
a
t
i
ng
m
ul
t
i
p
l
e
s
p
a
c
e
s
,
a
nd
f
i
l
t
e
r
i
ng
o
ut
e
m
p
t
y
s
t
ri
ngs
.
T
h
e
s
e
a
dj
us
t
m
e
nt
s
s
t
a
nd
a
rdi
z
e
d
t
h
e
t
e
xt
for
c
ons
i
s
t
e
nt
W
E
R
e
v
a
l
ua
t
i
on
.
T
h
e
ov
e
r
a
l
l
m
ode
l
pe
rfo
rm
a
nc
e
w
a
s
t
he
n
a
s
s
e
s
s
e
d
by
a
ve
r
a
gi
ng
t
he
W
E
R
a
c
ros
s
t
h
e
d
a
t
a
s
e
t
,
w
i
t
h
a
l
l
i
ndi
v
i
du
a
l
W
E
R
s
c
or
e
s
s
u
m
m
e
d
a
nd
di
vi
d
e
d
by
t
he
t
ot
a
l
nu
m
b
e
r
of
a
udi
o
fi
l
e
s
.
T
hi
s
e
v
a
l
ua
t
i
on
pr
oc
e
s
s
w
a
s
c
a
rr
i
e
d
o
ut
s
e
p
a
ra
t
e
l
y
for
J
a
va
ne
s
e
a
n
d
S
unda
ne
s
e
,
w
i
t
h
e
a
c
h
l
a
ng
ua
g
e
t
e
s
t
e
d
us
i
ng
i
t
s
re
s
p
e
c
t
i
v
e
f
i
ne
-
t
un
e
d
m
o
de
l
.
3.
R
ES
U
LTS
A
N
D
D
I
S
C
U
S
S
I
O
N
In
t
h
i
s
s
t
udy
,
a
l
l
pro
c
e
s
s
e
s
i
n
c
l
udi
ng
d
a
t
a
pr
e
-
pr
oc
e
s
s
i
ng
,
m
ode
l
fi
n
e
t
uni
ng,
a
nd
e
va
l
u
a
t
i
on
w
e
re
i
m
p
l
e
m
e
n
t
e
d
us
i
ng
P
y
t
hon
w
i
t
h
i
n
t
h
e
K
a
ggl
e
e
n
vi
ron
m
e
nt
,
l
e
v
e
ra
g
i
ng
4
c
or
e
s
of
a
n
N
V
ID
IA
T
e
s
l
a
P
1
00
G
P
U
w
i
t
h
29
G
B
of
RA
M
,
a
nd
a
n
Int
e
l
X
e
on
2.
20
G
H
z
CP
U
w
i
t
h
30
G
B
of
RA
M
.
L
i
bra
r
i
e
s
pri
m
a
ri
l
y
a
va
i
l
a
bl
e
i
n
t
h
e
K
a
g
gl
e
not
e
book
e
nvi
ronm
e
n
t
,
s
uc
h
a
s
t
r
a
ns
for
m
e
rs
,
hugg
i
ngf
a
c
e
_hub
,
j
i
w
e
r
,
t
or
c
h
,
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
6
,
N
o
.
3
,
N
ov
e
m
be
r
20
25
:
253
-
261
258
t
orc
h
a
ud
i
o,
a
nd
s
k
l
e
a
rn
,
w
e
r
e
us
e
d
e
x
t
e
ns
i
ve
l
y
t
hrough
out
t
he
re
s
e
a
r
c
h
.
F
or
t
h
e
W
hi
s
p
e
r
m
ode
l
g
e
ne
r
a
t
i
on
,
t
ra
ns
for
m
e
rs
a
nd
hug
gi
ng
fa
c
e
_h
ub
l
i
bra
r
i
e
s
w
e
r
e
ut
i
l
i
z
e
d.
B
e
for
e
fi
n
e
t
un
i
ng
,
t
h
e
T
ok
e
ni
z
e
r
a
nd
F
e
a
t
u
re
E
xt
r
a
c
t
or
f
rom
t
h
e
t
r
a
ns
for
m
e
rs
l
i
bra
r
y
w
e
r
e
i
m
po
rt
e
d;
t
h
e
F
e
a
t
ure
E
xt
ra
c
t
or
proc
e
s
s
e
d
t
h
e
ra
w
a
ud
i
o
i
npu
t
s
,
w
h
i
l
e
t
he
T
oke
ni
z
e
r
c
o
nve
r
t
e
d
m
od
e
l
out
p
ut
s
i
nt
o
r
e
a
da
bl
e
t
e
x
t
.
T
he
us
e
of
t
he
W
hi
s
p
e
r
S
m
a
l
l
m
ode
l
re
pr
e
s
e
n
t
s
a
nove
l
c
on
t
ri
bu
t
i
o
n
t
o
A
S
R
r
e
s
e
a
rc
h
for
J
a
va
n
e
s
e
a
nd
S
und
a
ne
s
e
.
P
r
e
vi
ous
s
t
udi
e
s
,
s
uc
h
a
s
t
h
os
e
by
Cr
ys
s
i
ove
r
a
nd
Z
a
hr
a
[23]
a
nd
A
ri
s
a
put
r
a
e
t
al
.
[24]
,
pri
m
a
ri
l
y
re
l
i
e
d
on
W
a
v2V
e
c
2
B
a
s
e
,
W
a
v2V
e
c
2
l
a
rg
e
,
or
X
L
S
-
R
m
od
e
l
s
,
oft
e
n
re
q
ui
r
i
ng
a
d
di
t
i
on
a
l
N
-
gr
a
m
l
a
ngu
a
g
e
m
od
e
l
s
t
o
boos
t
pe
r
form
a
n
c
e
.
I
n
c
ont
r
a
s
t
,
W
hi
s
p
e
r
i
s
pre
-
t
ra
i
ne
d
on
a
m
a
s
s
i
v
e
,
s
upe
rvi
s
e
d
d
a
t
a
s
e
t
a
nd
i
nt
e
gr
a
t
e
s
bo
t
h
a
c
ous
t
i
c
m
o
de
l
i
ng
a
n
d
l
a
ng
ua
g
e
ge
n
e
ra
t
i
on
i
nt
o
a
s
i
ngl
e
t
ra
ns
fo
rm
e
r
-
b
a
s
e
d
a
r
c
hi
t
e
c
t
u
re
.
T
hi
s
e
nd
-
to
-
e
nd
a
ppr
oa
c
h
e
l
i
m
i
na
t
e
s
t
h
e
n
e
e
d
for
pos
t
-
pro
c
e
s
s
i
ng
w
i
t
h
e
xt
e
rn
a
l
de
c
od
e
rs
,
w
hi
c
h
i
s
e
s
p
e
c
i
a
l
l
y
a
dva
nt
a
ge
ous
for
l
ow
-
re
s
o
urc
e
l
a
ngua
ge
s
w
he
r
e
t
e
xt
c
orpor
a
a
r
e
s
c
a
rc
e
.
M
or
e
ov
e
r,
un
l
i
k
e
m
o
de
l
s
l
i
k
e
B
E
R
T
or
m
BA
RT
t
ha
t
foc
us
on
t
e
xt
ua
l
l
a
ng
ua
g
e
und
e
rs
t
a
ndi
ng,
W
h
i
s
pe
r
i
s
op
t
i
m
i
z
e
d
fo
r
s
p
e
e
c
h
-
to
-
t
e
xt
t
ra
ns
c
ri
p
t
i
on
t
a
s
ks
a
nd
do
e
s
n
ot
r
e
qui
r
e
s
e
p
a
ra
t
e
t
ok
e
ni
z
e
r
s
or
A
S
R
-
s
pe
c
i
f
i
c
a
rc
hi
t
e
c
t
ur
e
s
.
T
hi
s
m
a
k
e
s
i
t
i
nhe
r
e
n
t
l
y
b
e
t
t
e
r
s
ui
t
e
d
for
di
re
c
t
t
r
a
ns
c
ri
p
t
i
on
i
n
l
a
ngu
a
ge
s
w
i
t
h
l
i
m
i
t
e
d
di
g
i
t
a
l
r
e
s
our
c
e
s
.
T
o
t
he
b
e
s
t
of
our
know
l
e
dge
,
t
hi
s
s
t
udy
re
p
re
s
e
nt
s
t
he
fi
rs
t
a
ppl
i
c
a
t
i
on
of
W
hi
s
pe
r
S
m
a
l
l
t
o
J
a
v
a
n
e
s
e
a
nd
S
und
a
ne
s
e
A
S
R,
e
s
t
a
b
l
i
s
h
i
ng
a
ne
w
b
e
nc
h
m
a
rk
f
or
t
he
s
e
l
a
n
gua
g
e
s
i
n
t
e
rm
s
o
f
s
i
m
pl
i
c
i
t
y
,
p
e
rfor
m
a
nc
e
,
a
nd
s
c
a
l
a
b
i
l
i
t
y
.
T
h
e
J
a
v
a
ne
s
e
a
nd
S
un
da
ne
s
e
d
a
t
a
s
e
t
s
,
c
o
l
l
e
c
t
e
d
f
ro
m
t
h
e
O
pe
nS
L
R
w
e
bs
i
t
e
,
w
e
r
e
dow
nl
oa
de
d
a
s
fi
l
e
s
nu
m
b
e
r
e
d
z
e
r
o
t
o
t
hr
e
e
a
l
on
g
w
i
t
h
t
he
i
r
r
e
s
p
e
c
t
i
v
e
t
r
a
n
s
c
r
i
p
t
i
on
fi
l
e
s
[1
1]
.
T
h
e
a
u
di
o
f
i
l
e
s
w
e
r
e
i
n
.
f
l
a
c
for
m
a
t
,
a
n
d
t
h
e
t
r
a
ns
c
r
i
p
t
i
on
fi
l
e
s
i
n
c
l
ud
e
d
c
ol
um
ns
s
u
c
h
a
s
F
i
l
e
ID
,
U
s
e
rID
,
a
n
d
T
r
a
ns
c
r
i
p
t
i
on
.
S
i
n
c
e
t
h
e
U
s
e
r
ID
c
o
l
u
m
n
w
a
s
i
r
re
l
e
v
a
n
t
t
o
t
hi
s
s
t
ud
y,
i
t
w
a
s
r
e
m
o
ve
d,
a
nd
t
h
e
F
i
l
e
ID
c
ol
u
m
n
w
a
s
r
e
n
a
m
e
d
t
o
“
A
ud
i
o
”
t
o
c
l
a
r
i
f
y
i
t
s
f
un
c
t
i
o
n
.
T
o
e
ns
ur
e
e
a
c
h
e
nt
ry
l
i
nk
e
d
c
o
rr
e
c
t
l
y
t
o
i
t
s
c
orr
e
s
p
on
di
ng
a
ud
i
o
fi
l
e
,
c
om
pl
e
t
e
fi
l
e
p
a
t
hs
w
e
r
e
a
dd
e
d
t
o
t
h
e
A
ud
i
o
c
o
l
u
m
n
.
T
he
d
a
t
a
s
e
t
w
a
s
s
p
l
i
t
by
a
l
l
oc
a
t
i
n
g
fo
l
d
e
rs
z
e
r
o
t
o
t
w
o
fo
r
t
r
a
i
n
i
ng
(
80%
)
a
n
d
va
l
i
da
t
i
on
(20
%)
,
a
s
de
s
c
r
i
b
e
d
i
n
T
a
b
l
e
1,
w
hi
l
e
t
h
e
f
i
l
e
s
i
n
f
o
l
d
e
r
t
hr
e
e
c
om
pr
i
s
e
d
t
he
t
e
s
t
i
ng
s
e
t
(10
0%
)
.
U
s
i
ng
t
h
e
m
od
i
fi
e
d
t
ra
ns
c
ri
pt
i
on
f
i
l
e
s
,
fe
a
t
ur
e
e
xt
ra
c
t
i
on
w
a
s
pe
rfor
m
e
d
w
i
t
h
t
h
e
W
hi
s
p
e
rP
roc
e
s
s
or
l
i
br
a
ry
,
w
h
i
c
h
s
uppor
t
s
b
ot
h
a
u
di
o
pr
e
-
pro
c
e
s
s
i
ng
a
nd
t
ok
e
ni
z
a
t
i
on.
A
ud
i
o
f
i
l
e
s
w
e
re
t
r
a
ns
for
m
e
d
i
nt
o
l
o
g
-
M
e
l
s
p
e
c
t
rog
r
a
m
s
,
c
a
p
t
uri
ng
fr
e
qu
e
nc
y
d
a
t
a
ov
e
r
t
i
m
e
t
o
be
t
t
e
r
pr
e
pa
r
e
t
he
d
a
t
a
for
m
o
de
l
fi
ne
-
t
uni
ng.
T
he
t
ra
ns
for
m
e
d
d
a
t
a
s
e
t
s
c
ons
i
s
t
e
d
o
f
t
w
o
m
a
i
n
c
o
m
pon
e
nt
s
:
i
np
ut
_fe
a
t
ure
s
a
nd
L
a
b
e
l
s
.
Inpu
t
_f
e
a
t
ure
s
c
o
nt
a
i
n
e
d
t
he
l
og
-
M
e
l
s
pe
c
t
rog
ra
m
s
de
ri
v
e
d
fr
om
t
h
e
on
e
-
di
m
e
ns
i
o
n
a
l
a
ud
i
o
a
rra
ys
,
w
h
i
c
h
re
p
re
s
e
nt
t
he
fre
que
nc
y
c
ont
e
nt
of
a
udi
o
s
i
gn
a
l
s
.
T
h
i
s
t
r
a
ns
for
m
a
t
i
on
prov
i
de
d
t
h
e
m
ode
l
w
i
t
h
i
nfo
rm
a
t
i
on
t
ha
t
s
pa
ns
b
ot
h
fre
q
ue
n
c
y
a
nd
t
i
m
e
,
c
r
uc
i
a
l
for
a
c
c
ura
t
e
s
pe
e
c
h
re
c
ogni
t
i
on.
L
a
be
l
s
,
ge
n
e
r
a
t
e
d
by
t
he
T
ok
e
n
i
z
e
r
fun
c
t
i
on
,
w
e
r
e
t
oke
n
i
z
e
d
re
pr
e
s
e
n
t
a
t
i
ons
of
t
he
t
r
a
ns
c
r
i
pt
i
on
t
e
x
t
,
e
n
c
od
i
ng
e
a
c
h
t
ra
ns
c
ri
pt
i
on
a
s
a
s
e
qu
e
nc
e
o
f
i
n
t
e
ge
r
ID
s
(L
a
b
e
l
ID
s
)
t
ha
t
t
h
e
m
od
e
l
c
oul
d
i
n
t
e
rpr
e
t
.
W
i
t
h
i
npu
t
_f
e
a
t
ur
e
s
a
nd
L
a
b
e
l
s
i
n
m
e
m
ory
,
t
h
e
m
o
de
l
s
w
e
re
r
e
a
d
y
for
f
i
ne
-
t
un
i
ng
.
T
h
e
s
a
m
e
pr
e
-
t
r
a
i
n
e
d
W
hi
s
p
e
r
ba
s
e
m
od
e
l
w
a
s
us
e
d
for
bo
t
h
l
a
ngua
ge
s
,
but
e
a
c
h
m
od
e
l
w
a
s
t
ra
i
ne
d
s
e
pa
r
a
t
e
l
y
w
i
t
h
l
a
ngu
a
g
e
-
s
pe
c
i
f
i
c
d
a
t
a
:
J
a
v
a
n
e
s
e
d
a
t
a
f
or
t
h
e
“
W
hi
s
p
e
r
-
s
m
a
l
l
-
j
v”
m
ode
l
a
n
d
S
unda
n
e
s
e
da
t
a
for
t
h
e
“
W
hi
s
p
e
r
-
s
m
a
l
l
-
s
u”
m
o
de
l
.
A
ft
e
r
fi
ne
-
t
uni
ng,
t
he
e
va
l
ua
t
i
on
ph
a
s
e
c
om
m
e
nc
e
d.
T
hi
s
p
ha
s
e
i
nv
ol
v
e
d
t
e
s
t
i
n
g
t
he
m
od
e
l
s
w
i
t
h
di
s
t
i
n
c
t
t
e
s
t
i
ng
d
a
t
a
s
e
t
s
s
p
e
c
i
fi
c
t
o
e
a
c
h
l
a
ngu
a
g
e
.
D
ur
i
ng
t
hi
s
ph
a
s
e
,
e
a
c
h
m
od
e
l
t
r
a
ns
c
ri
b
e
d
i
t
s
t
e
s
t
i
ng
da
t
a
s
e
t
,
w
i
t
h
t
ra
ns
c
ri
p
t
i
o
n
a
c
c
ur
a
c
y
m
e
a
s
ure
d
by
W
E
R
.
F
or
e
xa
m
pl
e
,
t
he
m
ode
l
fi
n
e
-
t
u
ne
d
on
J
a
v
a
n
e
s
e
da
t
a
t
ra
ns
c
ri
b
e
d
t
he
J
a
v
a
ne
s
e
t
e
s
t
s
e
t
,
w
h
i
l
e
t
h
e
m
od
e
l
fi
n
e
-
t
u
ne
d
on
S
und
a
n
e
s
e
d
a
t
a
h
a
nd
l
e
d
t
h
e
S
und
a
n
e
s
e
t
e
s
t
s
e
t
.
A
ft
e
r
t
r
a
ns
c
r
i
p
t
i
on
,
a
nor
m
a
l
i
z
a
t
i
on
s
t
e
p
w
a
s
a
ppl
i
e
d
t
o
bo
t
h
pr
e
di
c
t
e
d
a
nd
t
rue
t
r
a
ns
c
r
i
pt
i
ons
.
T
h
i
s
i
nc
l
ude
d
re
m
ov
i
ng
e
m
p
t
y
s
t
r
i
ngs
,
c
on
ve
r
t
i
ng
a
l
l
t
e
xt
t
o
l
ow
e
r
c
a
s
e
,
e
l
i
m
i
na
t
i
ng
e
x
t
ra
s
p
a
c
e
s
,
t
ri
m
m
i
ng
w
hi
t
e
s
p
a
c
e
,
a
nd
r
e
m
o
vi
ng
p
unc
t
ua
t
i
on
,
e
ns
uri
ng
c
ons
i
s
t
e
n
c
y
a
c
ros
s
W
E
R
c
a
l
c
u
l
a
t
i
ons
.
T
h
e
W
E
R
for
e
a
c
h
m
ode
l
w
a
s
c
a
l
c
ul
a
t
e
d
by
a
ve
ra
g
i
ng
i
nd
i
vi
du
a
l
W
E
Rs
a
c
r
os
s
a
l
l
t
r
a
ns
c
ri
b
e
d
f
i
l
e
s
.
T
h
e
J
a
va
n
e
s
e
da
t
a
s
e
t
c
ont
a
i
n
e
d
11
,
574
a
ud
i
o
fi
l
e
s
,
w
h
i
l
e
t
he
S
und
a
ne
s
e
d
a
t
a
s
e
t
h
a
d
1
3,
8
20
a
ud
i
o
f
i
l
e
s
.
T
h
e
a
v
e
ra
g
e
W
E
R
re
s
u
l
t
s
,
a
s
s
how
n
i
n
T
a
bl
e
3
,
re
v
e
a
l
e
d
t
h
a
t
t
he
“
W
hi
s
p
e
r
-
s
m
a
l
l
-
j
v
”
m
ode
l
a
c
hi
e
v
e
d
a
W
E
R
of
14
.
97
%
on
t
h
e
J
a
va
ne
s
e
t
e
s
t
s
e
t
,
w
hi
l
e
t
he
“
W
hi
s
p
e
r
-
s
m
a
l
l
-
s
u”
m
od
e
l
a
c
hi
e
ve
d
a
l
o
w
e
r
W
E
R
o
f
2.
03%
o
n
t
h
e
S
und
a
n
e
s
e
t
e
s
t
s
e
t
,
de
m
o
ns
t
ra
t
i
n
g
W
h
i
s
pe
r’s
e
ff
e
c
t
i
v
e
n
e
s
s
i
n
l
ow
-
re
s
our
c
e
A
S
R
for
t
he
s
e
l
a
ngu
a
g
e
s
.
T
he
pe
rf
orm
a
nc
e
o
f
t
h
e
W
hi
s
p
e
r
m
od
e
l
s
w
a
s
t
h
e
n
c
om
pa
re
d
t
o
pr
e
vi
o
us
re
s
e
a
r
c
h
[2
3],
[24]
,
a
s
s
how
n
i
n
T
a
bl
e
4
.
O
n
t
he
S
u
nda
n
e
s
e
da
t
a
s
e
t
,
t
he
W
h
i
s
pe
r
m
ode
l
d
e
m
ons
t
r
a
t
e
d
s
i
gni
f
i
c
a
nt
i
m
pro
ve
m
e
n
t
s
i
n
W
E
R
ov
e
r
t
h
e
W
a
v2V
e
c
2
B
a
s
e
a
nd
W
a
v2V
e
c
2
L
a
r
ge
m
od
e
l
s
,
w
hi
c
h
a
c
h
i
e
v
e
d
W
E
Rs
of
2
3.
5
%
a
n
d
24
%,
re
s
pe
c
t
i
ve
l
y.
H
ow
e
ve
r
,
de
s
p
i
t
e
t
he
s
e
ga
i
ns
,
t
h
e
W
hi
s
pe
r
m
ode
l
'
s
p
e
rfo
rm
a
nc
e
on
t
h
e
J
a
va
ne
s
e
d
a
t
a
s
e
t
f
e
l
l
s
hort
c
o
m
p
a
re
d
t
o
t
h
e
X
L
S
-
R
m
od
e
l
w
i
t
h
a
n
N
-
gra
m
l
a
ngu
a
ge
m
o
de
l
,
w
hi
c
h
r
e
por
t
e
d
a
l
ow
e
r
W
E
R
of
5
.
4%
.
W
hi
l
e
t
h
e
W
hi
s
p
e
r
m
od
e
l
’s
out
c
om
e
s
on
S
und
a
ne
s
e
d
a
t
a
m
a
rk
a
m
e
a
n
i
ngfu
l
a
dv
a
nc
e
m
e
n
t
i
n
l
ow
-
re
s
our
c
e
A
S
R,
t
h
e
fi
nd
i
ngs
i
n
di
c
a
t
e
t
ha
t
a
dd
i
t
i
ona
l
i
m
prov
e
m
e
n
t
c
o
ul
d
be
a
c
hi
e
ve
d
b
y
i
n
t
e
gr
a
t
i
ng
l
a
ngu
a
g
e
m
od
e
l
s
w
i
t
h
A
S
R
m
od
e
l
s
.
A
s
d
e
t
a
i
l
e
d
i
n
T
a
b
l
e
3,
t
he
W
h
i
s
pe
r
m
ode
l
s
a
c
hi
e
v
e
d
t
r
a
i
n
i
ng
W
E
Rs
of
21
.
4%
for
J
a
v
a
n
e
s
e
a
nd
2.
1%
for
S
un
da
n
e
s
e
,
w
i
t
h
t
e
s
t
i
ng
W
E
Rs
of
1
4.
97%
a
nd
2
.
03
%,
r
e
s
pe
c
t
i
ve
l
y.
T
h
e
s
i
gni
fi
c
a
n
t
di
ffe
r
e
n
c
e
be
t
w
e
e
n
W
E
Rs
for
S
unda
ne
s
e
a
nd
J
a
va
ne
s
e
i
s
b
e
l
i
e
v
e
d
t
o
s
t
e
m
,
i
n
p
a
r
t
,
fro
m
i
n
c
ons
i
s
t
e
nc
i
e
s
i
n
t
he
t
ra
ns
c
ri
p
t
i
o
n
qua
l
i
t
y
of
t
he
J
a
va
n
e
s
e
d
a
t
a
s
e
t
.
A
l
t
ho
ugh
t
h
e
da
t
a
s
e
t
’s
c
r
e
a
t
ors
m
a
nu
a
l
l
y
r
e
vi
e
w
e
d
t
ra
ns
c
ri
p
t
i
ons
,
s
o
m
e
e
rrors
m
a
y
s
t
i
l
l
be
pre
s
e
nt
[11
].
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
J
av
a
ne
s
e
and
s
unda
ne
s
e
s
p
e
e
c
h
r
e
c
og
ni
t
i
on
us
i
n
g
W
hi
s
pe
r
(
A
l
i
m
R
ahar
j
o
)
259
T
a
b
l
e
3
.
E
va
l
ua
t
i
o
n
r
e
s
ul
t
M
o
d
e
l
L
a
n
g
u
a
g
e
V
a
l
i
d
a
t
i
o
n
W
E
R
(%
)
T
e
s
t
i
n
g
W
E
R
(%
)
W
h
i
s
p
e
r
-
s
m
a
l
l
-
jv
J
a
v
a
n
e
s
e
2
1
.
4
1
4
.
9
7
W
h
i
s
p
e
r
-
s
m
a
l
l
-
su
S
u
n
d
a
n
e
s
e
2
.
1
2
.
0
3
T
a
b
l
e
4
.
P
r
e
vi
ous
w
ork
c
om
p
a
ri
s
on
P
re
v
i
o
u
s
w
o
rk
M
o
d
e
l
T
e
s
t
i
n
g
W
E
R
J
a
v
a
n
e
s
e
(%
)
S
u
n
d
a
n
e
s
e
(%
)
[2
4
]
X
L
S
R
-
300
m
+
N
-
g
ra
m
1
0
.
1
5
.
4
[2
3
]
W
a
v
2
V
e
c
2
Ba
s
e
-
2
3
.
5
W
a
v
2
V
e
c
2
L
a
rg
e
-
24
T
h
i
s
s
t
u
d
y
F
i
n
e
t
u
n
e
d
w
h
i
s
p
e
r
s
m
a
l
l
1
4
.
9
7
2
.
0
3
T
he
n
ot
i
c
e
a
b
l
e
g
a
p
i
n
W
E
R
b
e
t
w
e
e
n
t
h
e
t
w
o
l
a
n
gua
g
e
s
2
.
03%
for
S
un
da
n
e
s
e
v
e
rs
us
14
.
97%
for
J
a
v
a
n
e
s
e
c
a
n
b
e
a
t
t
r
i
but
e
d
t
o
s
e
ve
ra
l
f
a
c
t
ors
.
F
i
rs
t
,
t
r
a
ns
c
r
i
pt
i
on
q
ua
l
i
t
y
pl
a
ys
a
c
ri
t
i
c
a
l
r
ol
e
i
n
s
up
e
rvi
s
e
d
l
e
a
rni
n
g,
a
nd
t
h
e
J
a
va
ne
s
e
d
a
t
a
s
e
t
i
s
k
now
n
t
o
c
ont
a
i
n
m
ore
i
nc
o
ns
i
s
t
e
nt
a
nnot
a
t
i
ons
a
nd
o
rt
hog
ra
ph
i
c
va
ri
a
bi
l
i
t
y,
a
s
a
c
know
l
e
dg
e
d
b
y
t
h
e
da
t
a
s
e
t
’s
c
re
a
t
ors
[1
1].
T
he
s
e
i
n
c
ons
i
s
t
e
n
c
i
e
s
m
a
y
ha
ve
i
nt
ro
duc
e
d
no
i
s
e
duri
ng
t
ra
i
ni
ng,
m
a
ki
n
g
i
t
ha
rde
r
for
t
he
m
od
e
l
t
o
l
e
a
rn
re
l
i
a
b
l
e
m
a
ppi
ngs
be
t
w
e
e
n
s
pe
e
c
h
a
n
d
t
e
xt
.
In
c
ont
r
a
s
t
,
t
h
e
S
und
a
n
e
s
e
d
a
t
a
s
e
t
a
pp
e
a
rs
t
o
b
e
m
ore
s
t
a
n
da
rd
i
z
e
d,
w
h
i
c
h
l
i
ke
l
y
i
m
pro
ve
d
m
od
e
l
c
onv
e
rg
e
nc
e
a
nd
r
e
c
o
gni
t
i
on
a
c
c
u
ra
c
y.
S
e
c
ond
,
l
i
ngu
i
s
t
i
c
c
om
pl
e
xi
t
y
a
nd
va
r
i
a
bi
l
i
t
y
m
a
y
c
on
t
ri
b
ut
e
t
o
t
h
e
d
i
ffe
re
n
c
e
.
J
a
v
a
n
e
s
e
e
xh
i
bi
t
s
a
m
or
e
i
nt
r
i
c
a
t
e
s
p
e
e
c
h
l
e
v
e
l
s
ys
t
e
m
,
i
n
c
l
udi
ng
m
u
l
t
i
pl
e
r
e
gi
s
t
e
rs
s
u
c
h
a
s
N
goko
,
M
a
dy
a
,
a
nd
K
ra
m
a
,
e
a
c
h
w
i
t
h
i
t
s
ow
n
vo
c
a
bul
a
ry
a
nd
us
a
ge
ru
l
e
s
.
T
he
s
e
ho
nori
f
i
c
di
s
t
i
nc
t
i
o
ns
i
nt
rodu
c
e
vo
c
a
bul
a
ry
di
ve
rs
i
t
y
a
nd
c
on
t
e
x
t
ua
l
a
m
b
i
gui
t
y
,
w
hi
c
h
c
a
n
i
n
c
re
a
s
e
m
od
e
l
c
onfus
i
on
dur
i
ng
d
e
c
odi
n
g.
S
u
nda
n
e
s
e
,
w
hi
l
e
i
t
a
l
s
o
ha
s
po
l
i
t
e
f
orm
s
,
ge
n
e
r
a
l
l
y
fe
a
t
u
re
s
l
e
s
s
m
orp
hol
og
i
c
a
l
va
r
i
a
t
i
on
a
nd
a
fl
a
t
t
e
r
s
t
ruc
t
ure
i
n
s
pok
e
n
for
m
,
w
hi
c
h
m
a
y
s
i
m
pl
i
fy
t
he
l
e
a
rni
n
g
t
a
s
k
for
t
h
e
m
od
e
l
.
T
hi
rd
,
t
h
e
di
ffe
r
e
n
c
e
m
a
y
a
l
s
o
re
f
l
e
c
t
pronun
c
i
a
t
i
on
va
r
i
a
b
i
l
i
t
y
a
nd
di
a
l
e
c
t
a
l
i
nfl
u
e
nc
e
s
.
J
a
v
a
ne
s
e
i
s
s
poke
n
a
c
ros
s
a
b
roa
d
e
r
ge
ogra
phi
c
a
l
r
e
gi
on
w
i
t
h
s
t
ro
ng
e
r
di
a
l
e
c
t
a
l
v
a
r
i
a
t
i
on
s
uc
h
a
s
S
ol
o,
Y
ogy
a
ka
rt
a
,
Ba
nyu
m
a
s
,
a
n
d
E
a
s
t
e
rn
J
a
v
a
ne
s
e
v
a
ri
e
t
i
e
s
,
w
h
i
l
e
t
h
e
S
un
da
ne
s
e
da
t
a
s
e
t
i
s
m
or
e
re
gi
on
a
l
l
y
c
onc
e
nt
ra
t
e
d
a
round
W
e
s
t
J
a
va
,
p
a
rt
i
c
u
l
a
r
l
y
B
a
ndung
.
T
hi
s
re
g
i
on
a
l
fo
c
us
m
a
y
h
a
ve
r
e
du
c
e
d
va
r
i
a
b
i
l
i
t
y
i
n
s
pe
a
ke
r
a
c
c
e
n
t
s
a
nd
phon
e
t
i
c
re
a
l
i
z
a
t
i
o
ns
,
re
s
u
l
t
i
ng
i
n
a
c
l
e
a
n
e
r
t
r
a
i
n
i
ng
s
i
g
na
l
.
L
a
s
t
l
y
,
t
he
di
s
t
ri
bu
t
i
on
of
u
t
t
e
ra
n
c
e
l
e
ng
t
hs
a
nd
s
p
e
a
k
i
ng
ra
t
e
s
m
i
ght
ha
v
e
pl
a
ye
d
a
rol
e
.
P
re
l
i
m
i
n
a
r
y
i
ns
p
e
c
t
i
on
s
ugg
e
s
t
s
t
ha
t
J
a
v
a
n
e
s
e
r
e
c
ordi
n
gs
i
nc
l
ude
a
w
i
de
r
r
a
ng
e
of
s
p
e
a
k
i
ng
s
p
e
e
ds
a
nd
m
ore
c
o
m
pl
e
x
s
e
n
t
e
n
c
e
c
ons
t
ruc
t
i
o
ns
,
w
h
i
l
e
S
unda
ne
s
e
re
c
or
di
ngs
t
e
nd
t
o
be
m
or
e
c
on
c
i
s
e
a
nd
c
ons
i
s
t
e
n
t
.
T
h
e
s
e
s
ubt
l
e
d
i
ff
e
re
n
c
e
s
i
n
s
pe
a
ki
n
g
s
t
y
l
e
c
oul
d
a
ff
e
c
t
how
w
e
l
l
t
h
e
m
od
e
l
g
e
ne
r
a
l
i
z
e
s
a
c
ros
s
ut
t
e
r
a
nc
e
s
d
uri
n
g
d
e
c
odi
ng
.
4.
C
O
N
C
LU
S
I
O
N
T
hi
s
s
t
ud
y
s
e
t
o
ut
t
o
i
nv
e
s
t
i
ga
t
e
t
he
e
ffe
c
t
i
ve
n
e
s
s
of
t
h
e
W
h
i
s
pe
r
A
S
R
m
ode
l
for
t
w
o
l
ow
r
e
s
ourc
e
l
a
ngu
a
g
e
s
,
J
a
va
ne
s
e
a
n
d
S
und
a
n
e
s
e
,
a
i
m
i
ng
t
o
i
m
prov
e
t
r
a
ns
c
ri
p
t
i
on
a
c
c
u
ra
c
y
f
or
un
de
rr
e
pr
e
s
e
n
t
e
d
l
a
ngu
a
g
e
s
i
n
Ind
one
s
i
a
.
By
fi
ne
t
uni
n
g
W
h
i
s
pe
r
a
m
ode
l
pr
e
t
r
a
i
n
e
d
on
o
v
e
r
680
,
00
0
hours
of
m
ul
t
i
l
i
ng
ua
l
a
udi
o
d
a
t
a
t
hi
s
s
t
udy
e
v
a
l
u
a
t
e
d
t
he
m
ode
l
’s
c
a
p
a
c
i
t
y
t
o
r
e
c
ogni
z
e
a
nd
t
r
a
ns
c
r
i
be
s
pe
e
c
h
i
n
t
he
s
e
l
a
ngu
a
g
e
s
a
c
c
ura
t
e
l
y.
T
he
da
t
a
s
e
t
s
w
h
i
c
h
w
e
r
e
ob
t
a
i
ne
d
fro
m
t
h
e
ope
ns
l
r
(S
L
R
35
f
or
J
a
va
n
e
s
e
a
nd
S
L
R3
6
for
S
unda
n
e
s
e
)
w
e
re
us
e
d
for
t
he
t
ra
i
ni
ng
a
n
d
t
e
s
t
i
n
g
of
t
h
e
W
hi
s
pe
r
m
od
e
l
on
l
a
rge
c
ol
l
e
c
t
i
ons
of
a
udi
o
da
t
a
,
w
he
re
t
h
e
p
e
rfor
m
a
n
c
e
e
v
a
l
ua
t
i
on
w
a
s
b
a
s
e
d
on
t
he
W
E
R
m
e
t
r
i
c
.
T
he
re
s
u
l
t
s
d
e
m
o
ns
t
r
a
t
e
d
s
ubs
t
a
nt
i
a
l
i
m
pro
ve
m
e
n
t
s
i
n
W
E
R
f
or
S
und
a
ne
s
e
,
w
he
r
e
W
hi
s
p
e
r
ou
t
p
e
rfor
m
e
d
pre
v
i
ou
s
m
od
e
l
s
l
i
k
e
W
a
v2V
e
c
2
Ba
s
e
a
nd
W
a
v
2V
e
c
2
L
a
rg
e
,
a
c
hi
e
vi
ng
a
s
i
gn
i
fi
c
a
n
t
l
y
l
ow
e
r
W
E
R
on
t
h
e
t
e
s
t
i
ng
s
e
t
.
H
ow
e
ve
r
,
w
h
i
l
e
t
h
e
W
hi
s
p
e
r
m
ode
l
’s
pe
rfo
rm
a
nc
e
on
S
u
nda
n
e
s
e
da
t
a
w
a
s
not
e
w
or
t
hy
,
t
he
o
ut
c
om
e
s
fo
r
J
a
va
n
e
s
e
hi
g
hl
i
ght
e
d
a
r
e
a
s
for
furt
h
e
r
e
nh
a
n
c
e
m
e
n
t
.
C
o
m
p
a
r
e
d
t
o
t
he
X
L
S
-
R
m
od
e
l
,
w
hi
c
h
us
e
d
a
n
N
-
gr
a
m
l
a
ng
ua
g
e
m
o
de
l
a
nd
a
c
h
i
e
v
e
d
a
l
ow
e
r
W
E
R
on
J
a
v
a
ne
s
e
,
t
h
e
W
h
i
s
pe
r
m
ode
l
s
h
ow
e
d
s
l
i
ght
l
y
hi
ghe
r
W
E
R
s
c
o
re
s
,
s
ugg
e
s
t
i
ng
t
ha
t
a
dd
i
t
i
ona
l
l
a
ngu
a
g
e
-
s
pe
c
i
f
i
c
m
ode
l
i
n
g
s
t
ra
t
e
g
i
e
s
m
i
ght
e
nha
nc
e
W
hi
s
p
e
r’s
pe
rf
orm
a
nc
e
f
or
J
a
v
a
ne
s
e
.
F
ut
ure
i
m
pro
ve
m
e
n
t
for
J
a
va
ne
s
e
l
a
ngua
ge
s
p
e
c
i
fi
c
c
o
ul
d
be
do
ne
w
i
t
h
m
ore
h
ype
r
pa
r
a
m
e
t
e
r
t
uni
n
g
a
nd
m
or
e
t
e
s
t
i
ng.
T
hi
s
di
s
c
re
p
a
nc
y
b
e
t
w
e
e
n
S
und
a
ne
s
e
a
nd
J
a
v
a
ne
s
e
W
E
Rs
a
p
pe
a
rs
t
o
s
t
e
m
p
a
rt
l
y
from
t
ra
ns
c
ri
p
t
i
on
i
nc
ons
i
s
t
e
nc
i
e
s
w
i
t
h
i
n
t
he
J
a
v
a
ne
s
e
d
a
t
a
s
e
t
,
a
s
n
ot
e
d
b
y
t
he
da
t
a
s
e
t
'
s
c
re
a
t
ors
.
T
he
s
e
f
i
nd
i
ngs
un
de
rs
c
ore
t
h
e
i
m
por
t
a
nc
e
of
hi
gh
-
qu
a
l
i
t
y
,
c
ons
i
s
t
e
n
t
t
ra
ns
c
ri
p
t
i
o
ns
i
n
A
S
R
t
r
a
i
n
i
ng
da
t
a
,
e
s
pe
c
i
a
l
l
y
for
l
a
ngu
a
ge
s
w
i
t
h
l
i
m
i
t
e
d
r
e
s
our
c
e
s
.
I
m
prov
e
m
e
n
t
s
t
o
t
r
a
ns
c
r
i
pt
i
on
a
c
c
ur
a
c
y
i
n
t
r
a
i
n
i
ng
d
a
t
a
s
e
t
s
c
o
ul
d
furt
he
r
re
d
uc
e
W
E
R
i
n
fut
ur
e
A
S
R
m
od
e
l
s
for
l
ow
-
r
e
s
ourc
e
l
a
n
gua
g
e
s
.
T
h
e
pe
rf
orm
a
nc
e
d
i
s
pa
r
i
t
y
be
t
w
e
e
n
t
he
t
w
o
l
a
ngu
a
ge
s
unde
rs
c
ore
s
t
h
e
i
m
por
t
a
n
c
e
of
bo
t
h
t
r
a
ns
c
r
i
pt
i
on
qu
a
l
i
t
y
a
n
d
l
i
ngu
i
s
t
i
c
c
o
m
p
l
e
x
i
t
y
i
n
A
S
R
m
od
e
l
t
r
a
i
n
i
ng
.
F
ut
ure
w
o
rk
m
a
y
b
e
n
e
fi
t
fro
m
a
d
dre
s
s
i
n
g
t
h
e
s
e
i
s
s
ue
s
t
h
r
ough
i
m
prov
e
d
a
nno
t
a
t
i
on
pra
c
t
i
c
e
s
a
nd
m
or
e
di
a
l
e
c
t
-
a
w
a
r
e
m
od
e
l
i
ng
s
t
ra
t
e
g
i
e
s
,
p
a
r
t
i
c
ul
a
rl
y
for
J
a
v
a
n
e
s
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
6
,
N
o
.
3
,
N
ov
e
m
be
r
20
25
:
253
-
261
260
F
U
N
D
I
N
G
I
N
F
O
R
M
A
TI
O
N
N
o
fund
i
ng
i
nv
ol
v
e
d.
A
U
TH
O
R
C
O
N
TR
I
BU
TI
O
N
S
S
TA
T
EM
EN
T
T
hi
s
j
our
na
l
us
e
s
t
h
e
Cont
ri
bu
t
or
Rol
e
s
T
a
x
ono
m
y
(
C
Re
di
T
)
t
o
re
c
ogn
i
z
e
i
nd
i
vi
du
a
l
a
ut
hor
c
ont
r
i
bu
t
i
ons
,
r
e
du
c
e
a
ut
hors
h
i
p
di
s
pu
t
e
s
,
a
nd
fa
c
i
l
i
t
a
t
e
c
o
l
l
a
bora
t
i
on
.
N
ame
of
A
u
th
o
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
A
l
i
m
R
a
ha
r
j
o
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
A
m
a
l
i
a
Z
a
hr
a
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
ft
w
a
re
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
rm
a
l
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
s
o
u
rc
e
s
D
:
D
a
t
a
Cu
ra
t
i
o
n
O
:
W
ri
t
i
n
g
-
O
ri
g
i
n
a
l
D
ra
ft
E
:
W
ri
t
i
n
g
-
Re
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
s
u
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
rv
i
s
i
o
n
P
:
P
ro
j
e
c
t
a
d
m
i
n
i
s
t
ra
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
s
i
t
i
o
n
C
O
N
F
LI
C
T
O
F
I
N
T
ER
ES
T
S
TA
T
EM
EN
T
N
o
c
o
nfl
i
c
t
of
i
n
t
e
re
s
t
.
D
A
TA
A
V
A
I
LA
BI
LI
TY
T
he
da
t
a
t
ha
t
s
uppo
rt
t
he
fi
nd
i
ngs
of
t
h
i
s
s
t
udy
a
r
e
ope
n
l
y
a
va
i
l
a
bl
e
i
n
[
O
pe
nS
L
R
]
a
t
ht
t
ps
:
/
/
w
w
w
.
ope
ns
l
r
.
org/
re
s
our
c
e
s
.
php
,
re
f
e
re
n
c
e
nu
m
b
e
r
[1
1].
R
EF
ER
EN
C
ES
[1
]
D
.
Y
u
a
n
d
L
.
D
e
n
g
,
“
A
u
t
o
m
a
t
i
c
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
,
”
S
p
r
i
n
g
e
r
L
o
n
d
o
n
,
2
0
1
5
.
[2
]
Y
.
Z
h
a
n
g
e
t
a
l
.
,
“
G
o
o
g
l
e
U
S
M
:
s
c
a
l
i
n
g
a
u
t
o
m
a
t
i
c
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
b
e
y
o
n
d
1
0
0
l
a
n
g
u
a
g
e
s
,
”
a
r
X
i
v
Co
m
p
u
t
e
r
S
c
i
e
n
c
e
,
2
0
2
3
.
[3
]
J
.
L
i
,
“
Re
c
e
n
t
a
d
v
a
n
c
e
s
i
n
e
n
d
-
to
-
e
n
d
a
u
t
o
m
a
t
i
c
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
,
”
A
P
S
IP
A
T
r
a
n
s
a
c
t
i
o
n
s
o
n
S
i
g
n
a
l
a
n
d
In
f
o
r
m
a
t
i
o
n
P
r
o
c
e
s
s
i
n
g
,
v
o
l
.
1
1
,
n
o
.
1
,
2
0
2
2
,
d
o
i
:
h
t
t
p
s
:
/
/
d
o
i
.
o
rg
/
1
0
.
1
5
6
1
/
1
1
6
.
0
0
0
0
0
0
5
0
.
[4
]
D
.
A
m
o
d
e
i
e
t
a
l
.
,
“
D
e
e
p
s
p
e
e
c
h
2
:
e
nd
-
to
-
e
n
d
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
i
n
E
n
g
l
i
s
h
a
n
d
M
a
n
d
a
ri
n
,
”
3
3
r
d
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
e
n
c
e
o
n
M
a
c
h
i
n
e
L
e
a
r
n
i
n
g
,
ICM
L
2
0
1
6
,
v
o
l
.
1
,
p
p
.
3
1
2
–
3
2
1
,
2
0
1
6
.
[5
]
A
.
G
u
l
a
t
i
e
t
a
l
.
,
“
C
o
n
f
o
rm
e
r
:
c
o
n
v
o
l
u
t
i
o
n
-
a
u
g
m
e
n
t
e
d
t
r
a
n
s
f
o
r
m
e
r
f
o
r
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
,
”
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
A
n
n
u
a
l
C
o
n
f
e
r
e
n
c
e
o
f
t
h
e
I
n
t
e
r
n
a
t
i
o
n
a
l
S
p
e
e
c
h
Co
m
m
u
n
i
c
a
t
i
o
n
A
s
s
o
c
i
a
t
i
o
n
,
INT
E
R
S
P
E
E
CH
,
p
p
.
5
0
3
6
–
5
0
4
0
,
2
0
2
0
,
d
o
i
:
1
0
.
2
1
4
3
7
/
I
n
t
e
r
s
p
e
e
c
h
.
2
0
2
0
-
3
0
1
5
.
[6
]
S
.
K
ri
m
a
n
e
t
a
l
.
,
“
Q
u
a
rt
z
n
e
t
:
d
e
e
p
a
u
t
o
m
a
t
i
c
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
w
i
t
h
1
D
t
i
m
e
-
c
h
a
n
n
e
l
s
e
p
a
ra
b
l
e
c
o
n
v
o
l
u
t
i
o
n
s
,
”
i
n
ICA
S
S
P
,
IE
E
E
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
e
n
c
e
o
n
A
c
o
u
s
t
i
c
s
,
S
p
e
e
c
h
a
n
d
S
i
g
n
a
l
P
r
o
c
e
s
s
i
n
g
-
P
r
o
c
e
e
d
i
n
g
s
,
2
0
2
0
,
p
p
.
6
1
2
4
–
6
1
2
8
,
d
o
i
:
1
0
.
1
1
0
9
/
ICA
S
S
P
4
0
7
7
6
.
2
0
2
0
.
9
0
5
3
8
8
9
.
[7
]
A
.
Ra
d
fo
rd
,
J
.
W
.
K
i
m
,
T
.
X
u
,
G
.
Bro
c
k
m
a
n
,
C.
M
c
L
e
a
v
e
y
,
a
n
d
I.
S
u
t
s
k
e
v
e
r,
“
Ro
b
u
s
t
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
v
i
a
l
a
r
g
e
-
s
c
a
l
e
w
e
a
k
s
u
p
e
rv
i
s
i
o
n
,
”
P
r
o
c
e
e
d
i
n
g
s
o
f
M
a
c
h
i
n
e
L
e
a
r
n
i
n
g
R
e
s
e
a
r
c
h
,
v
o
l
.
2
0
2
,
p
p
.
2
8
4
9
2
–
2
8
5
1
8
,
2
0
2
3
.
[8
]
Z
.
Z
h
a
o
a
n
d
W
.
Q
.
Z
h
a
n
g
,
“
E
n
d
-
to
-
e
n
d
k
e
y
w
o
rd
s
e
a
r
c
h
b
a
s
e
d
o
n
a
t
t
e
n
t
i
o
n
a
n
d
e
n
e
rg
y
s
c
o
re
r
fo
r
l
o
w
re
s
o
u
rc
e
l
a
n
g
u
a
g
e
s
,
”
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
A
n
n
u
a
l
Co
n
f
e
r
e
n
c
e
o
f
t
h
e
In
t
e
r
n
a
t
i
o
n
a
l
S
p
e
e
c
h
Co
m
m
u
n
i
c
a
t
i
o
n
A
s
s
o
c
i
a
t
i
o
n
,
INT
E
R
S
P
E
E
CH
,
p
p
.
2
5
8
7
–
2
5
9
1
,
2
0
2
0
,
d
o
i
:
1
0
.
2
1
4
3
7
/
In
t
e
rs
p
e
e
c
h
.
2
0
2
0
-
2
6
1
3
.
[9
]
C.
M
.
R.
K
.
Ra
j
e
,
“
A
u
t
o
m
a
t
i
c
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
-
A
S
R
s
o
ft
w
a
re
m
a
rk
e
t
re
p
o
rt
2
0
2
4
,
”
c
o
g
n
i
t
i
v
e
m
a
r
k
e
t
r
e
s
e
a
r
c
h
.
c
o
m
,
2
0
2
5
.
[O
n
l
i
n
e
].
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
w
w
w
.
c
o
g
n
i
t
i
v
e
m
a
rk
e
t
re
s
e
a
rc
h
.
c
o
m
/
a
u
t
o
m
a
t
i
c
-
s
p
e
e
c
h
-
re
c
o
g
n
i
t
i
o
n
-
%
2
8
a
s
r%
2
9
-
s
o
ft
w
a
re
-
m
a
r
k
e
t
-
re
p
o
rt
. [
a
c
c
e
s
s
e
d
O
c
t
.
2
5
,
2
0
2
4
]
.
[1
0
]
D
.
M
.
E
b
e
rh
a
rd
,
G
.
F
.
S
i
m
o
n
s
,
a
n
d
C.
D
.
F
e
n
n
i
g
,
“
E
t
h
n
o
l
o
g
u
e
:
l
a
n
g
u
a
g
e
s
o
f
t
h
e
w
o
rl
d
.
t
w
e
n
t
y
-
s
e
v
e
n
t
h
e
d
i
t
i
o
n
,
”
L
a
n
g
u
a
g
e
s
o
f
t
h
e
W
o
r
l
d
,
2
0
2
4
.
[O
n
l
i
n
e
].
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
w
w
w
.
e
t
h
n
o
l
o
g
u
e
.
c
o
m
.
[1
1
]
O
.
K
j
a
rt
a
n
s
s
o
n
,
S
.
S
a
ri
n
,
K
.
P
i
p
a
t
s
ri
s
a
w
a
t
,
M
.
J
a
n
s
c
h
e
,
a
n
d
L
.
H
a
,
“
Cro
w
d
-
s
o
u
rc
e
d
s
p
e
e
c
h
c
o
rp
o
ra
fo
r
J
a
v
a
n
e
s
e
,
S
u
n
d
a
n
e
s
e
,
S
i
n
h
a
l
a
,
N
e
p
a
l
i
,
a
n
d
Ba
n
g
l
a
d
e
s
h
i
Be
n
g
a
l
i
,
”
6
t
h
W
o
r
k
s
h
o
p
o
n
S
p
o
k
e
n
L
a
n
g
u
a
g
e
T
e
c
h
n
o
l
o
g
i
e
s
f
o
r
U
n
d
e
r
-
R
e
s
o
u
r
c
e
d
L
a
n
g
u
a
g
e
s
,
S
L
T
U
2
0
1
8
,
p
p
.
5
2
–
5
5
,
2
0
1
8
,
d
o
i
:
1
0
.
2
1
4
3
7
/
S
L
T
U
.
2
0
1
8
-
11.
[1
2
]
A
.
Ba
e
v
s
k
i
,
H
.
Z
h
o
u
,
A
.
M
o
h
a
m
e
d
,
a
n
d
M
.
A
u
l
i
,
“
W
a
v
2
v
e
c
2
.
0
:
a
fra
m
e
w
o
rk
fo
r
s
e
l
f
-
s
u
p
e
rv
i
s
e
d
l
e
a
rn
i
n
g
o
f
s
p
e
e
c
h
re
p
re
s
e
n
t
a
t
i
o
n
s
,
”
A
d
v
a
n
c
e
s
i
n
Ne
u
r
a
l
In
f
o
r
m
a
t
i
o
n
P
r
o
c
e
s
s
i
n
g
S
y
s
t
e
m
s
,
2
0
2
0
.
[1
3
]
Z
.
M
a
x
w
e
l
l
-
S
m
i
t
h
a
n
d
B.
F
o
l
e
y
,
“
A
u
t
o
m
a
t
e
d
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
o
f
In
d
o
n
e
s
i
a
n
-
E
n
g
l
i
s
h
l
a
n
g
u
a
g
e
l
e
s
s
o
n
s
o
n
Y
o
u
T
u
b
e
u
s
i
n
g
t
ra
n
s
fe
r
l
e
a
rn
i
n
g
,
”
F
i
e
l
d
M
a
t
t
e
r
s
2
0
2
3
-
2
n
d
W
o
r
k
s
h
o
p
o
n
NL
P
A
p
p
l
i
c
a
t
i
o
n
s
t
o
F
i
e
l
d
L
i
n
g
u
i
s
t
i
c
s
,
P
r
o
c
e
e
d
i
n
g
s
,
p
p
.
1
–
1
6
,
2
0
2
3
,
d
o
i
:
1
0
.
1
8
6
5
3
/
v
1
/
2
0
2
3
.
fi
e
l
d
m
a
t
t
e
rs
-
1
.
1
.
[1
4
]
K
.
A
z
i
z
a
h
a
n
d
M
.
A
d
ri
a
n
i
,
“
H
i
e
ra
rc
h
i
c
a
l
t
ra
n
s
fe
r
l
e
a
rn
i
n
g
fo
r
t
e
x
t
-
to
-
s
p
e
e
c
h
i
n
In
d
o
n
e
s
i
a
n
,
J
a
v
a
n
e
s
e
,
a
n
d
S
u
n
d
a
n
e
s
e
l
a
n
g
u
a
g
e
s
,
”
2
0
2
0
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
e
n
c
e
o
n
A
d
v
a
n
c
e
d
Co
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
In
f
o
r
m
a
t
i
o
n
S
y
s
t
e
m
s
,
ICA
CS
IS
2
0
2
0
,
p
p
.
4
2
1
–
4
2
8
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
0
9
/
ICA
CS
I
S
5
1
0
2
5
.
2
0
2
0
.
9
2
6
3
0
8
6
.
[1
5
]
A
.
A
d
i
l
a
,
D
.
L
e
s
t
a
ri
,
A
.
P
u
rw
a
ri
a
n
t
i
,
D
.
T
a
n
a
y
a
,
K
.
A
z
i
z
a
h
,
a
n
d
S
.
S
a
k
t
i
,
“
E
n
h
a
n
c
i
n
g
In
d
o
n
e
s
i
a
n
a
u
t
o
m
a
t
i
c
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
:
e
v
a
l
u
a
t
i
n
g
m
u
l
t
i
l
i
n
g
u
a
l
m
o
d
e
l
s
w
i
t
h
d
i
v
e
rs
e
s
p
e
e
c
h
v
a
ri
a
b
i
l
i
t
i
e
s
,
”
2
0
2
4
2
7
t
h
Co
n
f
e
r
e
n
c
e
o
n
t
h
e
O
r
i
e
n
t
a
l
CO
C
O
S
D
A
In
t
e
r
n
a
t
i
o
n
a
l
Co
m
m
i
t
t
e
e
f
o
r
t
h
e
Co
-
O
r
d
i
n
a
t
i
o
n
a
n
d
S
t
a
n
d
a
r
d
i
s
a
t
i
o
n
o
f
S
p
e
e
c
h
D
a
t
a
b
a
s
e
s
a
n
d
A
s
s
e
s
s
m
e
n
t
T
e
c
h
n
i
q
u
e
s
,
O
-
CO
CO
S
D
A
2
0
2
4
-
P
r
o
c
e
e
d
i
n
g
s
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
O
-
CO
CO
S
D
A
6
4
3
8
2
.
2
0
2
4
.
1
0
8
0
0
3
3
6
.
[1
6
]
O
.
H
.
A
n
i
d
j
a
r,
R
.
M
a
rb
e
l
,
a
n
d
R.
Y
o
z
e
v
i
t
c
h
,
“
W
h
i
s
p
e
r
t
u
rn
s
s
t
ro
n
g
e
r:
a
u
g
m
e
n
t
i
n
g
W
a
v
2
V
e
c
2
.
0
fo
r
s
u
p
e
ri
o
r
A
S
R
i
n
l
o
w
-
re
s
o
u
rc
e
l
a
n
g
u
a
g
e
s
,
”
a
r
X
i
v
Co
m
p
u
t
e
r
S
c
i
e
n
c
e
,
2
0
2
4
.
[1
7
]
M.
-
H
.
H
s
u
a
n
d
H
.
L
e
e
,
“
S
M
IL
E
:
s
p
e
e
c
h
m
e
t
a
i
n
-
c
o
n
t
e
x
t
l
e
a
rn
i
n
g
fo
r
l
o
w
-
re
s
o
u
rc
e
l
a
n
g
u
a
g
e
a
u
t
o
m
a
t
i
c
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
,
”
a
r
X
i
v
E
l
e
c
t
r
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
a
n
d
S
y
s
t
e
m
s
S
c
i
e
n
c
e
,
2
0
2
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
J
av
a
ne
s
e
and
s
unda
ne
s
e
s
p
e
e
c
h
r
e
c
og
ni
t
i
on
us
i
n
g
W
hi
s
pe
r
(
A
l
i
m
R
ahar
j
o
)
261
[1
8
]
Y
.
L
i
u
,
X
.
Y
a
n
g
,
a
n
d
D
.
Q
u
,
“
E
x
p
l
o
ra
t
i
o
n
o
f
w
h
i
s
p
e
r
fi
n
e
-
t
u
n
i
n
g
s
t
ra
t
e
g
i
e
s
fo
r
l
o
w
-
re
s
o
u
rc
e
A
S
R,
”
E
u
r
a
s
i
p
J
o
u
r
n
a
l
o
n
A
u
d
i
o
,
S
p
e
e
c
h
,
a
n
d
M
u
s
i
c
P
r
o
c
e
s
s
i
n
g
,
n
o
.
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
8
6
/
s
1
3
6
3
6
-
024
-
0
0
3
4
9
-
3.
[1
9
]
L
.
Z
h
a
n
g
,
N
.
J
i
a
n
g
,
Q
.
W
a
n
g
,
Y
.
L
i
,
Q
.
L
u
,
a
n
d
L
.
X
i
e
,
“
W
h
i
s
p
e
r
-
S
V
:
a
d
a
p
t
i
n
g
w
h
i
s
p
e
r
fo
r
l
o
w
-
d
a
t
a
-
re
s
o
u
rc
e
s
p
e
a
k
e
r
v
e
ri
fi
c
a
t
i
o
n
,
”
S
p
e
e
c
h
Co
m
m
u
n
i
c
a
t
i
o
n
,
v
o
l
.
1
6
3
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
s
p
e
c
o
m
.
2
0
2
4
.
1
0
3
1
0
3
.
[2
0
]
V
.
T
i
m
m
e
l
,
C.
P
a
o
n
e
s
s
a
,
R.
K
a
k
o
o
e
e
,
M
.
V
o
g
e
l
,
a
n
d
D
.
P
e
rru
c
h
o
u
d
,
“
F
i
n
e
-
t
u
n
i
n
g
w
h
i
s
p
e
r
o
n
l
o
w
-
re
s
o
u
rc
e
l
a
n
g
u
a
g
e
s
fo
r
re
a
l
-
w
o
rl
d
a
p
p
l
i
c
a
t
i
o
n
s
,
”
a
r
X
i
v
Co
m
p
u
t
e
r
S
c
i
e
n
c
e
,
2024.
[2
1
]
D
.
K
.
G
e
t
e
e
t
a
l
.
,
“
W
h
i
s
p
e
ri
n
g
i
n
A
m
h
a
ri
c
:
f
i
n
e
-
t
u
n
i
n
g
w
h
i
s
p
e
r
fo
r
l
o
w
-
re
s
o
u
rc
e
l
a
n
g
u
a
g
e
,
”
a
r
X
i
v
Co
m
p
u
t
e
r
S
c
i
e
n
c
e
,
2025.
[2
2
]
X
.
d
e
Z
u
a
z
o
,
E
.
N
a
v
a
s
,
I
.
S
a
ra
t
x
a
g
a
,
a
n
d
I.
H
.
Ri
o
j
a
,
“
W
h
i
s
p
e
r
-
L
M
:
i
m
p
ro
v
i
n
g
A
S
R
m
o
d
e
l
s
w
i
t
h
l
a
n
g
u
a
g
e
m
o
d
e
l
s
fo
r
l
o
w
-
re
s
o
u
rc
e
l
a
n
g
u
a
g
e
s
,
”
a
r
X
i
v
Co
m
p
u
t
e
r
S
c
i
e
n
c
e
,
2
0
2
5
.
[2
3
]
A
.
Cry
s
s
i
o
v
e
r
a
n
d
A
.
Z
a
h
ra
,
“
S
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
m
o
d
e
l
d
e
s
i
g
n
fo
r
S
u
n
d
a
n
e
s
e
l
a
n
g
u
a
g
e
u
s
i
n
g
W
A
V
2
V
E
C
2
.
0
,
”
In
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
S
p
e
e
c
h
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
2
7
,
n
o
.
1
,
p
p
.
1
7
1
–
1
7
7
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
7
7
2
-
0
2
3
-
1
0
0
6
6
-
5.
[2
4
]
P
.
A
r
i
s
a
p
u
t
ra
,
A
.
T
.
H
a
n
d
o
y
o
,
a
n
d
A
.
Z
a
h
ra
,
“
X
L
S
-
R
d
e
e
p
l
e
a
rn
i
n
g
m
o
d
e
l
fo
r
m
u
l
t
i
l
i
n
g
u
a
l
A
S
R
o
n
l
o
w
-
re
s
o
u
rc
e
l
a
n
g
u
a
g
e
s
:
In
d
o
n
e
s
i
a
n
,
J
a
v
a
n
e
s
e
,
a
n
d
S
u
n
d
a
n
e
s
e
,
”
ICIC
E
x
p
r
e
s
s
L
e
t
t
e
r
s
,
P
a
r
t
B
:
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
5
,
n
o
.
6
,
p
p
.
5
5
1
–
5
5
9
,
2
0
2
4
.
[2
5
]
J
.
C.
V
á
s
q
u
e
z
-
Co
rre
a
a
n
d
A
.
Á
l
v
a
re
z
M
u
n
i
a
i
n
,
“
N
o
v
e
l
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
s
y
s
t
e
m
s
a
p
p
l
i
e
d
t
o
fo
re
n
s
i
c
s
w
i
t
h
i
n
c
h
i
l
d
e
x
p
l
o
i
t
a
t
i
o
n
:
W
a
v
2
v
e
c
2
.
0
v
s
.
w
h
i
s
p
e
r,
”
S
e
n
s
o
r
s
,
v
o
l
.
2
3
,
n
o
.
4
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
3
0
4
1
8
4
3
.
[2
6
]
D
.
R.
Y
e
rra
m
re
d
d
y
,
J
.
M
a
ra
s
a
n
i
,
P
.
S
.
V
.
G
o
w
t
h
a
m
,
G
.
H
a
rs
h
i
t
,
a
n
d
A
n
j
a
l
i
,
“
S
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
p
a
ra
d
i
g
m
s
:
a
c
o
m
p
a
ra
t
i
v
e
e
v
a
l
u
a
t
i
o
n
o
f
s
p
e
e
c
h
b
ra
i
n
,
w
h
i
s
p
e
r
a
n
d
W
a
v
2
V
e
c
2
m
o
d
e
l
s
,
”
2
0
2
4
IE
E
E
9
t
h
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
e
n
c
e
f
o
r
Co
n
v
e
r
g
e
n
c
e
i
n
T
e
c
h
n
o
l
o
g
y
,
I2
CT
2
0
2
4
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
I2
CT
6
1
2
2
3
.
2
0
2
4
.
1
0
5
4
4
1
3
3
.
[2
7
]
A
.
Ba
rc
o
v
s
c
h
i
,
R.
J
a
i
n
,
a
n
d
P
.
Co
rc
o
ra
n
,
“
A
c
o
m
p
a
ra
t
i
v
e
a
n
a
l
y
s
i
s
b
e
t
w
e
e
n
c
o
n
f
o
rm
e
r
-
t
ra
n
s
d
u
c
e
r,
w
h
i
s
p
e
r,
a
n
d
w
a
v
2
v
e
c
2
fo
r
i
m
p
ro
v
i
n
g
t
h
e
c
h
i
l
d
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
,
”
i
n
2
0
2
3
I
n
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
e
n
c
e
o
n
S
p
e
e
c
h
T
e
c
h
n
o
l
o
g
y
a
n
d
H
u
m
a
n
-
Co
m
p
u
t
e
r
D
i
a
l
o
g
u
e
,
S
p
e
D
2
0
2
3
,
2
0
2
3
,
p
p
.
4
2
–
4
7
,
d
o
i
:
1
0
.
1
1
0
9
/
S
p
e
D
5
9
2
4
1
.
2
0
2
3
.
1
0
3
1
4
8
6
7
.
[2
8
]
A
.
Ca
n
d
i
d
o
J
u
n
i
o
r
e
t
a
l
.
,
“
CO
RA
A
A
S
R
:
a
l
a
r
g
e
c
o
r
p
u
s
o
f
s
p
o
n
t
a
n
e
o
u
s
a
n
d
p
re
p
a
r
e
d
s
p
e
e
c
h
m
a
n
u
a
l
l
y
v
a
l
i
d
a
t
e
d
f
o
r
s
p
e
e
c
h
r
e
c
o
g
n
i
t
i
o
n
i
n
B
ra
z
i
l
i
a
n
P
o
rt
u
g
u
e
s
e
,
”
L
a
n
g
u
a
g
e
R
e
s
o
u
r
c
e
s
a
n
d
E
v
a
l
u
a
t
i
o
n
,
v
o
l
.
5
7
,
n
o
.
3
,
p
p
.
1
1
3
9
–
11
7
1
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
5
7
9
-
0
2
2
-
09621
-
4.
[2
9
]
W
a
g
i
a
t
i
,
N
.
D
a
rm
a
y
a
n
t
i
,
Y
.
Y
o
h
a
n
a
ri
s
a
g
a
rn
i
w
a
,
a
n
d
D
.
Z
e
i
n
,
“
M
a
p
p
i
n
g
t
h
e
d
i
m
e
n
s
i
o
n
s
o
f
l
i
n
g
u
i
s
t
i
c
d
i
s
t
a
n
c
e
:
a
s
t
u
d
y
o
n
q
u
a
n
t
i
t
a
t
i
v
e
a
n
d
q
u
a
l
i
t
a
t
i
v
e
g
e
o
l
i
n
g
u
i
s
t
i
c
s
o
f
Ba
n
j
a
r
S
u
n
d
a
n
e
s
e
d
i
a
l
e
c
t
,
”
E
u
r
o
p
e
a
n
J
o
u
r
n
a
l
o
f
L
a
n
g
u
a
g
e
a
n
d
Cu
l
t
u
r
e
S
t
u
d
i
e
s
,
v
o
l
.
2
,
n
o
.
4
,
p
p
.
8
–
1
7
,
2
0
2
3
,
d
o
i
:
1
0
.
2
4
0
1
8
/
e
j
l
a
n
g
.
2
0
2
3
.
2
.
4
.
8
7
.
[3
0
]
W
.
U
d
a
s
m
o
ro
e
t
a
l
.
,
“
T
h
e
p
re
s
e
rv
a
t
i
o
n
o
f
t
h
e
J
a
v
a
n
e
s
e
l
a
n
g
u
a
g
e
i
n
t
h
e
S
p
e
c
i
a
l
Re
g
i
o
n
o
f
Y
o
g
y
a
k
a
rt
a
,
”
In
d
o
n
e
s
i
a
n
J
o
u
r
n
a
l
o
f
G
e
o
g
r
a
p
h
y
,
v
o
l
.
5
5
,
n
o
.
1
,
p
p
.
5
9
–
5
9
,
F
e
b
.
2
0
2
3
,
d
o
i
:
h
t
t
p
s
:
/
/
d
o
i
.
o
rg
/
1
0
.
2
2
1
4
6
/
i
j
g
.
6
8
1
8
3
.
[3
1
]
R.
A
l
h
a
m
m
a
d
,
“
T
h
e
p
h
o
n
o
l
o
g
y
,
m
o
rp
h
o
l
o
g
y
,
a
n
d
s
y
n
t
a
x
o
f
S
u
n
d
a
n
e
s
e
,
”
F
o
r
u
m
f
o
r
L
i
n
g
u
i
s
t
i
c
S
t
u
d
i
e
s
,
v
o
l
.
5
,
n
o
.
3
,
D
e
c
.
2
0
2
3
,
d
o
i
:
h
t
t
p
s
:
/
/
d
o
i
.
o
rg
/
1
0
.
5
9
4
0
0
/
fl
s
.
v
5
i
3
.
1
9
4
5
.
[3
2
]
R.
J
a
i
n
,
A
.
Ba
rc
o
v
s
c
h
i
,
M
.
Y
i
w
e
re
,
P
.
Co
rc
o
ra
n
,
a
n
d
H
.
Cu
c
u
,
“
A
d
a
p
t
a
t
i
o
n
o
f
w
h
i
s
p
e
r
m
o
d
e
l
s
t
o
c
h
i
l
d
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
,
”
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
A
n
n
u
a
l
Co
n
f
e
r
e
n
c
e
o
f
t
h
e
In
t
e
r
n
a
t
i
o
n
a
l
S
p
e
e
c
h
Co
m
m
u
n
i
c
a
t
i
o
n
A
s
s
o
c
i
a
t
i
o
n
,
INT
E
R
S
P
E
E
CH
,
p
p
.
5
2
4
2
–
5
2
4
6
,
2
0
2
3
,
d
o
i
:
1
0
.
2
1
4
3
7
/
In
t
e
rs
p
e
e
c
h
.
2
0
2
3
-
9
3
5
.
[3
3
]
R.
J
a
i
n
,
A
.
Ba
rc
o
v
s
c
h
i
,
M
.
Y
.
Y
i
w
e
re
,
P
.
Co
rc
o
ra
n
,
a
n
d
H
.
Cu
c
u
,
“
E
x
p
l
o
ri
n
g
n
a
t
i
v
e
a
n
d
n
o
n
-
n
a
t
i
v
e
E
n
g
l
i
s
h
c
h
i
l
d
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
w
i
t
h
w
h
i
s
p
e
r,
”
IE
E
E
A
c
c
e
s
s
,
v
o
l
.
1
2
,
p
p
.
4
1
6
0
1
–
4
1
6
1
0
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
CCE
S
S
.
2
0
2
4
.
3
3
7
8
7
3
8
.
[3
4
]
W
.
L
i
u
,
Y
.
Q
i
n
,
Z
.
P
e
n
g
,
a
n
d
T
.
L
e
e
,
“
S
p
a
rs
e
l
y
s
h
a
re
d
l
o
ra
o
n
w
h
i
s
p
e
r
fo
r
c
h
i
l
d
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
,
”
ICA
S
S
P
,
I
E
E
E
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
e
n
c
e
o
n
A
c
o
u
s
t
i
c
s
,
S
p
e
e
c
h
a
n
d
S
i
g
n
a
l
P
r
o
c
e
s
s
i
n
g
-
P
r
o
c
e
e
d
i
n
g
s
,
p
p
.
1
1
7
5
1
–
1
1
7
5
5
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
ICA
S
S
P
4
8
4
8
5
.
2
0
2
4
.
1
0
4
4
7
0
0
4
.
[3
5
]
D
.
S
.
P
a
rk
e
t
a
l
.
,
“
S
p
e
c
a
u
g
m
e
n
t
:
a
s
i
m
p
l
e
d
a
t
a
a
u
g
m
e
n
t
a
t
i
o
n
m
e
t
h
o
d
fo
r
a
u
t
o
m
a
t
i
c
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
,
”
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
A
n
n
u
a
l
Co
n
f
e
r
e
n
c
e
o
f
t
h
e
In
t
e
r
n
a
t
i
o
n
a
l
S
p
e
e
c
h
Co
m
m
u
n
i
c
a
t
i
o
n
A
s
s
o
c
i
a
t
i
o
n
,
INT
E
R
S
P
E
E
CH
,
p
p
.
2
6
1
3
–
2
6
1
7
,
2
0
1
9
,
d
o
i
:
1
0
.
2
1
4
3
7
/
In
t
e
rs
p
e
e
c
h
.
2
0
1
9
-
2
6
8
0
.
BI
O
G
R
A
P
H
I
ES
O
F
A
U
T
H
O
R
S
A
l
i
m
R
ah
a
r
j
o
R
e
c
e
i
ve
d
a
ba
c
he
l
o
r
’
s
de
gr
e
e
i
n
c
o
m
pu
t
e
r
s
c
i
e
nc
e
f
r
om
t
h
e
F
a
c
ul
t
y
of
C
o
m
p
ut
e
r
S
c
i
e
n
c
e
,
B
i
na
N
us
a
nt
a
r
a
U
n
i
ve
r
s
i
t
y,
I
ndone
s
i
a
i
n
2023
.
C
u
r
r
e
nt
l
y
ge
t
t
i
n
g
h
i
s
m
a
s
t
e
r
’
s
d
e
gr
e
e
a
t
C
o
m
p
ut
e
r
S
c
i
e
nc
e
i
n
B
i
na
N
us
a
nt
a
r
a
U
ni
v
e
r
s
i
t
y,
I
nd
one
s
i
a
.
H
i
s
m
a
i
n
i
nt
e
r
e
s
t
s
i
n
c
l
ud
e
s
p
e
e
c
h
r
e
c
ogn
i
t
i
on
,
s
o
f
t
w
a
r
e
d
e
ve
l
op
m
e
nt
,
a
nd
m
a
c
hi
ne
l
e
a
r
ni
n
g.
H
e
c
a
n
b
e
c
ont
a
c
t
e
d
a
t
e
m
a
i
l
:
a
l
i
m
.
r
a
ha
r
j
o
@b
i
nu
s
.
a
c
.
i
d
.
A
m
a
l
i
a
Z
ah
r
a
i
s
a
l
e
c
t
ur
e
r
a
t
t
he
M
a
s
t
e
r
of
I
n
f
or
m
a
t
i
on
T
e
c
hn
ol
ogy
,
B
i
n
a
N
us
a
n
t
a
r
a
U
ni
ve
r
s
i
t
y
,
I
nd
one
s
i
a
.
S
he
r
e
c
e
i
v
e
d
he
r
b
a
c
h
e
l
o
r
’
s
de
g
r
e
e
i
n
c
o
m
pu
t
e
r
s
c
i
e
n
c
e
f
r
o
m
t
he
F
a
c
u
l
t
y
o
f
C
o
m
pu
t
e
r
S
c
i
e
nc
e
,
U
n
i
ve
r
s
i
t
y
of
I
ndo
ne
s
i
a
(
U
I
)
i
n
200
8.
S
h
e
doe
s
no
t
ha
ve
a
m
a
s
t
e
r
’
s
de
gr
e
e
.
H
e
r
P
h
.
D
.
w
a
s
obt
a
i
ne
d
f
r
o
m
t
he
S
c
hoo
l
o
f
C
o
m
p
ut
e
r
S
c
i
e
n
c
e
a
nd
I
nf
o
r
m
a
t
i
c
s
,
U
ni
ve
r
s
i
t
y
C
o
l
l
e
ge
D
u
bl
i
n
(
U
C
D
)
,
I
r
e
l
a
n
d
i
n
201
4.
H
e
r
r
e
s
e
a
r
c
h
i
n
t
e
r
e
s
t
s
c
o
ve
r
va
r
i
o
us
f
i
e
l
d
s
i
n
s
pe
e
c
h
t
e
c
hn
ol
ogy
,
s
uc
h
a
s
s
p
e
e
c
h
r
e
c
ogn
i
t
i
o
n,
s
poke
n
l
a
n
g
ua
ge
i
de
n
t
i
f
i
c
a
t
i
on
,
s
pe
a
ke
r
ve
r
i
f
i
c
a
t
i
on
,
s
p
e
e
c
h
e
m
ot
i
on
r
e
c
ogn
i
t
i
o
n,
a
nd
s
o
o
n.
A
dd
i
t
i
o
na
l
l
y
,
s
he
a
l
s
o
ha
s
a
n
i
nt
e
r
e
s
t
i
n
na
t
u
r
a
l
l
a
ngua
ge
pr
oc
e
s
s
i
ng
(
N
L
P
)
,
c
o
m
pu
t
a
t
i
o
na
l
l
i
ngu
i
s
t
i
c
s
,
m
a
c
hi
n
e
l
e
a
r
n
i
ng
,
a
n
d
a
r
t
i
f
i
c
i
a
l
i
nt
e
l
l
i
ge
n
c
e
.
S
he
c
a
n
be
c
on
t
a
c
t
e
d
a
t
e
m
a
i
l
:
a
m
a
l
i
a
.
z
a
hr
a
@b
i
nu
s
.
e
d
u
.
Evaluation Warning : The document was created with Spire.PDF for Python.