I
AE
S
I
nte
rna
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
9
,
No
.
1
,
Ma
r
ch
2
0
2
0
,
p
p
.
81
~
90
I
SS
N:
2252
-
8938
,
DOI
: 1
0
.
1
1
5
9
1
/i
j
ai.
v
9
.i
1
.
p
p
81
-
90
81
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
A deep
learning
b
a
sed techni
que
for pla
g
ia
ris
m
d
ete
ction: a
co
m
pa
ra
tive s
tud
y
E
l M
o
s
t
a
f
a
H
a
m
bi
,
F
a
o
uzia
B
ena
bb
o
u
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
l
o
g
y
a
n
d
M
o
d
e
li
n
g
L
a
b
o
ra
to
ry
S
c
ien
c
e
,
F
a
c
u
lt
y
o
f
sc
ien
c
e
Be
n
M
’sik
,
Ha
ss
a
n
II
Un
iv
e
rsit
y
o
f
Ca
sa
b
lan
c
a
,
M
o
ro
c
c
o
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Au
g
2
7
,
2
0
1
9
R
ev
i
s
ed
Oct
2
0
,
2
0
1
9
A
cc
ep
ted
No
v
5
,
2
0
1
9
T
h
e
e
a
se
o
f
a
c
c
e
ss
to
th
e
v
a
rio
u
s
re
so
u
rc
e
s
o
n
th
e
w
e
b
e
n
a
b
led
t
h
e
d
e
m
o
c
ra
ti
z
a
ti
o
n
o
f
a
c
c
e
ss
to
in
fo
rm
a
ti
o
n
b
u
t
a
t
th
e
sa
m
e
ti
m
e
a
ll
o
w
e
d
th
e
a
p
p
e
a
ra
n
c
e
o
f
e
n
o
r
m
o
u
s
p
lag
iarism
p
ro
b
lem
s.
M
a
n
y
tec
h
n
iq
u
e
s
o
f
p
lag
iaris
m
w
e
re
id
e
n
ti
f
ied
in
th
e
li
tera
tu
re
,
b
u
t
t
h
e
p
lag
iarism
o
f
id
e
a
ste
e
ls
th
e
f
o
re
m
o
st
tro
u
b
les
o
m
e
to
d
e
tec
t,
b
e
c
a
u
se
it
u
se
s
d
iff
e
r
e
n
t
tex
t
m
a
n
ip
u
lati
o
n
a
t
t
h
e
sa
m
e
ti
m
e
.
In
d
e
e
d
,
a
f
e
w
stra
teg
ie
s
h
a
v
e
b
e
e
n
p
r
o
p
o
se
d
to
p
e
rf
o
rm
th
e
se
m
a
n
ti
c
p
lag
iari
s
m
d
e
tec
ti
o
n
,
b
u
t
th
e
y
a
re
stil
l
n
u
m
e
ro
u
s
c
h
a
ll
e
n
g
e
s
to
o
v
e
rc
o
m
e
.
Un
li
k
e
t
h
e
e
x
isti
n
g
sta
tes
o
f
th
e
a
rt,
th
e
p
u
r
p
o
se
o
f
th
is
stu
d
y
is
to
g
iv
e
a
n
o
v
e
rv
i
e
w
o
f
d
iffere
n
t
p
ro
p
o
siti
o
n
f
o
r
p
lag
iaris
m
d
e
tec
ti
o
n
b
a
se
d
o
n
th
e
d
e
e
p
lea
rn
in
g
a
lg
o
rit
h
m
s.
T
h
e
m
a
in
g
o
a
l
o
f
th
e
se
a
p
p
ro
a
c
h
e
s
is
to
p
r
o
v
id
e
h
ig
h
q
u
a
li
ty
o
f
w
o
rld
s
o
r
se
n
ten
c
e
s
v
e
c
to
r
re
p
re
se
n
tatio
n
.
I
n
th
is
p
a
p
e
r,
w
e
p
ro
p
o
se
a
c
o
m
p
a
ra
ti
v
e
stu
d
y
b
a
s
e
d
o
n
a
se
t
o
f
c
rit
e
rio
n
s
li
k
e
:
V
e
c
to
r
re
p
re
se
n
tatio
n
m
e
th
o
d
,
L
e
v
e
l
T
re
a
t
m
e
n
t,
S
im
il
a
rit
y
M
e
th
o
d
a
n
d
Da
tas
e
t.
On
e
re
su
lt
o
f
th
is
stu
d
y
is
th
a
t
m
o
st
o
f
re
se
a
rc
h
e
s
a
re
b
a
se
d
o
n
w
o
rld
g
ra
n
u
larity
a
n
d
u
se
th
e
w
o
rd
2
v
e
c
m
e
th
o
d
f
o
r
wo
rd
v
e
c
to
r
re
p
re
se
n
tatio
n
,
w
h
ich
so
m
e
ti
m
e
s
is
n
o
t
su
it
a
b
le
to
k
e
e
p
th
e
m
e
a
n
in
g
o
f
th
e
w
h
o
le
se
n
ten
c
e
s.
Eac
h
te
c
h
n
iq
u
e
h
a
s
stre
n
g
th
s
a
n
d
w
e
a
k
n
e
ss
e
s;
h
o
w
e
v
e
r,
n
o
n
e
is q
u
it
e
m
a
tu
re
f
o
r
se
m
a
n
ti
c
p
lag
iaris
m
d
e
tec
ti
o
n
.
K
ey
w
o
r
d
s
:
Dee
p
L
ea
r
n
i
n
g
Do
c2
v
ev
Neu
r
al
n
et
w
o
r
k
P
lag
iar
is
m
P
r
e
-
p
r
o
ce
s
s
in
g
W
o
r
d
2
v
ec
T
h
is i
s
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
:
E
l M
o
s
taf
a
Ha
m
b
i,
I
n
f
o
r
m
a
tio
n
T
ec
h
n
o
lo
g
y
an
d
Mo
d
elin
g
L
ab
o
r
ato
r
y
Scie
n
ce
,
Facu
lt
y
o
f
s
cie
n
ce
B
en
M
’
s
i
k
,
Hass
a
n
I
I
Un
i
v
er
s
it
y
o
f
C
a
s
ab
lan
ca
,
C
asab
lan
ca
2
0
0
0
0
,
Mo
r
o
c
co
.
E
m
ail:
h
a
m
b
i
m
o
s
ta
f
a9
1
@
g
m
a
il.c
o
m
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
ad
v
an
ce
m
e
n
t
o
f
i
n
f
o
r
m
atio
n
tech
n
o
lo
g
y
(
I
T
)
an
d
p
ar
ticu
lar
l
y
t
h
e
W
eb
h
as
im
p
r
e
s
s
i
v
el
y
ex
p
an
d
ed
th
e
ac
ce
s
s
ib
ilit
y
o
f
d
ata
an
d
lead
s
t
h
u
s
to
th
e
r
is
i
n
g
o
f
p
lag
iar
i
s
m
.
P
lag
iar
is
m
i
s
a
p
r
ac
tice
o
f
ta
k
i
n
g
s
o
m
eo
n
e
el
s
e
'
s
w
o
r
k
o
r
id
ea
s
an
d
p
ass
i
n
g
t
h
e
m
o
f
f
as o
n
e
's
o
w
n
.
Sev
er
al
p
la
g
iar
is
m
tec
h
n
iq
u
es a
r
e
p
er
f
o
r
m
ed
b
y
s
o
m
e
d
is
h
o
n
e
s
t a
u
t
h
o
r
s
,
an
d
h
er
e
b
ello
w
s
o
m
e
o
f
th
e
m
[
1
-
2
]
:
C
o
p
y
-
p
as
te,
tex
t
u
all
y
(
w
o
r
d
b
y
w
o
r
d
)
:
th
e
co
n
ten
t
o
f
t
h
e
te
x
t
is
co
p
ied
f
r
o
m
o
n
e
o
r
m
o
r
e
s
o
u
r
ce
s
an
d
co
u
ld
b
e
s
lig
h
tl
y
m
o
d
if
ied
.
P
ar
ap
h
r
asin
g
:
t
h
e
g
r
a
m
m
ar
o
f
th
e
te
x
t
is
ch
a
n
g
ed
th
e,
w
o
r
d
s
ar
e
ch
an
g
ed
b
y
t
h
eir
s
y
n
o
n
y
m
s
.
T
h
e
s
en
te
n
ce
s
ar
e
r
e
o
r
g
an
ized
f
r
o
m
th
e
o
r
i
g
in
a
l
w
o
r
k
an
d
s
o
m
e
p
ar
ts
o
f
th
e
te
x
t a
r
e
d
elete
d
.
Fals
e
r
ef
er
e
n
ce
s
,
r
ef
er
e
n
ce
s
ar
e
ch
an
g
ed
an
d
s
o
m
e
ti
m
es a
r
e
f
alse o
r
th
at
d
o
n
o
t e
v
e
n
e
x
is
t.
P
lag
iar
is
m
w
i
th
tr
a
n
s
la
tio
n
,
t
h
e
co
n
ten
t
s
ar
e
tr
an
s
lated
an
d
u
s
ed
w
it
h
o
u
t r
ef
er
en
ce
to
t
h
e
o
r
ig
i
n
al
w
o
r
k
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
1
,
Ma
r
ch
20
20
:
81
–
90
82
P
lag
iar
is
m
o
f
id
ea
s
,
it
is
th
e
m
o
s
t
d
i
f
f
icu
lt
p
lag
iar
i
s
m
to
d
etec
t
b
ec
au
s
e
i
t
is
m
o
r
e
co
m
p
licated
th
a
n
t
h
e
p
r
ev
io
u
s
t
y
p
es,
b
ec
a
u
s
e
i
t
is
n
o
t
s
i
m
p
le
m
a
n
ip
u
lat
io
n
s
m
a
d
e
o
n
th
e
tex
t,
b
u
t
a
m
o
r
e
a
d
v
an
ce
d
f
o
r
m
w
h
ic
h
co
u
ld
in
cl
u
d
e
all
t
h
e
o
th
er
tech
n
iq
u
e
s
.
I
n
g
e
n
er
al,
w
e
ca
n
cla
s
s
i
f
y
t
h
e
p
lag
iar
is
m
tec
h
n
iq
u
es
o
n
th
r
ee
s
tr
ate
g
ie
s
:
le
x
ical,
s
y
n
t
ax
ial
a
n
d
s
e
m
a
n
tic
m
et
h
o
d
s
.
T
h
e
p
lag
ia
r
is
m
o
f
id
ea
s
m
o
s
t
o
f
te
n
i
n
co
r
p
o
r
ates
r
ef
o
r
m
u
latio
n
s
as
w
e
ll
as
s
e
m
a
n
tic
a
n
d
lex
ical
c
h
an
g
es
w
h
ic
h
m
ak
e
it
v
er
y
h
ar
d
to
d
etec
t [
3
]
:
T
h
e
L
ex
ical
m
et
h
o
d
s
co
n
s
id
e
r
tex
t
a
s
a
s
eq
u
e
n
ce
o
f
c
h
ar
a
cter
s
o
r
ter
m
s
[
4
]
.
T
h
e
p
r
e
-
p
r
o
ce
s
s
in
g
tech
n
iq
u
e
i
n
cl
u
d
es
to
k
e
n
izati
o
n
,
lo
w
er
ca
s
i
n
g
,
p
u
n
c
tu
at
io
n
r
e
m
o
v
al
an
d
s
te
m
m
i
n
g
[
5
]
.
T
h
e
m
o
r
e
co
m
m
o
n
ter
m
s
th
e
d
o
cu
m
e
n
ts
h
a
v
e,
t
h
e
m
o
r
e
s
i
m
ilar
t
h
e
y
ar
e.
Me
th
o
d
s
s
u
c
h
as
lo
n
g
est
co
m
m
o
n
s
u
b
s
eq
u
en
ce
,
n
-
g
r
a
m
s
an
d
f
in
g
er
p
r
in
t
ar
e
co
n
s
id
er
ed
as
th
is
k
i
n
d
o
f
m
et
h
o
d
s
.
T
h
e
co
m
p
ar
is
o
n
u
n
its
ad
o
p
ted
in
clu
d
e
w
o
r
d
s
,
s
en
te
n
ce
s
,
h
u
m
a
n
d
efin
ed
s
lid
in
g
w
i
n
d
o
w
o
r
an
n
-
g
r
a
m
[
6
-
1
2
]
.
T
h
e
Sy
n
tactica
l
m
eth
o
d
s
u
s
e
tex
t
’
s
s
y
n
tactica
l
u
n
i
ts
f
o
r
co
m
p
ar
in
g
t
h
e
s
i
m
ila
r
it
y
b
et
w
ee
n
d
o
cu
m
e
n
ts
.
I
m
p
l
icitl
y
,
w
e
co
n
s
id
er
th
at
s
i
m
ila
r
d
o
cu
m
e
n
t
s
w
o
u
ld
h
av
e
s
i
m
ilar
s
y
n
tac
tical
s
tr
u
ct
u
r
e.
T
h
is
m
et
h
o
d
m
a
k
es
u
s
e
o
f
c
h
ar
ac
ter
is
tic
s
s
u
c
h
a
s
P
OS
tag
to
co
m
p
ar
e
th
e
s
i
m
ilar
it
y
b
et
w
ee
n
d
i
f
f
er
e
n
t
d
o
cu
m
e
n
ts
[
1
3
]
.
T
h
e
Se
m
a
n
tic
m
et
h
o
d
s
u
s
e
a
s
e
m
a
n
tic
s
i
m
il
ar
it
y
f
o
r
co
m
p
ar
i
n
g
d
o
cu
m
en
ts
.
I
n
th
i
s
ap
p
r
o
ac
h
,
d
if
f
er
e
n
t
s
e
m
a
n
tic
f
ea
t
u
r
es
wh
ich
in
cl
u
d
e
(
S
y
n
o
n
y
m
s
,
h
y
p
o
n
y
m
s
,
h
y
p
er
n
y
m
s
,
s
e
m
a
n
tic
d
ep
en
d
en
cie
s
)
[
2
-
3
]
ar
e
ex
tr
ac
ted
f
r
o
m
t
h
e
s
o
u
r
ce
d
o
cu
m
e
n
ts
a
n
d
th
e
n
u
s
ed
to
tr
ac
e
o
u
t
t
h
e
p
lag
iar
is
m
ca
s
e
f
r
o
m
t
h
e
co
r
p
u
s
.
T
h
e
p
lag
iar
i
s
m
d
etec
tio
n
is
co
n
s
id
er
ed
as
a
p
ar
t
o
f
Natu
r
al
L
a
n
g
u
a
g
e
P
r
o
ce
s
s
in
g
(
NL
P
)
.
He
n
ce
,
b
ased
o
n
N
L
P
tec
h
n
iq
u
es
m
a
n
y
s
o
l
u
tio
n
s
h
av
e
b
ee
n
p
r
o
p
o
s
ed
f
o
r
le
x
ical
o
r
S
y
n
tactica
l p
la
g
iar
is
m
,
an
d
m
o
s
t a
r
e
b
ased
o
n
th
e
co
n
ce
p
t e
x
tr
ac
tio
n
u
s
i
n
g
a
co
r
p
u
s
lik
e
W
o
r
d
Net
[
1
4
-
1
6
]
.
W
ith
th
e
clas
s
ical
ap
p
r
o
ac
h
es,
t
w
o
d
o
cu
m
e
n
ts
t
h
at
s
h
ar
e
th
e
s
a
m
e
w
o
r
d
s
ar
e
co
n
s
id
er
ed
s
im
ilar
,
an
d
th
e
w
o
r
d
o
r
d
er
is
n
o
t r
esp
ec
te
d
w
h
ich
w
ill
m
a
k
e
lo
s
s
o
f
t
h
e
tr
u
e
m
ea
n
i
n
g
o
f
a
d
o
cu
m
e
n
t.
I
n
r
ec
en
t
y
ea
r
s
,
d
ee
p
lear
n
in
g
tec
h
n
iq
u
es
h
a
v
e
b
ee
n
th
e
s
u
b
j
ec
t
o
f
s
e
v
er
al
r
ese
ar
ch
es
an
d
i
n
d
if
f
er
en
t
d
o
m
ain
s
,
f
r
o
m
p
atter
n
r
ec
o
g
n
itio
n
to
N
L
P
p
r
o
b
le
m
s
.
T
h
e
h
i
g
h
p
er
f
o
r
m
a
n
ce
o
b
tain
ed
ar
e
v
er
y
en
co
u
r
ag
i
n
g
a
n
d
m
ak
e
it
p
o
s
s
ib
le
to
co
n
s
id
er
th
e
u
s
e
o
f
t
h
ese
tec
h
n
iq
u
es
i
n
t
h
e
f
ie
ld
o
f
p
lag
ia
r
is
m
d
etec
tio
n
[
1
7
-
1
8
]
.
T
h
e
tech
n
iq
u
es
b
ased
o
n
Dee
p
L
ea
r
n
i
n
g
f
o
r
p
lag
iar
is
m
d
etec
tio
n
,
in
cl
u
d
e
n
o
t
o
n
l
y
t
h
e
co
n
tex
t
u
al
(
s
e
m
an
t
ic)
lev
el
o
f
th
e
d
o
cu
m
e
n
t
b
u
t
o
ls
o
th
e
s
y
n
tactica
l
an
d
le
x
ic
al
lev
el
in
v
ec
to
r
r
ep
r
esen
tat
i
o
n
.
T
h
e
r
e
m
ain
d
er
o
f
th
is
p
a
p
er
is
o
r
g
an
ized
as
f
o
llo
w
s
.
T
h
e
f
ir
s
t
s
ec
tio
n
p
r
esen
t
s
b
ac
k
g
r
o
u
n
d
co
n
ce
p
t.
T
h
e
s
ec
o
n
d
s
ec
tio
n
d
ef
i
n
es
r
elat
ed
w
o
r
k
.
T
h
e
th
ir
d
s
ec
tio
n
co
n
tai
n
s
a
d
ee
p
an
al
y
s
e
co
n
ce
r
n
i
n
g
o
u
r
co
m
p
ar
is
o
n
s
tu
d
y
.
T
h
e
last
s
ec
t
io
n
i
n
tr
o
d
u
ce
s
t
h
e
co
n
cl
u
s
io
n
an
d
f
u
tu
r
e
w
o
r
k
.
2.
RE
S
E
ARCH
M
E
T
H
O
D
I
n
t
h
is
s
ec
tio
n
w
e
w
ill
m
en
tio
n
t
h
e
d
i
f
f
er
en
t
tec
h
n
iq
u
es
u
s
e
d
b
y
t
h
e
p
la
g
iar
is
m
d
etec
tio
n
ap
p
r
o
ac
h
es,
w
h
et
h
er
in
ter
m
s
o
f
it
s
r
ep
r
esen
tatio
n
o
f
its
te
x
t
s
o
r
th
e
m
et
h
o
d
s
th
o
s
e
ca
lcu
late
t
h
e
s
i
m
i
lar
it
y
:
a.
Neu
r
al
n
et
w
o
r
k
b
ased
m
o
d
els
W
o
r
d
em
b
ed
d
in
g
s
ar
e
a
t
y
p
e
o
f
w
o
r
d
r
ep
r
esen
tatio
n
w
h
ich
s
to
r
es
th
e
co
n
te
x
t
u
al
i
n
f
o
r
m
at
io
n
in
a
lo
w
-
d
i
m
en
s
io
n
al
v
ec
to
r
.
T
h
is
ap
p
r
o
ac
h
g
ai
n
ed
e
x
tr
e
m
e
p
o
p
u
lar
it
y
w
i
th
t
h
e
i
n
tr
o
d
u
ctio
n
o
f
W
o
r
d
2
Vec
in
2
0
1
3
,
g
r
o
u
p
s
o
f
m
o
d
el
s
to
lear
n
t
h
e
w
o
r
d
e
m
b
ed
d
in
g
s
i
n
a
co
m
p
u
tatio
n
a
ll
y
ef
f
ic
ien
t
w
a
y
.
A
n
d
Do
c2
Vec
ca
n
b
e
s
ee
n
a
n
ex
te
n
s
io
n
o
f
W
o
r
d
2
Vec
w
h
o
s
e
g
o
al
i
s
to
cr
ea
te
a
r
e
p
r
esen
tatio
n
al
v
ec
to
r
o
f
a
d
o
cu
m
e
n
t o
r
p
ar
ag
r
ap
h
.
Wo
rd2
v
ec
:
is
a
m
o
d
el
u
s
i
n
g
n
eu
r
al
n
et
w
o
r
k
u
s
ed
to
p
r
o
d
u
ce
a
d
is
tr
ib
u
ted
r
ep
r
esen
tati
o
n
o
f
w
o
r
d
.
So
m
e
r
esear
ch
er
s
a
y
s
t
h
at
i
s
n
o
t
d
ee
p
lear
n
i
n
g
tec
h
n
iq
u
e,
b
ec
au
s
e
it
is
s
i
m
p
le
b
i
-
la
y
er
ed
n
eu
r
al
n
et
w
o
r
k
ar
ch
itect
u
r
e.
T
h
is
m
o
d
el
is
s
h
allo
w
,
t
w
o
-
la
y
er
n
eu
r
al
n
et
w
o
r
k
s
th
a
t
ar
e
tr
ain
ed
to
r
ec
o
n
s
tr
u
ct
lin
g
u
is
t
ic
co
n
tex
t
s
o
f
w
o
r
d
s
.
W
o
r
d
2
v
ec
tak
es a
s
its
in
p
u
t a
l
ar
g
e
co
r
p
u
s
o
f
te
x
t a
n
d
p
r
o
d
u
ce
s
a
v
ec
to
r
s
p
ac
e,
t
y
p
icall
y
o
f
s
ev
er
al
h
u
n
d
r
ed
d
im
e
n
s
io
n
s
,
w
it
h
ea
c
h
u
n
iq
u
e
w
o
r
d
in
t
h
e
co
r
p
u
s
[
1
9
]
.
Do
c2
v
ec
:
Do
c
2
v
ec
is
an
u
n
s
u
p
er
v
is
ed
alg
o
r
ith
m
to
g
e
n
er
at
e
v
ec
to
r
s
r
ep
r
esen
tatio
n
o
f
s
e
n
ten
ce
s
,
p
ar
ag
r
ap
h
s
an
d
d
o
cu
m
e
n
ts
[
2
0
]
.
I
ts
m
o
d
el
is
b
ased
o
n
W
o
r
d
2
Vec
,
w
it
h
o
n
l
y
ad
d
in
g
an
o
t
h
er
v
ec
to
r
(
p
ar
ag
r
ap
h
I
D)
to
th
e
in
p
u
t.
T
h
e
ar
ch
itect
u
r
e
o
f
Do
c
2
Vec
m
o
d
el
is
s
h
o
w
n
Fi
g
u
r
e
1
.
Fig
u
r
e
1
.
Do
c2
v
ec
ar
ch
itectu
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
A
d
ee
p
lea
r
n
in
g
b
a
s
ed
tec
h
n
iq
u
e
fo
r
p
la
g
ia
r
is
m
d
etec
tio
n
:
a
co
mp
a
r
a
tive
s
tu
d
y
(
E
l Mo
s
ta
fa
Ha
mb
i
)
83
I
n
s
tead
o
f
u
s
in
g
j
u
s
t
n
ea
r
b
y
wo
r
d
s
to
p
r
ed
ict
th
e
w
o
r
d
,
w
e
a
ls
o
ad
d
ed
an
o
th
er
f
ea
t
u
r
e
v
ec
t
o
r
,
w
h
ic
h
is
d
o
cu
m
e
n
t
-
u
n
iq
u
e.
b.
Dee
p
lear
n
in
g
b
ased
m
o
d
els
Dee
p
lear
n
in
g
is
a
s
et
o
f
lear
n
in
g
m
et
h
o
d
s
atte
m
p
ti
n
g
to
m
o
d
el
d
ata
w
it
h
co
m
p
le
x
ar
ch
itect
u
r
es
co
m
b
i
n
i
n
g
d
if
f
er
e
n
t
n
o
n
-
li
n
ea
r
tr
a
n
s
f
o
r
m
atio
n
s
.
T
h
e
ele
m
e
n
tar
y
b
r
ick
s
o
f
d
ee
p
lear
n
i
n
g
ar
e
t
h
e
n
e
u
r
al
n
et
w
o
r
k
s
th
at
ar
e
co
m
b
i
n
ed
to
f
o
r
m
t
h
e
d
ee
p
n
eu
r
al
n
et
w
o
r
k
s
.
T
h
er
e
ex
is
t se
v
er
al
t
y
p
es o
f
ar
ch
itec
tu
r
es
f
o
r
n
eu
r
al
n
et
w
o
r
k
s
:
Rec
urs
iv
e
neura
l
net
w
o
rk
s
(
RNN)
:
h
av
e
b
ee
n
s
u
cc
e
s
s
f
u
l,
f
o
r
in
s
tan
ce
,
i
n
lear
n
i
n
g
s
eq
u
en
ce
a
n
d
tr
ee
s
tr
u
ct
u
r
es
in
n
at
u
r
al
la
n
g
u
a
g
e
p
r
o
ce
s
s
in
g
,
m
ai
n
l
y
p
h
r
ase
a
n
d
s
e
n
te
n
ce
co
n
t
in
u
o
u
s
r
ep
r
esen
tat
io
n
s
b
ased
o
n
w
o
r
d
e
m
b
ed
d
in
g
[
2
1
]
.
Sia
m
ese
L
ST
M
f
o
r
L
ea
rning
do
cu
m
e
nts
Si
m
ila
rit
y
:
L
S
T
M
is
a
k
in
g
o
f
r
ec
u
r
r
en
t
n
e
u
r
al
n
et
w
o
r
k
an
d
it
i
s
g
r
ea
t
w
h
en
w
e
h
av
e
a
n
e
n
ti
r
e
s
eq
u
en
ce
o
f
w
o
r
d
s
o
r
s
e
n
ten
ce
s
.
T
h
is
i
s
b
ec
au
s
e
R
N
Ns
ca
n
m
o
d
el
a
n
d
r
e
m
e
m
b
er
th
e
r
elatio
n
s
h
ip
s
b
et
w
ee
n
d
if
f
er
en
t
w
o
r
d
s
an
d
s
en
te
n
ce
s
.
Ma
n
h
at
tan
L
ST
M
m
o
d
els
h
a
v
e
t
w
o
n
et
w
o
r
k
s
L
ST
Mle
f
t
a
n
d
L
S
T
Mr
ig
h
t
w
h
ich
p
r
o
ce
s
s
o
n
e
o
f
th
e
s
e
n
te
n
ce
s
i
n
a
g
i
v
e
n
p
air
in
d
ep
en
d
en
tl
y
.
Sia
m
e
s
e
L
ST
M,
a
v
er
s
io
n
o
f
Ma
n
h
a
ttan
L
ST
M
w
h
er
e
b
o
th
L
ST
Mle
f
t a
n
d
L
ST
Mr
ig
h
t
h
av
e
s
a
m
e
tied
w
ei
g
h
ts
s
u
c
h
t
h
at
L
ST
Mle
f
t
=
L
ST
Mr
ig
h
t.
S
u
c
h
a
m
o
d
el
i
s
u
s
e
f
u
l
f
o
r
task
s
li
k
e
d
u
p
licate
q
u
e
r
y
d
etec
tio
n
a
n
d
q
u
er
y
r
an
k
i
n
g
.
Her
e,
d
u
p
licate
d
ete
ctio
n
tas
k
i
s
p
er
f
o
r
m
ed
to
f
i
n
d
if
t
w
o
d
o
cu
m
e
n
ts
ar
e
s
i
m
ilar
o
r
n
o
t.
Si
m
i
lar
m
o
d
el
ca
n
b
e
tr
ai
n
ed
f
o
r
q
u
er
y
r
an
k
i
n
g
u
s
i
n
g
h
it
d
ata
f
o
r
a
g
iv
e
n
q
u
er
y
a
n
d
it
s
m
atch
in
g
r
esu
lt
s
as
a
p
r
o
x
y
f
o
r
s
i
m
ilar
it
y
[
2
1
]
.
Co
nv
o
lutio
na
l
neura
l
net
w
o
rk
:
C
NN
is
a
clas
s
o
f
d
ee
p
,
f
ee
d
-
f
o
r
w
ar
d
ar
tif
icial
n
eu
r
al
n
et
w
o
r
k
s
th
a
t
u
s
es
a
v
ar
iatio
n
o
f
m
u
lt
ila
y
er
p
er
ce
p
tio
n
s
d
esi
g
n
ed
to
r
eq
u
ir
e
m
i
n
i
m
al
p
r
ep
r
o
ce
s
s
in
g
.
T
h
ese
ar
e
in
s
p
ir
ed
b
y
a
n
i
m
al
v
is
u
al
co
r
tex
.
C
NN
s
ar
e
g
e
n
er
all
y
u
s
ed
i
n
co
m
p
u
ter
v
is
io
n
;
h
o
w
e
v
er
,
th
e
y
h
av
e
r
ec
e
n
t
l
y
b
ee
n
ap
p
lied
to
v
ar
io
u
s
N
L
P
tas
k
s
li
k
e
a
tex
t c
lass
i
f
icatio
n
[
2
1
]
.
Dee
p
Str
uct
ured
Se
m
a
ntic
M
o
del
(
DSSM
)
:
DSSM
s
tan
d
s
f
o
r
Dee
p
Stru
ct
u
r
ed
Se
m
a
n
tic
Mo
d
el,
o
r
m
o
r
e
g
en
er
al,
Dee
p
Se
m
a
n
tic
Si
m
ilar
it
y
Mo
d
el.
I
t
is
a
d
ee
p
n
eu
r
al
n
et
w
o
r
k
(
DNN)
m
o
d
el
lin
g
tec
h
n
iq
u
e
f
o
r
r
ep
r
esen
tin
g
te
x
t
s
tr
i
n
g
s
(
s
e
n
t
en
ce
s
,
q
u
er
ies,
p
r
ed
icate
s
,
e
n
ti
t
y
m
e
n
tio
n
s
,
etc.
)
i
n
a
co
n
ti
n
u
o
u
s
s
e
m
a
n
tic
s
p
ac
e
an
d
m
o
d
ellin
g
s
e
m
a
n
tic
s
i
m
ila
r
it
y
b
et
w
ee
n
t
w
o
te
x
t stri
n
g
s
.
c.
Oth
er
m
o
d
els
Oth
er
m
e
th
o
d
s
u
s
ed
to
co
n
s
tr
u
ct
a
v
ec
to
r
r
ep
r
esen
tatio
n
o
f
a
g
iv
e
n
te
x
t c
an
b
e
f
o
u
n
d
:
G
L
O
V
E
:
i
s
a
n
u
n
s
u
p
er
v
i
s
ed
lear
n
in
g
al
g
o
r
ith
m
f
o
r
o
b
tain
i
n
g
v
ec
to
r
r
ep
r
esen
tatio
n
s
f
o
r
w
o
r
d
s
.
T
r
ain
in
g
i
s
p
er
f
o
r
m
ed
o
n
a
g
g
r
e
g
ated
g
l
o
b
al
w
o
r
d
-
w
o
r
d
co
-
o
cc
u
r
r
en
ce
s
tatis
tic
s
f
r
o
m
a
co
r
p
u
s
,
an
d
th
e
r
esu
lt
in
g
r
ep
r
esen
tatio
n
s
s
h
o
w
ca
s
e
in
ter
esti
n
g
li
n
ea
r
s
u
b
s
tr
u
ct
u
r
es o
f
t
h
e
w
o
r
d
v
ec
to
r
s
p
ac
e
[
2
2
]
.
I
nfe
rSent
:
is
a
s
e
n
ten
ce
e
m
b
e
d
d
in
g
s
m
et
h
o
d
th
at
p
r
o
v
id
es
s
e
m
an
tic
r
ep
r
esen
tatio
n
s
f
o
r
E
n
g
l
is
h
s
e
n
ten
ce
s
.
I
t
is
tr
ain
ed
o
n
n
a
tu
r
al
la
n
g
u
a
g
e
in
f
er
en
ce
d
ata
an
d
g
e
n
er
alize
s
w
ell
to
m
an
y
d
if
f
er
e
n
t ta
s
k
s
[
2
2
]
.
d.
Si
m
i
lar
it
y
m
et
h
o
d
s
Fin
d
i
n
g
s
i
m
ilar
it
y
b
et
w
ee
n
e
le
m
e
n
ts
i
s
th
e
co
r
e
o
f
s
en
te
n
ce
s
i
m
ilar
it
y
.
I
n
th
e
liter
at
u
r
e,
th
er
e
ar
e
m
an
y
m
etr
ics
f
o
r
ca
lcu
lati
n
g
s
i
m
il
ar
it
y
.
T
h
is
s
ec
tio
n
s
h
o
w
s
d
if
f
er
en
t
ap
p
r
o
ac
h
es
u
s
ed
to
ca
lcu
late
s
i
m
i
lar
it
y
b
et
w
ee
n
ele
m
e
n
t
s
:
Co
s
ine
s
i
m
ila
rit
y
:
is
a
m
ea
s
u
r
e
o
f
s
i
m
ilar
it
y
b
et
w
ee
n
t
w
o
n
o
n
-
ze
r
o
v
ec
to
r
s
o
f
a
n
i
n
n
er
p
r
o
d
u
ct
s
p
ac
e
th
a
t
m
ea
s
u
r
es
t
h
e
co
s
in
e
o
f
th
e
a
n
g
le
b
et
w
ee
n
th
e
m
.
T
h
e
co
s
in
e
o
f
0
°
is
1
,
an
d
it
is
less
th
an
1
f
o
r
an
y
an
g
le
i
n
th
e
in
ter
v
a
l [
0
,
π]
r
ad
ian
s
[
2
3
]
.
J
a
cc
a
rd
ind
ex
:
also
k
n
o
w
n
a
s
I
n
ter
s
ec
tio
n
o
v
er
U
n
io
n
a
n
d
t
h
e
J
ac
ca
r
d
s
i
m
ilar
it
y
co
ef
f
icie
n
t (
o
r
ig
i
n
all
y
g
i
v
e
n
th
e
Fre
n
c
h
n
a
m
e
co
ef
f
icie
n
t
d
e
co
m
m
u
n
it
y
b
y
P
au
l
J
ac
ca
r
d
)
,
is
a
s
tati
s
tic
u
s
ed
f
o
r
g
au
g
i
n
g
t
h
e
s
i
m
ilar
it
y
a
n
d
d
iv
er
s
it
y
o
f
s
a
m
p
le
s
et
s
.
T
h
e
J
ac
ca
r
d
co
ef
f
icien
t
m
ea
s
u
r
e
s
s
i
m
ilar
it
y
b
et
w
ee
n
f
i
n
ite
s
a
m
p
le
s
ets an
d
i
s
d
ef
i
n
ed
as th
e
s
ize
o
f
th
e
i
n
ter
s
ec
tio
n
d
iv
id
ed
b
y
t
h
e
s
ize
o
f
t
h
e
u
n
io
n
o
f
t
h
e
s
a
m
p
le
s
ets [
2
3
]
.
E
ucli
dea
n
Dis
t
a
nce:
r
ef
er
s
to
E
u
clid
ea
n
d
is
tan
ce
.
W
h
e
n
d
ata
is
d
en
s
e
o
r
c
o
n
tin
u
o
u
s
,
th
is
i
s
th
e
b
est
p
r
o
x
i
m
it
y
m
ea
s
u
r
e.
T
h
e
E
u
cli
d
ea
n
d
is
ta
n
ce
b
et
w
ee
n
t
w
o
p
o
in
ts
is
t
h
e
le
n
g
t
h
o
f
th
e
p
at
h
c
o
n
n
ec
ti
n
g
t
h
e
m
,
a
n
d
it is
o
b
tain
ed
w
it
h
th
e
P
y
t
h
a
g
o
r
ea
n
T
h
eo
r
em
[
2
3
]
.
L
o
ng
est
co
mm
o
n
s
u
bs
e
qu
en
ce
(
L
CS)
m
et
ho
d
:
co
n
s
i
s
ts
o
f
f
i
n
d
in
g
th
e
lo
n
g
est
s
u
b
s
eq
u
e
n
ce
co
m
m
o
n
to
al
l
s
eq
u
en
ce
s
i
n
a
s
et
o
f
s
eq
u
en
ce
s
.
T
h
e
lo
n
g
e
s
t
co
m
m
o
n
s
u
b
s
eq
u
en
ce
p
r
o
b
le
m
i
s
a
clas
s
ic
co
m
p
u
ter
s
c
ien
ce
p
r
o
b
lem
,
th
e
b
asi
s
o
f
d
ata
co
m
p
ar
i
s
o
n
p
r
o
g
r
a
m
s
s
u
ch
as
t
h
e
d
if
f
u
tili
t
y
an
d
h
a
s
ap
p
licati
o
n
s
in
co
m
p
u
tatio
n
al
lin
g
u
i
s
tic
s
an
d
b
io
in
f
o
r
m
atic
s
[
2
4
]
.
Wo
rd
M
o
v
er
’
s
Dis
t
a
nce
(
WM
D)
:
u
s
es
w
o
r
d
e
m
b
ed
d
in
g
s
to
ca
lcu
late
th
e
s
i
m
ilar
itie
s
,
an
d
p
r
ec
is
el
y
,
it
u
s
es
n
o
r
m
alize
d
B
ag
-
of
-
w
o
r
d
s
an
d
w
o
r
d
E
m
b
ed
d
in
g
s
to
ca
lcu
late
th
e
d
is
ta
n
ce
b
et
w
ee
n
d
o
cu
m
e
n
ts
[
2
5
]
.
3.
RE
L
AT
E
D
WO
RK
Ou
r
s
t
u
d
y
f
o
cu
s
e
s
o
n
t
h
e
d
et
ec
tio
n
o
f
s
e
m
a
n
tic
p
lag
iar
is
m
m
o
r
e
p
r
ec
is
el
y
t
h
e
id
en
ti
f
ica
tio
n
o
f
t
h
e
p
lag
iar
is
m
o
f
id
ea
s
b
et
w
ee
n
t
w
o
g
iv
e
n
te
x
t
s
,
as
il
lu
s
tr
ated
b
elo
w
w
e
d
u
g
o
n
m
e
th
o
d
s
t
h
a
t
d
etec
t
th
i
s
t
y
p
e
o
f
p
lag
iar
is
m
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
1
,
Ma
r
ch
20
20
:
81
–
90
84
In
[
2
6
]
p
r
o
p
o
s
ed
a
p
lag
iar
is
m
d
etec
tio
n
s
y
s
te
m
,
w
h
ich
r
el
y
o
n
u
s
e
s
en
te
n
ce
s
co
m
p
ar
is
o
n
i
n
t
w
o
p
h
ases
.
T
h
e
y
f
ir
s
t
ex
tr
ac
t
w
o
r
d
v
ec
to
r
s
b
y
w
o
r
d
2
v
ec
alg
o
r
i
th
m
,
a
n
d
th
e
n
r
e
m
o
v
e
P
er
s
ia
n
s
to
p
w
o
r
d
s
w
h
ile
tex
t
p
r
e
-
p
r
o
ce
s
s
i
n
g
.
Af
ter
th
at
,
f
o
r
ea
ch
s
en
te
n
ce
an
av
er
a
g
e
o
f
all
w
o
r
d
v
ec
to
r
s
is
ca
l
cu
l
ated
.
A
f
ter
f
ea
tu
r
e
ex
tr
ac
tio
n
,
in
p
h
a
s
e
1
,
ea
ch
s
en
te
n
ce
in
a
s
u
s
p
icio
u
s
d
o
c
u
m
e
n
t
is
co
m
p
ar
ed
w
i
th
all
t
h
e
s
en
te
n
ce
s
in
th
e
s
o
u
r
ce
d
o
cu
m
en
t
s
.
C
o
s
in
e
s
i
m
ilar
it
y
i
s
u
s
ed
as
a
co
m
p
ar
is
o
n
m
e
tr
ic.
A
f
ter
th
is
s
tep
w
h
i
ch
h
elp
s
to
f
i
n
d
th
e
n
ea
r
est
s
en
ten
ce
s
i
n
r
ea
l
ti
m
e,
in
p
h
ase
2
,
le
x
ical
s
i
m
ilar
i
t
y
o
f
t
w
o
s
en
te
n
ce
s
is
ev
al
u
a
ted
b
y
t
h
e
J
ac
ca
r
d
s
i
m
ilar
it
y
m
ea
s
u
r
e.
T
w
o
s
e
n
te
n
ce
s
w
h
ic
h
p
ass
J
ac
ca
r
d
s
i
m
i
l
ar
it
y
th
r
e
s
h
o
ld
co
n
s
id
er
ed
as
p
lag
iar
is
m
at
f
in
a
l
s
tep
.
In
[
2
7
]
p
r
o
p
o
s
ed
th
e
u
s
e
w
o
r
d
2
v
ec
m
o
d
el
in
o
r
d
er
to
co
m
p
u
te
v
ec
to
r
o
f
f
ea
tu
r
e
s
f
o
r
ev
er
y
w
o
r
d
.
T
h
ey
ch
o
o
s
e
d
o
cu
m
en
ts
f
r
o
m
t
h
e
c
o
r
p
u
s
its
e
lf
,
h
o
w
e
v
er
t
h
e
d
o
cu
m
en
ts
u
s
ed
f
o
r
test
i
n
g
w
a
s
p
r
o
ce
s
s
ed
an
d
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
t
h
at
w
as
m
ad
e
is
s
to
p
w
o
r
d
s
r
em
o
v
al.
T
h
e
s
i
m
ilar
it
y
b
et
w
ee
n
v
ec
to
r
s
w
a
s
co
m
p
u
ted
b
y
u
s
i
n
g
co
s
in
e
s
i
m
ilar
it
y
.
[
2
4
]
T
h
e
ai
m
o
f
t
h
is
ap
p
r
o
ac
h
i
s
ev
alu
ati
n
g
th
e
v
al
id
it
y
o
f
u
s
i
n
g
th
e
d
is
tr
ib
u
te
d
r
ep
r
esen
tatio
n
to
d
ef
i
n
e
t
h
e
w
o
r
d
s
i
m
ilar
it
y
.
T
h
e
y
in
tr
o
d
u
ce
th
r
ee
m
et
h
o
d
s
b
ased
o
n
th
e
f
o
llo
w
i
n
g
t
h
r
ee
d
o
cu
m
en
t
s
i
m
ilar
ities
:
f
o
r
t
w
o
d
o
cu
m
e
n
t
s
:
T
h
e
len
g
t
h
o
f
th
e
lo
n
g
est
co
m
m
o
n
s
u
b
s
eq
u
e
n
c
e
(
L
C
S)
d
iv
id
ed
b
y
th
e
le
n
g
t
h
o
f
t
h
e
s
h
o
r
ter
d
o
cu
m
e
n
t,
th
e
lo
ca
l
m
a
x
i
m
al
v
al
u
e
o
f
t
h
e
le
n
g
th
o
f
L
C
S,
an
d
th
e
lo
ca
l
m
a
x
i
m
a
l
v
alu
e
o
f
t
h
e
w
e
ig
h
ted
le
n
g
th
o
f
L
C
S.
T
h
e
d
is
tr
ib
u
ted
r
ep
r
esen
tat
io
n
w
as
o
b
tai
n
ed
f
r
o
m
n
o
p
ar
ticu
lar
d
a
ta
b
y
w
o
r
d
2
v
ec
.
An
o
th
er
ap
p
r
o
ac
h
u
s
e
s
t
h
e
p
r
i
n
cip
le
o
f
Dee
p
Str
u
ct
u
r
ed
Se
m
an
tic
Mo
d
el
(
DS
SM)
p
r
o
p
o
s
ed
b
y
[
2
8
]
.
DSSM i
s
a
d
ee
p
lear
n
in
g
-
b
ase
d
tech
n
iq
u
e
th
a
t is p
r
o
p
o
s
ed
f
o
r
s
e
m
an
tic
u
n
d
er
s
ta
n
d
in
g
o
f
t
ex
tu
a
l d
ata.
I
t
m
ap
s
s
h
o
r
t
tex
t
u
al
s
tr
i
n
g
s
,
s
u
c
h
as
s
en
te
n
ce
s
,
to
f
ea
t
u
r
e
v
ec
to
r
s
in
a
lo
w
-
d
i
m
e
n
s
io
n
al
s
e
m
a
n
ti
c
s
p
ac
e.
T
h
en
th
e
v
ec
to
r
r
ep
r
esen
tatio
n
s
ar
e
u
tili
ze
d
f
o
r
d
o
cu
m
en
t r
etr
iev
al
b
y
co
m
p
ar
i
n
g
th
e
s
i
m
ilar
it
y
b
et
wee
n
d
o
cu
m
en
t
s
a
n
d
q
u
er
ies.
Af
ter
o
b
tain
i
n
g
th
e
s
e
m
an
tic
f
ea
t
u
r
e
v
ec
to
r
s
f
o
r
ea
ch
p
air
ed
s
n
ip
p
e
ts
o
f
tex
t,
c
o
s
in
e
s
i
m
ilar
it
y
is
u
tili
ze
d
to
m
ea
s
u
r
e
t
h
e
s
e
m
a
n
tic
s
i
m
i
lar
it
y
b
et
w
ee
n
th
e
p
a
ir
.
Si
m
ilar
l
y
,
w
it
h
th
e
p
r
ev
io
u
s
m
et
h
o
d
s
,
in
[
2
9
]
d
ee
p
lear
n
in
g
d
o
cu
m
e
n
ts
o
r
t
ex
ts
ca
n
b
e
r
ep
r
esen
ted
as
v
e
cto
r
s
b
y
t
h
e
u
s
i
n
g
d
o
cu
m
en
t
to
v
ec
to
r
tech
n
iq
u
e
(
d
o
c2
v
ec
)
.
A
n
d
t
h
e
d
etec
tio
n
o
f
p
lag
iar
is
m
w
i
ll
b
e
d
o
n
e
b
y
a
s
i
m
p
le
co
m
p
ar
is
o
n
b
et
w
e
en
all
s
e
n
ten
ce
s
o
f
ea
ch
t
w
o
d
o
cu
m
e
n
ts
a
n
al
y
s
ed
.
T
h
e
ap
p
r
o
ac
h
p
r
o
p
o
s
ed
in
[
3
0
]
is
b
ased
o
n
co
n
v
er
ti
n
g
a
p
ar
ag
r
ap
h
to
v
ec
to
r
s
a
n
d
it's
i
n
s
p
i
r
ed
b
y
th
e
m
et
h
o
d
s
f
o
r
lear
n
in
g
t
h
e
w
o
r
d
v
e
cto
r
s
.
T
h
e
in
s
p
ir
atio
n
i
s
th
at
th
e
w
o
r
d
v
ec
to
r
s
ar
e
ask
e
d
to
co
n
tr
ib
u
te
to
a
p
r
ed
ictio
n
task
ab
o
u
t
th
e
n
e
x
t
w
o
r
d
in
t
h
e
s
e
n
te
n
ce
.
So
,
d
es
p
ite
th
e
f
ac
t
t
h
at
t
h
e
w
o
r
d
v
ec
to
r
s
ar
e
in
itialized
r
an
d
o
m
l
y
,
t
h
e
y
ca
n
ev
e
n
t
u
all
y
ca
p
tu
r
e
s
e
m
a
n
tic
s
as
a
n
i
n
d
ir
ec
t
r
esu
lt
o
f
t
h
e
p
r
ed
ictio
n
t
ask
.
I
t
w
i
ll
u
s
e
t
h
is
id
ea
in
th
eir
p
ar
ag
r
ap
h
v
ec
to
r
s
in
a
s
i
m
i
lar
m
a
n
n
er
.
T
h
e
p
ar
ag
r
ap
h
v
ec
to
r
s
ar
e
also
as
k
ed
to
co
n
tr
ib
u
te
to
th
e
p
r
ed
ictio
n
task
o
f
t
h
e
n
e
x
t
w
o
r
d
g
iv
e
n
m
a
n
y
co
n
te
x
ts
s
a
m
p
le
d
f
r
o
m
th
e
p
ar
ag
r
ap
h
.
T
h
ese
ap
p
r
o
ac
h
es
[
2
9
-
3
0
]
ar
e
u
s
ed
to
p
er
f
o
r
m
s
i
m
ilar
it
y
d
etec
tio
n
b
et
w
ee
n
t
h
e
d
o
cu
m
e
n
t
v
ec
to
r
s
b
u
t
also
u
s
e
t
h
e
co
s
in
e
to
co
m
p
ar
e
t
h
e
v
ec
to
r
s
.
I
n
p
ap
er
[
3
1
]
th
ey
r
ep
r
esen
t
ea
c
h
w
o
r
d
w
b
y
a
v
ec
to
r
.
I
t
co
n
s
tr
u
ct
s
t
h
ese
w
o
r
d
v
ec
to
r
s
u
s
in
g
G
lo
Ve.
T
h
is
ap
p
r
o
ac
h
u
s
e
s
t
h
e
r
ec
u
r
s
i
v
e
n
eu
r
al
n
et
w
o
r
k
s
al
g
o
r
ith
m
to
h
av
e
a
v
ec
to
r
r
ep
r
esen
tatio
n
o
f
a
s
e
n
te
n
ce
a
n
d
u
s
e
t
h
e
co
s
in
e
f
o
r
ca
lcu
la
te
t
h
e
s
i
m
ilar
i
t
y
.
I
n
[
3
2
]
t
w
o
in
p
u
t
s
en
te
n
ce
s
ar
e
p
r
o
ce
s
s
ed
i
n
p
ar
allel
b
y
id
en
tica
l
n
e
u
r
al
n
e
t
w
o
r
k
s
,
o
u
tp
u
tti
n
g
s
en
ten
ce
r
ep
r
esen
tatio
n
s
.
T
h
e
s
en
te
n
ce
r
ep
r
esen
tatio
n
s
ar
e
co
m
p
ar
ed
b
y
t
h
e
s
tr
u
ct
u
r
ed
s
i
m
ilar
it
y
m
ea
s
u
r
e
m
e
n
t
la
y
er
.
T
h
e
s
i
m
ilar
it
y
f
ea
t
u
r
es
ar
e
th
en
p
ass
ed
to
a
f
u
ll
y
-
co
n
n
ec
ted
la
y
er
f
o
r
co
m
p
u
ti
n
g
th
e
s
i
m
ilar
it
y
s
co
r
e.
C
o
s
i
n
e
d
is
t
an
ce
m
ea
s
u
r
es
t
h
e
d
is
tan
ce
o
f
t
w
o
v
ec
to
r
s
ac
co
r
d
in
g
to
th
e
an
g
le
b
et
w
ee
n
t
h
e
m
.
T
h
e
u
s
e
o
f
co
s
i
n
e
to
d
etec
t
s
i
m
ilar
it
y
b
et
w
ee
n
s
en
te
n
ce
s
r
e
m
ain
s
a
s
o
lu
tio
n
th
at
ca
r
r
ies
m
a
n
y
r
is
k
s
.
I
n
f
er
Sen
t
[
2
2
]
is
an
NL
P
tech
n
iq
u
e
f
o
r
u
n
iv
er
s
a
l
s
en
te
n
ce
r
ep
r
esen
tatio
n
d
ev
el
o
p
ed
b
y
Face
b
o
o
k
th
at
u
s
es
s
u
p
er
v
i
s
ed
tr
ain
i
n
g
to
p
r
o
d
u
c
e
h
ig
h
tr
an
s
f
er
ab
le
r
ep
r
esen
tatio
n
s
.
T
h
e
y
u
s
ed
a
B
i
-
d
ir
ec
tio
n
al
L
ST
M
w
it
h
atte
n
tio
n
t
h
at
co
n
s
is
te
n
tl
y
s
u
r
p
ass
ed
m
a
n
y
u
n
s
u
p
er
v
is
ed
t
r
ain
i
n
g
m
et
h
o
d
s
s
u
c
h
as
th
e
S
k
ip
T
h
o
u
g
h
t
v
ec
to
r
s
.
T
h
e
y
also
p
r
o
v
id
e
a
P
y
to
r
ch
i
m
p
le
m
en
ta
tio
n
th
at
th
e
y
u
s
ed
to
g
en
er
ate
s
e
n
ten
ce
e
m
b
ed
d
i
n
g
.
So
,
t
h
is
a
p
p
r
o
ac
h
n
ee
d
s
to
d
e
f
i
n
e
a
s
i
m
ilar
it
y
m
ea
s
u
r
e
to
co
m
p
ar
e
t
w
o
v
ec
to
r
s
,
a
n
d
f
o
r
th
at
g
o
al,
it
’
ll b
e
th
e
co
s
i
n
e
s
i
m
ilar
it
y
.
T
h
e
au
th
o
r
s
i
n
[
3
3
]
u
s
ed
w
o
r
d
e
m
b
ed
d
in
g
,
v
ec
to
r
r
e
p
r
esen
tatio
n
s
o
f
ter
m
s
,
co
m
p
u
ted
f
r
o
m
u
n
lab
elled
d
ata,
t
h
at
r
ep
r
esen
t
ter
m
s
i
n
a
s
e
m
a
n
tic
s
p
ac
e
i
n
w
h
ic
h
p
r
o
x
i
m
it
y
o
f
v
ec
to
r
s
ca
n
b
e
in
ter
p
r
eted
as
s
e
m
a
n
tic
s
i
m
ilar
it
y
.
T
h
e
y
p
r
o
p
o
s
e
to
g
o
f
r
o
m
w
o
r
d
-
le
v
el
to
tex
t
-
le
v
el
s
e
m
a
n
tics
b
y
co
m
b
in
i
n
g
i
n
s
ig
h
t
s
f
r
o
m
m
et
h
o
d
s
b
ased
o
n
ex
ter
n
al
s
o
u
r
ce
s
o
f
s
e
m
a
n
tic
k
n
o
w
led
g
e
w
it
h
w
o
r
d
e
m
b
ed
d
in
g
.
T
h
e
y
d
er
iv
e
m
u
ltip
le
t
y
p
es
o
f
m
eta
-
f
ea
t
u
r
es
f
r
o
m
t
h
e
co
m
p
ar
i
s
o
n
o
f
th
e
w
o
r
d
v
ec
to
r
s
f
o
r
s
h
o
r
t
tex
t
p
air
s
,
a
n
d
f
r
o
m
th
e
v
ec
to
r
m
ea
n
s
o
f
th
eir
r
esp
ec
ti
v
e
w
o
r
d
e
m
b
ed
d
in
g
.
T
h
e
f
ea
t
u
r
es
r
ep
r
esen
ti
n
g
lab
elled
s
h
o
r
t
te
x
t
p
air
s
ar
e
u
s
ed
to
tr
ai
n
a
s
u
p
er
v
i
s
ed
lear
n
i
n
g
al
g
o
r
ith
m
.
In
[
2
5
]
p
r
esen
t
t
h
e
W
o
r
d
Mo
v
er
’
s
Di
s
tan
ce
(
W
MD
)
,
a
n
o
v
el
d
is
tan
ce
f
u
n
ctio
n
b
et
w
ee
n
tex
t
d
o
cu
m
en
t
s
.
T
h
i
s
w
o
r
k
i
s
b
ased
o
n
r
ec
en
t
r
e
s
u
lt
s
in
w
o
r
d
e
m
b
ed
d
in
g
th
a
t
lear
n
s
e
m
a
n
ticall
y
m
ea
n
in
g
f
u
l
r
ep
r
esen
tatio
n
s
f
o
r
w
o
r
d
s
f
r
o
m
lo
ca
l
co
-
o
cc
u
r
r
e
n
ce
s
in
s
e
n
te
n
ce
s
.
T
h
e
W
MD
d
is
tan
ce
m
ea
s
u
r
es
th
e
d
is
s
i
m
ilar
it
y
b
et
w
ee
n
t
w
o
tex
t
d
o
cu
m
e
n
t
s
as
th
e
m
in
i
m
u
m
a
m
o
u
n
t
o
f
d
is
ta
n
c
e
t
h
at
t
h
e
e
m
b
ed
d
ed
w
o
r
d
s
o
f
o
n
e
d
o
cu
m
en
t
n
ee
d
to
“
tr
av
el”
to
r
ea
ch
th
e
e
m
b
ed
d
ed
w
o
r
d
s
o
f
an
o
th
er
d
o
cu
m
e
n
t.
T
h
is
ar
ticle
[
3
4
]
p
r
o
p
o
s
ed
an
in
n
o
v
a
tiv
e
w
o
r
d
e
m
b
ed
d
in
g
-
b
ased
s
y
s
te
m
d
e
v
o
ted
to
ca
lcu
la
tin
g
t
h
e
s
e
m
an
tic
s
i
m
ilar
it
y
i
n
A
r
ab
ic
s
e
n
te
n
ce
s
.
T
h
e
m
ai
n
i
d
ea
is
to
ex
p
lo
it
v
ec
to
r
s
a
s
wo
r
d
r
ep
r
esen
tatio
n
s
i
n
a
m
u
lti
d
i
m
en
s
io
n
al
s
p
ac
e
i
n
o
r
d
er
to
ca
p
tu
r
e
th
e
s
e
m
a
n
tic
an
d
s
y
n
tactic
p
r
o
p
er
ties
o
f
wo
r
d
s
.
I
DF
w
e
ig
h
ti
n
g
a
n
d
P
ar
t
-
of
-
Sp
ee
c
h
ta
g
g
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
A
d
ee
p
lea
r
n
in
g
b
a
s
ed
tec
h
n
iq
u
e
fo
r
p
la
g
ia
r
is
m
d
etec
tio
n
:
a
co
mp
a
r
a
tive
s
tu
d
y
(
E
l Mo
s
ta
fa
Ha
mb
i
)
85
ar
e
ap
p
lied
o
n
th
e
ex
a
m
i
n
ed
s
en
te
n
ce
s
to
s
u
p
p
o
r
t
th
e
id
en
t
if
icatio
n
o
f
w
o
r
d
s
t
h
at
ar
e
h
i
g
h
l
y
d
escr
ip
ti
v
e
i
n
ea
ch
s
e
n
ten
ce
.
I
n
p
ap
er
[
3
5
]
th
e
y
ad
d
r
ess
th
e
is
s
u
e
o
f
f
in
d
i
n
g
an
e
f
f
ec
ti
v
e
v
ec
to
r
r
ep
r
esen
tatio
n
f
o
r
a
v
er
y
s
h
o
r
t
tex
t
f
r
ag
m
e
n
t.
B
y
e
f
f
ec
ti
v
e
t
h
e
y
m
ea
n
t
h
at
t
h
e
r
ep
r
esen
tat
io
n
s
h
o
u
ld
g
r
asp
m
o
s
t
o
f
t
h
e
s
e
m
an
tic
i
n
f
o
r
m
a
tio
n
i
n
th
at
f
r
ag
m
e
n
t.
Fo
r
th
i
s
,
t
h
e
y
u
s
e
s
e
m
a
n
tic
w
o
r
d
e
m
b
ed
d
i
n
g
to
r
ep
r
esen
t
i
n
d
iv
id
u
al
w
o
r
d
s
,
an
d
w
e
lear
n
h
o
w
to
w
e
ig
h
e
v
er
y
w
o
r
d
in
t
h
e
tex
t
th
r
o
u
g
h
t
h
e
u
s
e
o
f
tf
-
id
f
(
ter
m
f
r
eq
u
e
n
c
y
-
i
n
v
er
s
e
d
o
cu
m
e
n
t
f
r
eq
u
en
c
y
)
in
f
o
r
m
atio
n
to
ar
r
iv
e
at
an
o
v
er
all
r
ep
r
esen
tatio
n
o
f
t
h
e
f
r
ag
m
e
n
t
co
m
p
ar
in
g
t
w
o
t
f
-
id
f
v
ec
to
r
s
i
s
d
o
n
e
th
r
o
u
g
h
a
s
tan
d
ar
d
co
s
in
e
s
i
m
ilar
it
y
.
[
3
6
]
T
h
is
p
ap
er
in
v
esti
g
a
tes
t
h
e
ef
f
ec
ti
v
e
n
es
s
o
f
s
ev
er
al
s
u
c
h
n
ai
v
e
tech
n
iq
u
es,
as
w
ell
as
tr
ad
itio
n
al
t
f
-
id
f
s
i
m
ilar
it
y
,
f
o
r
f
r
a
g
m
en
ts
o
f
d
if
f
er
en
t le
n
g
th
s
.
T
h
is
m
ai
n
co
n
tr
ib
u
tio
n
i
s
a
f
ir
s
t
s
tep
to
w
ar
d
s
a
h
y
b
r
id
m
eth
o
d
th
at
co
m
b
i
n
e
s
th
e
s
tr
en
g
th
o
f
d
en
s
e
d
is
tr
ib
u
ted
r
ep
r
esen
tatio
n
s
-
as
o
p
p
o
s
ed
to
s
p
ar
s
e
ter
m
m
atc
h
i
n
g
-
w
i
th
th
e
s
tr
e
n
g
th
o
f
tf
-
id
f
b
ased
m
et
h
o
d
s
to
a
u
to
m
atica
ll
y
r
ed
u
ce
t
h
e
i
m
p
ac
t
o
f
les
s
i
n
f
o
r
m
ati
v
e
ter
m
s
.
T
h
is
ap
p
r
o
ac
h
o
u
tp
er
f
o
r
m
s
t
h
e
ex
is
t
in
g
tec
h
n
iq
u
es
in
a
to
y
e
x
p
er
i
m
e
n
tal
s
et
-
u
p
,
lead
in
g
to
th
e
co
n
cl
u
s
io
n
th
a
t
th
e
co
m
b
in
at
io
n
o
f
w
o
r
d
e
m
b
ed
d
i
n
g
a
n
d
t
f
-
id
f
i
n
f
o
r
m
at
io
n
m
ig
h
t
lead
to
a
b
etter
m
o
d
el
f
o
r
s
e
m
a
n
tic
co
n
ten
t
w
ith
in
v
er
y
s
h
o
r
t
tex
t
f
r
ag
m
e
n
t
s
.
B
et
w
ee
n
t
w
o
s
u
c
h
r
ep
r
esen
tatio
n
s
th
e
y
th
en
ca
lc
u
late
t
h
e
co
s
i
n
e
s
i
m
il
ar
it
y
.
I
n
t
h
e
ar
ch
itectu
r
e
p
r
o
p
o
s
ed
in
[
3
7
]
,
w
o
r
d
e
m
b
ed
d
in
g
is
f
i
r
s
t
tr
ain
ed
o
n
A
P
I
d
o
cu
m
en
t
s
,
tu
to
r
ials
,
an
d
r
ef
er
en
ce
d
o
cu
m
e
n
ts
,
an
d
th
en
ag
g
r
eg
ated
in
o
r
d
er
to
e
s
ti
m
ate
s
e
m
a
n
tic
s
i
m
ilar
ities
b
et
w
ee
n
d
o
cu
m
e
n
ts
w
h
er
e
th
e
s
i
m
ilar
it
y
b
et
w
ee
n
v
ec
to
r
s
is
u
s
u
all
y
d
ef
i
n
ed
as
co
s
in
e
s
i
m
i
lar
it
y
.
I
n
p
ap
er
[
3
8
]
,
th
e
y
p
r
o
p
o
s
e
to
co
m
b
i
n
e
ex
p
lic
it
s
e
m
a
n
tic
an
a
l
y
s
i
s
(
E
S
A
)
r
ep
r
esen
tatio
n
s
a
n
d
w
o
r
d
2
v
ec
r
ep
r
esen
tatio
n
s
a
s
a
w
a
y
to
g
en
er
ate
d
en
s
er
r
ep
r
esen
tatio
n
s
an
d
,
co
n
s
eq
u
e
n
tl
y
,
a
b
etter
s
im
i
la
r
it
y
m
ea
s
u
r
e
b
et
w
ee
n
s
h
o
r
t
tex
ts
.
I
n
[
3
9
]
th
e
y
p
r
o
p
o
s
ed
a
s
em
a
n
tic
s
i
m
ilar
it
y
ap
p
r
o
ac
h
f
o
r
p
ar
ap
h
r
ase
id
en
ti
f
icatio
n
in
A
r
ab
ic
tex
ts
b
y
co
m
b
i
n
i
n
g
d
i
f
f
er
e
n
t
tech
n
iq
u
es
o
f
Nat
u
r
al
L
a
n
g
u
a
g
e
P
r
o
ce
s
s
i
n
g
N
L
P
s
u
ch
as
:
T
er
m
Fre
q
u
e
n
c
y
I
n
v
er
s
e
Do
c
u
m
en
t
Fre
q
u
en
c
y
T
F
-
I
DF
tec
h
n
iq
u
e.
T
h
e
g
o
al
is
t
o
r
ep
r
esen
t
a
w
o
r
d
v
ec
to
r
u
s
in
g
w
o
r
d
2
v
ec
.
An
d
also
,
to
g
en
er
ate
a
s
e
n
te
n
ce
v
ec
to
r
r
ep
r
esen
tatio
n
a
n
d
af
t
er
ap
p
ly
i
n
g
a
s
i
m
i
lar
it
y
m
ea
s
u
r
e
m
e
n
t
o
p
er
atio
n
b
ased
o
n
d
if
f
er
e
n
t
m
etr
i
cs
o
f
co
m
p
ar
is
o
n
,
s
u
c
h
as:
C
o
s
i
n
e
Si
m
i
lar
it
y
an
d
E
u
clid
ea
n
Dis
t
an
ce
.
T
h
is
ap
p
r
o
ac
h
w
as
e
v
al
u
ated
o
n
t
h
e
Op
e
n
So
u
r
ce
A
r
ab
ic
C
o
r
p
u
s
OS
AC
an
d
o
b
tain
ed
a
p
r
o
m
is
i
n
g
r
ate.
[
4
0
]
T
h
is
p
ap
er
p
r
o
p
o
s
es
a
n
o
v
el
d
ee
p
n
eu
r
al
n
et
w
o
r
k
-
b
ase
d
ap
p
r
o
ac
h
th
at
r
elie
s
o
n
co
ar
s
e
-
g
r
ain
ed
s
en
te
n
ce
m
o
d
ellin
g
u
s
i
n
g
a
c
o
n
v
o
l
u
tio
n
al
n
e
u
r
al
n
e
t
w
o
r
k
an
d
a
lo
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
m
o
d
el,
co
m
b
i
n
ed
w
it
h
a
s
p
ec
i
f
ic
f
i
n
e
-
g
r
ai
n
ed
w
o
r
d
-
le
v
el
s
i
m
i
lar
it
y
m
atc
h
i
n
g
m
o
d
el.
I
n
t
h
is
co
m
p
o
n
en
t,
th
e
y
r
ep
r
esen
t
ev
er
y
s
en
te
n
ce
u
s
in
g
th
eir
j
o
in
t
C
N
N
an
d
L
ST
M
ar
ch
itect
u
r
e.
T
h
e
C
NN
is
ab
le
to
lear
n
th
e
lo
ca
l
f
ea
t
u
r
es
f
r
o
m
w
o
r
d
s
to
p
h
r
ases
f
r
o
m
th
e
t
ex
t,
w
h
ile
t
h
e
L
ST
M
lear
n
s
th
e
lo
n
g
-
ter
m
d
ep
en
d
en
c
ies
o
f
th
e
tex
t.
Mo
r
e
s
p
ec
if
icall
y
,
t
h
e
y
f
ir
s
tl
y
tak
e
th
e
w
o
r
d
e
m
b
ed
d
in
g
a
s
i
n
p
u
t
to
th
eir
C
NN
m
o
d
el,
in
w
h
i
ch
v
ar
io
u
s
t
y
p
e
s
o
f
co
n
v
o
lu
tio
n
s
an
d
p
o
o
lin
g
tech
n
iq
u
e
s
ar
e
ap
p
lied
to
ca
p
tu
r
e
th
e
m
ax
i
m
u
m
in
f
o
r
m
a
tio
n
f
r
o
m
t
h
e
te
x
t.
Nex
t,
th
e
en
co
d
ed
f
ea
t
u
r
es
ar
e
u
s
ed
as
i
n
p
u
t
to
t
h
e
L
ST
M
n
et
w
o
r
k
.
Fi
n
all
y
,
t
h
e
lo
n
g
-
ter
m
d
ep
en
d
e
n
cies
lear
n
ed
b
y
th
e
L
ST
M
b
ec
o
m
e
s
th
e
s
e
m
a
n
tic
s
en
te
n
ce
r
ep
r
esen
tatio
n
.
[
4
1
]
T
h
is
ap
p
r
o
ac
h
p
r
o
p
o
s
es
to
ex
p
lici
tl
y
m
o
d
el
p
air
w
i
s
e
w
o
r
d
i
n
ter
ac
tio
n
s
a
n
d
p
r
ese
n
t
a
n
o
v
el
s
i
m
ilar
it
y
f
o
cu
s
m
ec
h
an
i
s
m
t
o
id
en
tify
i
m
p
o
r
tan
t
co
r
r
esp
o
n
d
en
ce
s
f
o
r
b
etter
s
im
ilar
it
y
m
ea
s
u
r
e
m
e
n
t.
T
h
e
y
u
s
ed
Glo
Ve
w
o
r
d
e
m
b
ed
d
in
g
s
f
o
r
v
ec
to
r
r
ep
r
esen
tatio
n
o
f
w
o
r
d
an
d
t
h
eir
m
o
d
el
c
o
n
tain
s
f
o
u
r
m
aj
o
r
co
m
p
o
n
e
n
t
s
:
1
.
B
id
ir
ec
tio
n
al
L
o
n
g
S
h
o
r
t
-
T
er
m
Me
m
o
r
y
Net
-
w
o
r
k
s
(
B
i
-
L
ST
Ms)
ar
e
u
s
ed
f
o
r
co
n
tex
t
m
o
d
eli
n
g
o
f
i
n
p
u
t
s
en
ten
ce
s
.
2
.
A
n
o
v
e
l
p
air
w
is
e
w
o
r
d
i
n
ter
ac
tio
n
m
o
d
eli
n
g
tech
n
iq
u
e
en
co
u
r
a
g
es
d
ir
ec
t
co
m
p
ar
is
o
n
s
b
et
w
ee
n
w
o
r
d
co
n
tex
t
s
ac
r
o
s
s
s
e
n
te
n
ce
s
.
C
o
s
in
e
d
is
ta
n
ce
(
co
s
)
m
ea
s
u
r
es
t
h
e
d
is
ta
n
ce
o
f
t
w
o
v
ec
to
r
s
b
y
t
h
e
an
g
le
b
et
w
e
en
th
e
m
,
w
h
ile
L
2
E
u
clid
ea
n
d
is
ta
n
ce
(
L
2
E
u
c
lid
)
an
d
d
o
tp
r
o
d
u
ct
d
is
tan
ce
(
Do
tP
r
o
d
u
ct)
m
ea
s
u
r
e
m
a
g
n
it
u
d
e
d
if
f
er
en
ce
s
.
W
e
u
s
e
t
h
r
ee
s
i
m
ilar
it
y
f
u
n
c
tio
n
s
f
o
r
r
ich
er
m
ea
s
u
r
e
m
e
n
t.
3
.
A
n
o
v
el
s
i
m
ilar
it
y
f
o
cu
s
la
y
er
h
e
lp
s
th
e
m
o
d
el
id
en
t
if
y
i
m
p
o
r
tan
t
p
air
w
is
e
w
o
r
d
i
n
ter
ac
tio
n
s
ac
r
o
s
s
s
en
ten
ce
s
.
4
.
A
la
y
er
d
ee
p
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
et
w
o
r
k
(
C
o
n
v
Net)
co
n
v
er
ts
t
h
e
s
i
m
ilar
i
t
y
m
ea
s
u
r
e
m
en
t
p
r
o
b
le
m
in
to
a
p
atter
n
r
ec
o
g
n
itio
n
p
r
o
b
lem
f
o
r
f
in
al
cla
s
s
i
f
icatio
n
.
T
h
e
m
o
d
el
o
f
[
4
2
]
is
a
p
p
lied
t
o
ass
ess
s
e
m
an
tic
s
i
m
ilar
it
y
b
et
w
ee
n
s
e
n
ten
ce
s
.
Fo
r
th
ese
a
p
p
licatio
n
s
,
th
e
y
p
r
o
v
id
e
w
o
r
d
-
e
m
b
ed
d
in
g
v
ec
to
r
s
w
o
r
d
2
v
ec
to
t
h
e
L
ST
Ms,
w
h
ich
u
s
e
a
f
i
x
ed
s
iz
e
v
e
cto
r
to
en
co
d
e
th
e
u
n
d
er
l
y
in
g
m
ea
n
i
n
g
ex
p
r
es
s
e
d
in
a
s
en
ten
ce
(
ir
r
esp
ec
tiv
e
o
f
t
h
e
p
ar
ticu
lar
w
o
r
d
in
g
/
s
y
n
tax
)
.
B
y
r
es
tr
ictin
g
s
u
b
s
eq
u
en
t
o
p
er
atio
n
s
to
r
el
y
o
n
a
s
i
m
p
le
Ma
n
h
a
ttan
m
etr
ic,
th
e
y
co
m
p
el
t
h
e
s
en
te
n
ce
r
ep
r
esen
tat
io
n
s
lear
n
ed
b
y
t
h
eir
m
o
d
el
to
f
o
r
m
a
h
ig
h
l
y
s
tr
u
ct
u
r
ed
s
p
a
ce
w
h
o
s
e
g
eo
m
etr
y
r
ef
lec
ts
co
m
p
lex
s
e
m
a
n
tic
r
elatio
n
s
h
ip
s
.
[
4
3
]
T
h
is
p
ap
er
p
r
o
p
o
s
es
a
m
o
d
el
f
o
r
co
m
-
p
ar
i
n
g
s
e
n
te
n
ce
s
t
h
at
u
s
e
s
a
m
u
ltip
lic
it
y
o
f
p
er
s
p
ec
tiv
es.
W
e
f
ir
s
t
m
o
d
el
ea
ch
s
en
ten
ce
u
s
in
g
a
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
et
w
o
r
k
th
at
ex
tr
ac
ts
f
ea
tu
r
es
at
m
u
ltip
le
le
v
el
s
o
f
g
r
an
u
lar
i
t
y
a
n
d
u
s
e
s
m
u
ltip
le
t
y
p
es
o
f
p
o
o
lin
g
.
W
e
th
e
n
co
m
p
ar
e
o
u
r
s
en
ten
ce
r
ep
r
esen
tatio
n
s
a
t
s
e
v
er
al
g
r
a
n
u
lar
i
ties
u
s
i
n
g
m
u
lt
ip
le
s
i
m
il
ar
it
y
m
etr
ics
(
co
s
,
L
E
u
clid
)
.
W
e
ap
p
ly
o
u
r
m
o
d
el
to
th
r
ee
task
s
,
in
c
lu
d
i
n
g
t
h
e
Mic
r
o
s
o
f
t
R
esear
c
h
p
ar
ap
h
r
ase
id
en
tif
icat
io
n
tas
k
an
d
t
w
o
Se
m
E
v
al
s
e
m
an
tic
tex
t
u
al
s
i
m
ilar
it
y
tas
k
s
.
I
n
th
is
p
ap
er
[
4
4
]
,
th
e
y
p
r
esen
t c
o
n
v
o
l
u
tio
n
a
l n
e
u
r
al
n
et
w
o
r
k
ar
ch
i
tectu
r
e
f
o
r
r
er
an
k
i
n
g
p
a
ir
s
o
f
s
h
o
r
t
tex
ts
,
w
h
er
e
t
h
e
y
lear
n
t
h
e
o
p
ti
m
al
r
ep
r
esen
tat
io
n
o
f
te
x
t
p
air
s
an
d
a
s
i
m
ilar
it
y
f
u
n
ctio
n
to
r
elate
t
h
e
m
i
n
a
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
1
,
Ma
r
ch
20
20
:
81
–
90
86
s
u
p
er
v
i
s
ed
w
a
y
f
r
o
m
t
h
e
a
v
ail
ab
le
tr
ain
i
n
g
d
ata.
T
h
eir
n
et
wo
r
k
tak
e
s
o
n
l
y
w
o
r
d
s
i
n
t
h
e
i
n
p
u
t,
th
u
s
r
eq
u
ir
in
g
m
i
n
i
m
al
p
r
ep
r
o
ce
s
s
in
g
.
I
n
p
ar
ticu
lar
,
th
e
y
co
n
s
id
er
th
e
tas
k
o
f
r
er
an
k
i
n
g
s
h
o
r
t
te
x
t
p
air
s
w
h
er
e
ele
m
e
n
t
s
o
f
th
e
p
air
ar
e
s
en
te
n
ce
s
.
T
h
e
y
test
o
u
r
d
ee
p
lear
n
in
g
s
y
s
te
m
o
n
t
w
o
p
o
p
u
lar
r
etr
iev
al
t
ask
s
f
r
o
m
T
R
E
C
:
Qu
est
io
n
A
n
s
w
er
i
n
g
a
n
d
Mic
r
o
b
lo
g
R
etr
iev
al.
[
4
5
]
T
h
is
s
y
s
te
m
co
m
b
i
n
es
co
n
v
o
l
u
tio
n
a
n
d
r
ec
u
r
r
en
t
n
e
u
r
al
n
et
w
o
r
k
s
to
m
ea
s
u
r
e
t
h
e
s
e
m
an
tic
s
i
m
ilar
it
y
o
f
s
e
n
ten
ce
s
.
I
t
u
s
e
s
a
co
n
v
o
l
u
tio
n
n
et
w
o
r
k
to
tak
e
ac
co
u
n
t
o
f
th
e
lo
ca
l
co
n
tex
t
o
f
w
o
r
d
s
an
d
an
L
ST
M
to
co
n
s
id
er
th
e
g
lo
b
al
co
n
tex
t
o
f
s
e
n
te
n
ce
s
.
T
h
is
co
m
b
in
a
tio
n
o
f
n
et
w
o
r
k
s
h
elp
s
to
p
r
eser
v
e
th
e
r
elev
a
n
t
in
f
o
r
m
a
tio
n
o
f
s
en
te
n
ce
s
a
n
d
i
m
p
r
o
v
es
t
h
e
ca
lcu
latio
n
o
f
t
h
e
s
i
m
ilar
it
y
b
et
w
ee
n
s
e
n
te
n
ce
s
.
A
cc
o
r
d
in
g
to
t
h
i
s
s
tate
o
f
t
h
e
ar
t
w
e
h
a
v
e
b
ee
n
ab
le
to
d
ete
ct
th
e
s
tr
e
n
g
t
h
s
a
n
d
w
ea
k
n
ess
e
s
o
f
ea
ch
ap
p
r
o
ac
h
th
at
h
elp
ed
u
s
to
b
u
ild
o
u
r
ap
p
r
o
ac
h
.
T
h
e
T
ab
le
1
r
e
p
r
esen
ts
a
s
u
m
m
ar
y
co
m
p
ar
ed
to
th
e
m
eth
o
d
s
ab
o
v
e:
T
ab
le
1
.
C
o
m
p
ar
ativ
e
tab
le
A
p
p
r
o
a
c
h
V
e
c
t
o
r
r
e
p
r
e
s
e
n
t
a
t
i
o
n
L
e
v
e
l
t
r
e
a
t
me
n
t
S
i
mi
l
a
r
i
t
y
me
t
h
o
d
D
a
t
a
se
t
/
r
e
so
u
r
c
e
s
C
r
i
t
i
c
a
l
W
o
r
d
S
e
n
t
e
n
c
e
[
2
6
]
W
o
r
d
2
v
e
c
A
v
e
r
a
g
e
se
n
t
e
n
c
e
C
o
si
n
e
,
Jac
c
a
r
d
P
A
N
2
0
1
6
L
o
ss o
f
t
h
e
me
a
n
i
n
g
o
f
t
h
e
se
n
t
e
n
c
e
.
[
2
7
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
o
si
n
e
O
S
A
C
A
r
a
b
i
c
c
o
r
p
u
s
Tw
o
d
o
c
u
me
n
t
s
sh
a
r
e
t
h
e
s
a
me
v
e
c
t
o
r
s c
o
u
l
d
b
e
n
o
n
-
p
l
a
g
i
a
r
i
z
e
d
.
T
h
e
u
se
o
f
c
o
si
n
e
t
o
d
e
t
e
c
t
si
mi
l
a
r
i
t
y
b
e
t
w
e
e
n
se
n
t
e
n
c
e
s
r
e
mai
n
s
a
so
l
u
t
i
o
n
t
h
a
t
c
a
r
r
i
e
s
man
y
r
i
sk
s
.
U
se
c
o
si
n
e
si
m
i
l
a
r
i
t
y
t
o
c
o
mp
u
t
e
a
si
m
i
l
a
r
i
t
y
b
e
t
w
e
e
n
se
n
t
e
n
c
e
s.
[
3
2
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
o
si
n
e
M
i
c
r
o
so
f
t
R
e
se
a
r
c
h
P
a
r
a
p
h
r
a
se
C
o
r
p
u
s
[
3
3
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
o
si
n
e
M
i
c
r
o
so
f
t
R
e
se
a
r
c
h
P
a
r
a
p
h
r
a
se
C
o
r
p
u
s
d
a
t
a
se
t
[
3
4
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
o
si
n
e
M
i
c
r
o
so
f
t
R
e
se
a
r
c
h
V
i
d
e
o
D
e
scri
p
t
i
o
n
C
o
r
p
u
s
[
3
5
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
o
si
n
e
W
i
k
i
p
e
d
i
a
d
a
t
a
se
t
[
3
6
]
W
o
r
d
2
v
e
c
tf
-
i
d
f
-
w
o
r
d
C
o
si
n
e
W
i
k
i
p
e
d
i
a
d
a
t
a
se
t
[
3
7
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
o
si
n
e
W
i
k
i
c
o
r
p
u
s
[
3
8
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
o
si
n
e
-
[
3
9
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
o
si
n
e
Eu
c
l
i
d
e
a
n
D
i
st
a
n
c
e
A
r
a
b
i
c
C
o
r
p
u
s
O
S
A
C
[
2
4
]
W
o
r
d
2
v
e
c
-
w
o
r
d
L
C
S
P
A
N
2
0
1
3
L
C
S
p
r
o
b
l
e
m se
e
k
s a
l
o
n
g
e
st
su
b
se
q
u
e
n
c
e
o
f
e
v
e
r
y
me
mb
e
r
o
f
a
g
i
v
e
n
se
t
o
f
v
e
c
t
o
r
s,
l
o
se
t
h
e
se
ma
n
t
i
c
a
sp
e
c
t
.
[
2
8
]
D
e
e
p
S
t
r
u
c
t
u
r
e
d
S
e
man
t
i
c
M
o
d
e
l
(
D
S
S
M
)
-
w
o
r
d
C
o
si
n
e
S
e
mEv
a
l
2
0
1
5
En
g
l
i
sh
S
T
S
T
h
e
t
r
e
a
t
me
n
t
i
s
a
t
t
h
e
l
e
v
e
l
o
f
se
n
t
e
n
c
e
o
r
smal
l
t
e
x
t
s.
[
2
9
]
-
D
o
c
2
v
e
c
se
n
t
e
n
c
e
C
o
si
n
e
-
S
l
o
w
n
e
ss o
f
t
h
e
sy
st
e
m.
T
h
e
se
man
t
i
c
a
sp
e
c
t
o
f
a
p
a
r
a
g
r
a
p
h
i
s
l
o
st
b
e
c
a
u
se
t
h
e
c
o
mp
a
r
i
so
n
i
s
d
o
n
e
se
n
t
e
n
c
e
b
y
se
n
t
e
n
c
e
.
[
3
0
]
-
D
o
c
2
v
e
c
se
n
t
e
n
c
e
C
o
si
n
e
-
S
t
a
n
f
o
r
d
se
n
t
i
me
n
t
t
r
e
e
b
a
n
k
d
a
t
a
se
t
-
I
M
D
B
d
a
t
a
se
t
[
3
1
]
G
l
o
V
e
R
e
c
u
r
si
v
e
n
e
u
r
a
l
n
e
t
w
o
r
k
s
se
n
t
e
n
c
e
C
o
si
n
e
S
e
mEv
a
l
-
2
0
1
5
T
a
sk
2
U
se
o
f
d
o
c
2
v
e
c
i
s
b
e
t
t
e
r
t
h
e
n
u
se
s R
N
N
.
T
h
e
se
man
t
i
c
a
sp
e
c
t
o
f
a
p
a
r
a
g
r
a
p
h
i
s
l
o
st
.
[
2
2
]
W
o
r
d
2
v
e
c
I
n
f
e
r
S
e
n
t
se
n
t
e
n
c
e
C
o
si
n
e
-
T
h
e
u
se
o
f
c
o
si
n
e
t
o
d
e
t
e
c
t
si
mi
l
a
r
i
t
y
b
e
t
w
e
e
n
se
n
t
e
n
c
e
s
r
e
mai
n
s
a
so
l
u
t
i
o
n
t
h
a
t
c
a
r
r
i
e
s
man
y
r
i
sk
s
.
T
h
e
c
o
mp
a
r
i
so
n
i
s
d
o
n
e
a
t
t
h
e
se
n
t
e
n
c
e
l
e
v
e
l
,
so
w
e
a
l
w
a
y
s e
n
c
o
u
n
t
e
r
t
h
e
p
r
o
b
l
e
m o
f
l
o
ss o
f
t
h
e
se
man
t
i
c
a
sp
e
c
t
o
f
t
h
e
p
a
r
a
g
r
a
p
h
o
r
t
e
x
t
a
n
a
l
y
se
d
.
A
c
c
o
r
d
i
n
g
t
o
t
h
e
st
u
d
y
d
o
n
e
b
y
[
3
1
]
,
[
3
2
]
h
e
f
o
u
n
d
t
h
a
t
t
h
e
u
se
o
f
d
o
c
2
v
e
c
g
i
v
e
s t
r
a
mp
l
i
n
g
r
e
su
l
t
s.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
A
d
ee
p
lea
r
n
in
g
b
a
s
ed
tec
h
n
iq
u
e
fo
r
p
la
g
ia
r
is
m
d
etec
tio
n
:
a
co
mp
a
r
a
tive
s
tu
d
y
(
E
l Mo
s
ta
fa
Ha
mb
i
)
87
T
ab
le
1
.
C
o
m
p
ar
ativ
e
tab
le
(
C
o
n
tin
u
e)
A
p
p
r
o
a
c
h
V
e
c
t
o
r
r
e
p
r
e
s
e
n
t
a
t
i
o
n
L
e
v
e
l
t
r
e
a
t
me
n
t
S
i
mi
l
a
r
i
t
y
me
t
h
o
d
D
a
t
a
se
t
/
r
e
so
u
r
c
e
s
C
r
i
t
i
c
a
l
W
o
r
d
S
e
n
t
e
n
c
e
[
2
5
]
W
o
r
d
2
v
e
c
-
w
o
r
d
W
M
D
B
B
C
S
P
O
R
-
T
T
h
i
s me
t
h
o
d
i
s
u
se
d
j
u
s
t
t
o
d
e
t
e
c
t
t
h
e
s
i
mi
l
a
r
i
t
y
b
e
t
w
e
e
n
smal
l
se
n
t
e
n
c
e
s.
[
4
0
]
G
l
o
v
e
-
w
o
r
d
C
N
N
-
R
N
N
S
e
mEv
a
l
2
0
1
5
T
h
e
se
me
t
h
o
d
s a
r
e
b
a
se
d
o
n
t
h
e
v
e
c
t
o
r
r
e
p
r
e
se
n
t
a
t
i
o
n
o
f
w
o
r
d
s,
so
t
h
e
y
a
r
e
u
se
d
o
n
l
y
f
o
r
t
h
e
d
e
t
e
c
t
i
o
n
o
f
s
i
mi
l
a
r
i
t
y
b
e
t
w
e
e
n
se
n
t
e
n
c
e
s
b
u
t
n
o
t
t
e
x
t
s.
[
4
5
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
N
N
-
L
S
T
M
S
I
C
K
[
4
1
]
G
l
o
v
e
-
w
o
r
d
L
st
m
C
n
n
D
o
t
P
r
o
d
u
c
t
L
2
Eu
c
l
i
d
-
2
0
1
4
S
e
mEv
a
l
-
M
i
c
r
o
so
f
t
V
i
d
e
o
P
a
r
a
p
h
r
a
se
C
o
r
p
u
s
-
W
i
k
i
Q
A
[
4
2
]
W
o
r
d
2
v
e
c
-
w
o
r
d
L
S
T
M
S
I
C
K
[
4
3
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
N
N
-
S
e
mEv
a
l
-
M
i
c
r
o
so
f
t
R
e
se
a
r
c
h
p
a
r
a
p
h
r
a
se
A
l
w
a
y
s
w
e
e
n
c
o
u
n
t
e
r
t
h
e
p
r
o
b
l
e
m o
f
l
e
v
e
l
r
e
p
r
e
s
e
n
t
a
t
i
o
n
o
f
t
h
e
a
n
a
l
y
se
d
d
a
t
a
;
t
h
e
r
e
p
r
e
se
n
t
a
t
i
o
n
b
y
w
o
r
d
p
o
se
s
t
h
e
p
r
o
b
l
e
m t
h
a
t
w
e
c
a
n
j
u
st
a
n
a
l
y
se
t
h
e
smal
l
se
n
t
e
n
c
e
s.
C
N
N
's u
se
o
f
t
r
e
a
t
i
n
g
t
h
e
si
mi
l
a
r
i
t
y
b
e
t
w
e
e
n
l
i
st
o
f
w
o
r
d
p
o
se
s se
v
e
r
a
l
p
r
o
b
l
e
ms l
i
k
e
t
h
e
l
o
ss o
f
se
man
t
i
c
s
l
e
v
e
l
o
f
t
h
e
se
n
t
e
n
c
e
c
o
n
st
r
u
c
t
.
[
4
4
]
W
o
r
d
2
v
e
c
-
w
o
r
d
C
N
N
T
R
EC
:
A
n
sw
e
r
i
n
g
a
n
d
M
i
c
r
o
b
l
o
g
R
e
t
r
i
e
v
a
l
I
n
ad
d
itio
n
to
th
at
w
e
co
u
ld
d
etec
t
th
e
m
o
s
t
p
o
w
er
f
u
l
m
et
h
o
d
s
u
s
ed
f
o
r
t
h
e
r
ep
r
esen
tatio
n
o
f
a
te
x
t.
I
t
h
as
b
ee
n
f
o
u
n
d
t
h
at
t
h
e
u
s
e
o
f
th
e
d
o
c2
v
ec
p
r
in
cip
le
r
e
m
ain
s
t
h
e
m
o
s
t
r
ele
v
an
t
s
o
l
u
tio
n
f
r
o
m
t
h
e
[
2
9
-
3
0
]
s
tu
d
y
,
an
d
th
e
n
w
e
w
e
n
t
f
u
r
th
er
a
n
d
to
o
k
i
n
s
p
ir
atio
n
f
r
o
m
i
t
to
b
u
ild
o
u
r
lear
n
i
n
g
s
y
s
te
m
th
at
d
etec
t
s
p
lag
iar
is
m
b
et
w
ee
n
t
h
e
d
o
cu
m
en
ts
.
4.
RE
SU
L
T
S
A
ND
D
I
SCU
SS
I
O
N
I
n
t
h
is
p
ar
t
w
e
w
ill
an
al
y
s
e
t
h
e
r
es
u
lt
s
f
o
u
n
d
i
n
th
e
s
t
u
d
y
ca
r
r
ied
o
u
t
ab
o
v
e,
f
ir
s
t
w
e
w
i
ll
ill
u
s
tr
ate
th
e
m
o
s
t i
m
p
o
r
tan
t c
o
m
p
ar
is
o
n
cr
iter
ia
d
ef
i
n
ed
:
Vec
t
o
r
re
presenta
t
io
n:
T
h
is
is
a
tr
ea
t
m
e
n
t
p
er
f
o
r
m
ed
o
n
a
tex
t
th
at
w
ill
tr
an
s
f
o
r
m
it
to
li
s
t
o
f
v
ec
to
r
s
w
h
ic
h
k
ee
p
th
e
s
e
m
a
n
tic
a
n
d
s
y
n
tacti
c
asp
ec
t o
f
f
er
ed
b
y
th
e
u
s
e
o
f
d
ee
p
lear
n
in
g
al
g
o
r
ith
m
s
.
L
ev
el
t
re
a
t
m
ent
:
th
i
s
cr
iter
io
n
d
ef
i
n
es
t
h
e
lev
el
o
f
th
e
tr
ea
t
m
en
t
o
f
a
tex
t,
m
o
r
e
ex
ac
tl
y
i
f
th
e
te
x
t
is
tr
ea
ted
b
y
w
o
r
d
o
r
b
y
s
e
n
te
n
ce
.
Si
m
ila
rit
y
m
et
ho
d:
T
h
is
p
ar
t
d
ea
ls
w
it
h
t
h
e
ap
p
r
o
ac
h
es
u
s
ed
f
o
r
ca
lcu
lati
n
g
t
h
e
s
i
m
i
lar
it
y
b
et
w
ee
n
t
h
e
v
ec
to
r
s
t
h
at
r
ep
r
esen
t
t
h
e
te
x
t
s
,
w
h
ic
h
w
i
ll
g
iv
e
u
s
a
g
lo
b
al
v
is
ib
il
it
y
to
d
etec
t
th
e
s
tr
en
g
t
h
s
a
n
d
w
ea
k
n
es
s
es
o
f
ea
ch
m
eth
o
d
.
I
n
ad
d
itio
n
,
w
e
ar
e
g
o
i
n
g
to
tal
k
ab
o
u
t
t
h
e
cr
itical
p
o
in
t
f
o
r
ea
ch
ap
p
r
o
ac
h
il
lu
s
tr
ated
i
n
t
h
e
p
ar
a
g
r
ap
h
ab
o
v
e.
Star
ti
n
g
f
r
o
m
t
h
e
m
e
th
o
d
s
u
s
ed
f
o
r
t
h
e
v
ec
to
r
r
ep
r
esen
tatio
n
o
f
a
tex
t,
ac
co
r
d
in
g
to
t
h
e
an
al
y
s
is
it
t
u
r
n
s
o
u
t
th
at
m
o
s
t
o
f
th
e
ap
p
r
o
ac
h
es
u
s
e
eit
h
er
th
e
w
o
r
d
2
v
ec
o
r
th
e
d
o
c
2
v
ec
f
o
r
its
v
ec
to
r
tr
an
s
f
o
r
m
atio
n
,
s
o
w
e
d
is
tin
g
u
is
h
t
h
at
th
e
m
i
k
o
lo
v
r
ep
r
esen
tatio
n
s
ar
e
th
e
b
est
m
et
h
o
d
s
u
s
ed
to
k
ee
p
th
e
s
e
m
a
n
tic
asp
ec
t
o
f
a
g
i
v
en
tex
t.
I
n
R
ev
e
n
g
e,
E
ac
h
A
p
p
r
o
ac
h
tr
ea
ts
t
h
e
te
x
t
w
it
h
it
s
o
w
n
wa
y
,
t
h
er
e
ar
e
s
o
m
e
w
h
o
tr
an
s
f
o
r
m
it
i
n
to
a
li
s
t
o
f
w
o
r
d
s
a
n
d
s
o
m
eo
n
e
i
n
to
a
lis
t
o
f
s
e
n
te
n
ce
s
,
t
h
ese
r
ep
r
ese
n
t
at
io
n
s
y
ield
r
es
u
lt
s
th
at
d
if
f
er
f
r
o
m
o
n
e
ap
p
r
o
ac
h
to
an
o
th
er
b
u
t
th
e
tr
a
n
s
f
o
r
m
at
io
n
o
f
a
tex
t
to
a
lis
t
o
f
s
e
n
ten
ce
s
in
o
u
r
o
p
in
io
n
r
e
m
ain
s
th
e
m
o
s
t
r
ele
v
an
t
s
in
ce
th
e
m
ea
n
i
n
g
o
f
th
e
te
x
t
tr
e
ated
r
em
ai
n
s
i
n
co
n
s
id
er
atio
n
.
W
ith
r
eg
ar
d
to
th
e
m
et
h
o
d
s
u
s
ed
f
o
r
th
e
s
i
m
ilar
it
y
ca
lc
u
latio
n
,
th
e
p
r
ec
ed
in
g
p
ar
ag
r
ap
h
s
m
e
n
tio
n
th
e
d
i
f
f
er
en
t
w
a
y
s
u
s
ed
to
d
etec
t
w
h
et
h
er
th
er
e
is
a
s
i
m
i
lar
it
y
o
r
n
o
t
b
et
w
ee
n
th
e
a
n
al
y
s
ed
te
x
ts
.
T
h
er
e
ar
e
also
m
a
n
y
ap
p
r
o
ac
h
es
t
h
at
w
o
r
k
w
it
h
C
N
N
an
d
R
NN
o
n
its
p
lag
iar
is
m
d
etec
tio
n
ar
ch
it
ec
t
u
r
e,
b
u
t
m
o
s
t
o
f
t
h
e
m
u
s
e
t
h
e
w
o
r
d
lev
el
f
o
r
its
v
ec
to
r
r
ep
r
esen
tatio
n
,
s
o
t
h
e
y
ar
e
u
s
ed
o
n
l
y
f
o
r
th
e
d
etec
tio
n
o
f
s
i
m
ilar
it
y
b
et
w
ee
n
s
e
n
te
n
c
es b
u
t n
o
t te
x
ts
.
I
n
co
n
clu
s
io
n
,
w
e
f
o
u
n
d
th
at
al
m
o
s
t
o
f
th
e
s
e
ap
p
r
o
ac
h
es
u
s
e
th
e
co
s
i
n
e
to
ca
lcu
la
te
th
e
s
i
m
i
lar
it
y
b
et
w
ee
n
d
o
cu
m
e
n
ts
,
s
o
it
w
as
f
o
u
n
d
t
h
at
th
e
s
e
m
et
h
o
d
s
p
er
f
o
r
m
its
s
i
m
ilar
it
y
a
n
al
y
s
es
i
n
w
o
r
d
-
by
-
w
o
r
d
o
r
s
en
te
n
ce
-
by
-
s
e
n
te
n
ce
,
w
h
ic
h
w
il
l
p
o
s
e
a
f
ter
r
eliab
ili
t
y
p
r
o
b
lem
o
f
th
e
s
e
r
es
u
lt
s
,
s
in
ce
w
e
ca
n
f
i
n
d
t
w
o
d
o
cu
m
en
ts
t
h
at
s
h
ar
e
t
h
e
s
a
m
e
w
o
r
d
o
r
th
e
s
a
m
e
s
e
n
ten
ce
s
b
u
t
th
e
y
ar
e
n
o
t
s
e
m
a
n
tical
l
y
s
i
m
ilar
,
in
ad
d
itio
n
to
th
at
w
e
ca
n
lo
s
e
th
e
s
e
m
a
n
t
ic
asp
ec
t
w
h
e
n
t
h
e
d
o
cu
m
e
n
ts
ar
e
tr
ea
tin
g
v
ia
a
li
s
t
o
f
s
e
n
te
n
ce
s
o
r
w
o
r
d
s
.
So
,
y
o
u
h
a
v
e
to
th
i
n
k
o
f
a
m
et
h
o
d
th
at
m
a
n
ag
e
s
th
is
p
r
o
b
lem
b
y
p
r
o
p
o
s
in
g
an
ap
p
r
o
ac
h
th
at
will
r
ep
r
esen
t
a
tex
t
b
y
a
l
is
t
o
f
s
e
n
ten
ce
s
t
h
at
w
ill
ev
en
t
u
all
y
b
e
tr
an
s
f
o
r
m
ed
in
t
o
a
lis
t
o
f
v
ec
to
r
s
,
an
d
in
ad
d
it
io
n
to
th
at
w
e
m
u
s
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
1
,
Ma
r
ch
20
20
:
81
–
90
88
u
s
e
a
tr
ea
t
m
en
t
th
a
t
k
ee
p
s
t
h
e
s
e
m
an
t
ic
asp
ec
t
o
f
t
h
is
lis
t
o
f
s
e
n
te
n
ce
s
,
s
o
it
w
ell
b
e
a
m
a
n
ip
u
la
tio
n
th
at
p
r
o
ce
s
s
e
s
a
lis
t o
f
s
e
n
te
n
ce
s
to
d
etec
t a
s
i
m
i
lar
it
y
u
s
in
g
a
n
al
g
o
r
ith
m
li
k
e
t
h
e
R
N
N
th
a
t
w
il
l k
ee
p
th
e
s
e
m
a
n
tic
asp
ec
t o
f
a
tex
t.
5.
CO
NCLU
SI
O
N
I
n
th
i
s
p
ap
er
,
w
e
h
av
e
m
e
n
tio
n
ed
m
a
n
y
d
if
f
er
en
t
m
et
h
o
d
s
u
s
ed
i
n
d
etec
tio
n
o
f
p
lag
iar
i
s
m
o
f
id
ea
s
th
at
s
tan
d
f
o
r
th
e
p
r
i
n
cip
al
o
f
Dee
p
L
ea
r
n
i
n
g
,
an
d
b
y
th
is
b
r
illi
an
t
s
t
u
d
y
w
e
co
u
ld
co
n
s
tr
u
ct
o
u
r
cr
itical
b
ase
o
f
t
h
e
p
r
ev
io
u
s
w
ea
k
n
ess
e
s
wh
ich
w
e
h
a
v
e
s
ee
n
d
u
r
in
g
o
u
r
s
tu
d
y
.
T
h
is
h
elp
ed
u
s
to
g
et
a
g
en
er
al
id
ea
ab
o
u
t
th
e
d
if
f
er
e
n
t
m
eth
o
d
s
o
f
d
ee
p
lear
n
in
g
u
s
ed
f
o
r
p
lag
iar
i
s
m
d
etec
tio
n
o
r
esp
ec
iall
y
s
e
m
an
tic
p
lag
iar
i
s
m
d
etec
tio
n
.
I
n
ad
d
itio
n
to
t
h
i
s
,
t
h
is
s
t
u
d
y
h
as
g
i
v
e
n
u
s
t
h
e
p
at
h
s
to
f
o
llo
w
f
o
r
th
e
co
n
s
tr
u
cti
o
n
o
f
o
u
r
ap
p
r
o
ac
h
b
y
b
en
e
f
iti
n
g
f
r
o
m
t
h
e
s
tr
e
n
g
t
h
s
o
f
ea
c
h
m
eth
o
d
a
n
d
b
y
p
ass
in
g
t
h
e
w
ea
k
p
o
in
ts
o
f
ea
c
h
m
eth
o
d
.
C
o
n
ce
r
n
in
g
th
e
f
u
t
u
r
e
w
o
r
k
co
n
s
is
t
s
o
f
co
n
s
tr
u
ct
an
d
p
u
t
tin
g
i
n
to
p
r
ac
tice
o
u
r
ap
p
r
o
ac
h
an
d
co
m
p
ar
in
g
i
t
w
it
h
th
e
o
t
h
er
m
et
h
o
d
s
u
s
ed
at
t
h
e
lev
el
o
f
th
e
p
h
ase
r
elate
d
w
o
r
k
.
RE
F
E
R
E
NC
E
S
[1
]
T
u
o
m
o
Ka
k
k
o
n
e
n
,
M
a
x
i
m
M
o
z
g
o
v
o
y
.
He
r
m
e
ti
c
a
n
d
W
e
b
P
lag
iaris
m
De
te
c
ti
o
n
S
y
ste
m
s
f
o
r
S
tu
d
e
n
t
Essa
y
sa
n
Ev
a
lu
a
ti
o
n
Of
T
h
e
S
tate
-
Of
-
T
h
e
-
A
rt.
J
o
u
rn
a
l
o
f
Ed
u
c
a
ti
o
n
a
l
Co
mp
u
t
in
g
Res
e
a
rc
h
,
v
4
2
n
2
p
1
3
5
-
1
5
9
2
0
1
0
.
Un
iv
e
rsit
y
o
f
Jo
e
n
su
u
,
F
in
lan
d
,
U
n
iv
e
rsity
o
f
A
izu
,
Ja
p
a
n
[
e
n
li
g
n
e
]
2
0
1
0
.
[2
]
A
h
m
e
d
Ja
b
r
A
h
m
e
d
M
u
f
tah
.
D
o
c
u
m
e
n
t
P
lag
iarism
De
tec
ti
o
n
A
l
g
o
rit
h
m
Us
in
g
S
e
m
a
n
ti
c
Ne
t
w
o
rk
s.
A
p
ro
jec
t
re
p
o
rt
su
b
m
it
ted
in
p
a
rt
ial
f
u
lf
il
lm
e
n
t
o
f
th
e
re
q
u
irem
e
n
ts
f
o
r
th
e
a
w
a
rd
o
f
th
e
d
e
g
re
e
o
f
M
a
s
ter
o
f
S
c
ien
c
e
(Co
m
p
u
ter
S
c
ien
c
e
).
F
a
c
u
lt
y
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
In
f
o
rm
a
ti
o
n
S
y
ste
m
s
Un
iv
e
rsit
y
Tec
h
n
o
lo
g
y
M
a
la
y
sia
(2
0
0
9
).
[3
]
Erf
a
n
e
h
G
h
a
ra
v
i,
Ka
y
v
a
n
Bij
a
ri
e
t
Kia
ra
sh
Zah
irn
iaA
De
e
p
L
e
a
r
n
in
g
A
p
p
ro
a
c
h
to
P
e
rsia
n
P
lag
iarism
De
tec
ti
o
n
.
J
o
u
rn
a
l
o
f
M
a
c
h
i
n
e
L
e
a
rn
in
g
Res
e
a
rc
h
(2
0
1
1
).
[4
]
S
.
M
.
A
lza
h
ra
n
i,
N.
S
a
li
m
,
a
n
d
A
.
A
b
ra
h
a
m
,
“
Un
d
e
rsta
n
d
in
g
p
la
g
iaris
m
li
n
g
u
isti
c
p
a
tt
e
rn
s,
tex
tu
a
l
fe
a
tu
re
s,
a
n
d
d
e
tec
ti
o
n
m
e
th
o
d
s,”
T
ra
n
s.
S
y
s.M
a
n
Cy
b
e
r
P
a
rt
C,
v
o
l
.
4
2
,
n
o
.
2
,
p
p
.
1
3
3
–
1
4
9
,
M
a
r.
2
0
1
2
.
[
On
l
in
e
].
A
v
a
il
a
b
le:
h
tt
p
:
//
d
x
.
d
o
i.
o
rg
/1
0
.
1
1
0
9
/T
S
M
C
C.
2
0
1
1
.
2
1
3
4
8
4
7
.
[5
]
M
.
Ch
o
n
g
a
n
d
L
.
S
p
e
c
ia,
“
Lex
i
c
a
l
g
e
n
e
ra
li
sa
ti
o
n
f
o
r
w
o
rd
-
lev
e
l
m
a
tch
in
g
in
p
lag
iaris
m
d
e
tec
ti
o
n
,
”
in
RA
N
L
P
,
2
0
1
1
,
p
p
.
7
0
4
–
7
0
9
.
[6
]
S
.
Brin
,
J.
Da
v
is,
a
n
d
H.
G
a
rc
ia
-
M
o
li
n
a
,
“
Co
p
y
d
e
tec
ti
o
n
m
e
c
h
a
n
ism
s
f
o
r
d
ig
it
a
l
d
o
c
u
m
e
n
ts,”
in
S
IG
M
OD
Co
n
f
e
re
n
c
e
,
1
9
9
5
,
p
p
.
3
9
8
–
4
0
9
.
[7
]
D.
R.
W
h
it
e
a
n
d
M
.
Jo
y
,
“
S
e
n
ten
c
e
-
b
a
se
d
n
a
tu
ra
l
la
n
g
u
a
g
e
p
lag
iaris
m
d
e
tec
ti
o
n
,
”
ACM
J
o
u
r
n
a
l
o
f
Ed
u
c
a
t
io
n
a
l
Res
o
u
rc
e
s in
Co
m
p
u
t
in
g
,
v
o
l.
4
,
n
o
.
4
,
p
p
.
1
–
2
0
,
2
0
0
4
.
[8
]
S
.
Nie
z
g
o
d
a
a
n
d
T
.
P
.
W
a
y
,
“
S
n
it
c
h
:
a
so
f
t
w
a
r
e
to
o
l
f
o
r
d
e
tec
ti
n
g
c
u
t
a
n
d
p
a
ste
p
lag
iaris
m
,
”
in
S
I
GCSE
,
2
0
0
6
,
p
p
.
51
–
5
5
.
[9
]
A
.
Ba
rr
´
o
n
-
Ce
d
e
n
o
a
n
d
P
.
Ro
ss
o
,
“
On
a
u
to
m
a
ti
c
p
lag
iarism
d
e
t
e
c
ti
o
n
b
a
se
d
o
n
n
-
g
ra
m
s
c
o
m
p
a
riso
n
,
”
i
n
ECIR
,
2
0
0
9
,
p
p
.
6
9
6
–
7
0
0
.
[1
0
]
M
.
S
.
P
e
ra
a
n
d
Y.
-
K.
Ng
,
“
A
n
a
ıv
e
b
a
y
e
s
c
la
ss
ifi
e
r
f
o
r
w
e
b
d
o
c
u
m
e
n
t
su
m
m
a
ries
c
re
a
ted
b
y
u
sin
g
w
o
rd
si
m
il
a
rit
y
a
n
d
sig
n
ifi
c
a
n
t
f
a
c
to
rs,”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
n
Arti
fi
c
ia
l
In
tell
ig
e
n
c
e
T
o
o
ls,
v
o
l.
1
9
,
n
o
.
4
,
p
p
.
4
6
5
–
4
8
6
,
2
0
1
0
.
[1
1
]
E.
S
tam
a
tato
s,
“
P
lag
iarism
d
e
tec
ti
o
n
u
sin
g
sto
p
w
o
rd
n
-
g
ra
m
s,”
J
A
S
IS
T
,
v
o
l.
6
2
,
n
o
.
1
2
,
p
p
.
2
5
1
2
–
2
5
2
7
,
2
0
1
1
.
[1
2
]
J.
G
r
m
a
n
a
n
d
R.
R
a
v
a
s,
“
I
m
p
ro
v
e
d
im
p
le
m
e
n
tatio
n
f
o
r
fi
n
d
in
g
te
x
t
si
m
il
a
rit
ies
in
larg
e
se
ts
o
f
d
a
ta
-
n
o
teb
o
o
k
f
o
r
p
a
n
a
t
c
lef
2
0
1
1
,
”
i
n
CL
EF
(No
te
b
o
o
k
P
a
p
e
rs/L
a
b
s/W
o
rk
sh
o
p
),
2
0
1
1
.
[1
3
]
Uz
u
n
e
r,
O.,
a
n
d
Ka
tz,
B.
,
a
n
d
Na
h
n
se
n
,
T
.
:
Us
in
g
S
y
n
ta
c
ti
c
I
n
fo
r
ma
ti
o
n
t
o
I
d
e
n
ti
fy
Pl
a
g
ia
rism
.
In
:
2
n
d
W
o
rk
sh
o
p
o
n
B
u
il
d
in
g
E
d
u
c
a
ti
o
n
a
l
A
p
p
li
c
a
t
io
n
s
u
sin
g
NL
P
(
2
0
0
5
).
[1
4
]
A
h
m
e
d
Ha
m
z
a
O
s
m
a
n
,
Na
o
m
ie
S
a
li
m
,
a
n
d
A
lb
a
ra
a
A
b
u
o
b
ie
d
a
.
S
u
rv
e
y
o
f
T
e
x
t
P
lag
iarism
De
tec
ti
o
n
.
Co
m
p
u
ter
En
g
in
e
e
rin
g
a
n
d
A
p
p
li
c
a
ti
o
n
s
Vo
l.
1
,
No
.
1
,
Ju
n
e
2
0
1
2
.
Un
iv
e
rsiti
T
e
k
n
o
lo
g
i
M
a
la
y
sia
,
F
a
c
u
lt
y
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
In
f
o
rm
a
ti
o
n
S
y
ste
m
s
,
S
k
u
d
a
i,
Jo
h
o
r,
M
a
lay
sia
,
In
tern
a
ti
o
n
a
l
Un
iv
e
rsity
o
f
Af
rica
,
F
a
c
u
l
ty
o
f
Co
m
p
u
ter
S
tu
d
ies
,
Kh
a
rto
u
m
,
S
u
d
a
n
.
[1
5
]
V
a
n
i
Ka
n
ji
ra
n
g
a
t,
De
e
p
a
G
u
p
ta.
A
S
tu
d
y
o
n
Ex
tri
n
sic
Tex
t
P
lag
ia
rism
De
te
c
ti
o
n
T
e
c
h
n
iq
u
e
s
a
n
d
T
o
o
ls.
Jo
u
rn
a
l
o
f
En
g
in
e
e
rin
g
S
c
ien
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
Re
v
ie
w
9
(5
)
(2
0
1
6
)
9
–
2
3
.
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter S
c
ien
c
e
&
En
g
in
e
e
rin
g
,
Am
rit
a
S
c
h
o
o
l
o
f
En
g
in
e
e
rin
g
,
A
m
rit
a
Un
iv
e
rsit
y
,
Am
rit
a
V
ish
w
a
V
id
y
a
p
e
e
th
a
m
,
B
a
n
g
a
lo
re
,
In
d
ia.
De
p
a
rtme
n
t
o
f
M
a
th
e
m
a
ti
c
s,
Am
rit
a
S
c
h
o
o
l
o
f
E
n
g
in
e
e
rin
g
,
Am
rit
a
Un
iv
e
rsit
y
,
Am
rit
a
V
ish
w
a
V
id
y
a
p
e
e
th
a
m
,
Ba
n
g
a
lo
re
,
In
d
ia.
[1
6
]
Ch
risti
n
a
Kra
u
s.
P
lag
iarism
De
tec
ti
o
n
-
S
tate
-
of
-
th
e
-
a
rt
sy
ste
m
s
(2
0
1
6
)
a
n
d
e
v
a
lu
a
ti
o
n
m
e
th
o
d
s
.
a
rX
iv
:1
6
0
3
.
0
3
0
1
4
v
1
[
c
s.IR
]
8
M
a
r
2
0
1
6
.
T
e
c
h
n
isc
h
e
Un
iv
e
rsität
Be
rli
n
Da
tab
a
se
S
y
ste
m
s
a
n
d
I
n
f
o
rm
a
ti
o
n
M
a
n
a
g
e
m
e
n
t
G
ro
u
p
.
[1
7
]
Co
ll
o
b
e
rt,
R.
a
n
d
W
e
sto
n
,
J.
A
u
n
if
ied
a
rc
h
i
tec
tu
re
fo
r
n
a
tu
r
a
l
la
n
g
u
a
g
e
p
r
o
c
e
ss
in
g
:
De
e
p
n
e
u
ra
l
n
e
two
rk
s
wit
h
mu
lt
it
a
sk
le
a
rn
i
n
g
.
In
P
ro
c
e
e
d
in
g
s o
f
th
e
2
5
th
in
tern
a
ti
o
n
a
l
c
o
n
f
e
re
n
c
e
o
n
M
a
c
h
in
e
lea
rn
in
g
A
CM
,
1
6
0
-
1
6
7
(2
0
0
8
).
[1
8
]
Ch
o
n
g
,
M
.
Y
.
M
.
,
.
A
stu
d
y
o
n
p
lag
iaris
m
d
e
tec
ti
o
n
a
n
d
p
lag
iarism
d
irec
ti
o
n
id
e
n
t
if
ica
ti
o
n
u
sin
g
n
a
tu
ra
l
lan
g
u
a
g
e
p
ro
c
e
ss
in
g
tec
h
n
i
q
u
e
s (2
0
1
3
).
[1
9
]
M
ik
o
lo
v
,
T
.
,
Ch
e
n
,
K.,
Co
rra
d
o
,
G
.
,
a
n
d
De
a
n
,
J.,
2
0
1
3
.
Ef
f
icie
n
t
e
stim
a
ti
o
n
o
f
w
o
rd
re
p
re
se
n
tatio
n
s
in
v
e
c
to
r
sp
a
c
e
.
a
rX
iv
p
re
p
ri
n
t
a
r
X
iv
:1
3
0
1
.
3
7
8
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
A
d
ee
p
lea
r
n
in
g
b
a
s
ed
tec
h
n
iq
u
e
fo
r
p
la
g
ia
r
is
m
d
etec
tio
n
:
a
co
mp
a
r
a
tive
s
tu
d
y
(
E
l Mo
s
ta
fa
Ha
mb
i
)
89
[2
0
]
Qu
o
c
L
e
a
n
d
T
o
m
a
s
M
ik
o
lo
v
.
Distrib
u
ted
Re
p
re
se
n
tatio
n
s
o
f
S
e
n
ten
c
e
s
a
n
d
Do
c
u
m
e
n
ts.
G
o
o
g
le
In
c
,
1
6
0
0
Am
p
h
it
h
e
a
tre P
a
rk
w
a
y
,
M
o
u
n
tain
V
iew
,
C
A
9
4
0
4
3
.
[2
1
]
S
h
a
o
ji
e
Ba
i,
J.
Zi
c
o
Ko
lt
e
r,
V
la
d
len
Ko
lt
u
n
.
A
n
Em
p
iri
c
a
l
Ev
a
lu
a
ti
o
n
o
f
G
e
n
e
ric
Co
n
v
o
lu
ti
o
n
a
l
a
n
d
Re
c
u
rre
n
t
Ne
tw
o
rk
s f
o
r
S
e
q
u
e
n
c
e
M
o
d
e
li
n
g
.
a
rX
iv
:1
8
0
3
.
0
1
2
7
1
v
2
[
c
s.L
G
]
1
9
A
p
r
2
0
1
8
.
[2
2
]
Ch
risti
a
n
S
.
P
e
ro
n
e
.
P
r
iv
a
c
y
-
p
re
s
e
rv
in
g
se
n
ten
c
e
se
m
a
n
ti
c
sim
il
a
ri
t
y
u
sin
g
In
f
e
rS
e
n
t
e
m
b
e
d
d
in
g
s
a
n
d
se
c
u
re
tw
o
-
p
a
rty
c
o
m
p
u
tatio
n
.
[2
3
]
M
a
m
d
o
u
h
F
a
ro
u
k
.
M
e
a
su
ri
n
g
S
e
n
ten
c
e
s
S
im
il
a
rit
y
:
A
S
u
rv
e
y
.
In
d
ian
Jo
u
rn
a
l
o
f
S
c
ien
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
V
o
l
1
2
(
2
5
),
DO
I:
1
0
.
1
7
4
8
5
/i
jst
/2
0
1
9
/v
1
2
i2
5
/1
4
3
9
7
7
,
J
u
ly
2
0
1
9
.
De
p
a
rtme
n
t
o
f
T
h
e
r
m
o
tec
h
n
ik
Co
m
p
u
ter
S
c
ien
c
e
,
A
s
siu
t
Un
iv
e
rsity
,
M
a
rk
a
z
El
-
F
a
th
,
A
ss
iu
t
G
o
v
e
rn
o
ra
te 7
1
5
1
5
,
Eg
y
p
t.
[2
4
]
Ke
n
su
k
e
Ba
b
a
,
T
e
tsu
y
a
Na
k
a
to
h
a
n
d
T
o
sh
iro
M
in
a
m
i.
Pl
a
g
ia
ris
m
d
e
tec
ti
o
n
u
si
n
g
d
o
c
u
me
n
t
simi
la
rity
b
a
se
d
o
n
d
istrib
u
ted
re
p
re
se
n
ta
ti
o
n
.
8
t
h
In
t
e
rn
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
A
d
v
a
n
c
e
s
in
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
l
o
g
y
,
I
AIT
2
0
1
6
,
1
9
-
22
De
c
e
m
b
e
r
2
0
1
6
,
M
a
c
a
u
,
Ch
in
a
.
F
u
ji
tsu
L
a
b
o
ra
to
ries
,
Ka
w
a
sa
k
i,
Ja
p
a
n
Ky
u
sh
u
Un
iv
e
rsity
,
F
u
k
u
o
k
a
,
Ja
p
a
n
.
[2
5
]
M
a
tt
J.
Ku
sn
e
r,
Yu
S
u
n
,
Nic
h
o
l
a
s
I.
Ko
lk
in
a
n
d
Kili
a
n
Q.
W
e
in
b
e
rg
e
r.
Fro
m
W
o
rd
Emb
e
d
d
i
n
g
s
T
o
Do
c
u
me
n
t
Dista
n
c
e
s
.
P
ro
c
e
e
d
in
g
s
o
f
th
e
3
2
n
d
I
n
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
M
a
c
h
in
e
L
e
a
rn
in
g
,
L
il
le,
F
ra
n
c
e
,
2
0
1
5
.
JML
R:
W
&
CP
v
o
lu
m
e
3
7
.
W
a
sh
in
g
to
n
Un
iv
e
rsit
y
in
S
t.
L
o
u
is,
1
Bro
o
k
in
g
s Dr.,
S
t.
L
o
u
is,
M
O
6
3
1
3
0
.
[2
6
]
Erf
a
n
e
h
G
h
a
ra
v
i,
Ka
y
v
a
n
Bij
a
ri
a
n
d
Kia
ra
sh
Zah
ir
n
ia.
A
De
e
p
L
e
a
r
n
in
g
A
p
p
r
o
a
c
h
to
Per
sia
n
P
la
g
ia
rism
De
tec
ti
o
n
.
DO
I:
1
0
.
1
1
0
9
/ICT
CS
.
2
0
1
7
.
4
2
Co
n
f
e
re
n
c
e
:
Co
n
f
e
r
e
n
c
e
:
T
h
e
In
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
n
e
w
T
re
n
d
s
in
Co
m
p
u
ti
n
g
S
c
ien
c
e
s
(ICT
CS
2
0
1
7
).
U
n
iv
e
rsity
o
f
T
e
h
ra
n
F
a
c
u
lt
y
o
f
n
e
w
S
c
ien
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
Da
ta
&
S
ig
n
a
l
p
ro
c
e
ss
in
g
L
a
b
2
0
1
7
.
[2
7
]
Dim
a
S
u
leim
a
n
,
A
ra
fa
t
Aw
a
jan
a
n
d
A
ra
f
a
t
Aw
a
jan
.
De
e
p
L
e
a
rn
in
g
Ba
se
d
Tec
h
n
iq
u
e
f
o
r
P
lag
iaris
m
De
te
c
ti
o
n
in
A
ra
b
ic
Tex
ts.
2
0
1
7
In
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
Ne
w
T
re
n
d
s
in
Co
m
p
u
ti
n
g
S
c
ien
c
e
s.
Co
m
p
u
ter
S
c
ien
c
e
De
p
a
rtme
n
t
P
rin
c
e
ss
S
u
m
a
y
a
Un
iv
e
rsit
y
f
o
r
T
e
c
h
n
o
l
o
g
y
2
0
1
7
.
[2
8
]
Na
v
e
e
d
Afz
a
l,
Ya
n
sh
a
n
W
a
n
g
a
n
d
Ho
n
g
f
a
n
g
L
iu
.
M
a
y
o
N
L
P
a
t
S
e
m
E
v
a
l
-
2
0
1
6
T
a
sk
1
:
S
e
m
a
n
ti
c
T
e
x
tu
a
l
S
im
il
a
rit
y
b
a
se
d
o
n
L
e
x
ica
l
S
e
m
a
n
ti
c
Ne
t
a
n
d
De
e
p
L
e
a
rn
in
g
S
e
m
a
n
ti
c
M
o
d
e
l.
P
r
o
c
e
e
d
in
g
s
o
f
S
e
m
Ev
a
l
-
2
0
1
6
,
p
a
g
e
s
6
7
4
–
6
7
9
,
S
a
n
Die
g
o
,
Ca
li
f
o
rn
ia,
J
u
n
e
1
6
-
1
7
,
2
0
1
6
.
2
0
1
6
A
ss
o
c
iatio
n
f
o
r
Co
m
p
u
tati
o
n
a
l
L
in
g
u
isti
c
s.
De
p
a
rtm
e
n
t
o
f
He
a
lt
h
S
c
ien
c
e
s Re
se
a
rc
h
M
a
y
o
Cli
n
ic,
R
o
c
h
e
ste
r,
M
N.
[2
9
]
T
e
d
o
V
rb
a
n
e
c
a
n
d
A
n
a
M
e
stro
v
ic.
T
h
e
S
tru
g
g
le
w
it
h
A
c
a
d
e
m
ic
P
lag
iarism
:
A
p
p
ro
a
c
h
e
s
b
a
se
d
o
n
S
e
m
a
n
ti
c
S
im
il
a
rit
y
.
M
IP
R
O
2
0
1
7
,
M
a
y
2
2
-
2
6
,
2
0
1
7
,
Op
a
ti
ja,
Cr
o
a
ti
a
.
F
a
c
u
lt
y
o
f
T
e
a
c
h
e
r
Ed
u
c
a
ti
o
n
,
Un
iv
e
rsity
o
f
Za
g
re
b
,
Cro
a
ti
a
De
p
a
rtm
e
n
t
o
f
[
n
f
o
rm
a
ti
c
s,
Un
iv
e
rsity
o
f
Rij
e
k
a
,
Cro
a
ti
a
.
[3
0
]
Qu
o
c
L
e
a
n
d
T
o
m
a
s
M
ik
o
lo
v
.
Distrib
u
ted
Re
p
re
se
n
tatio
n
s
o
f
S
e
n
ten
c
e
s
a
n
d
Do
c
u
m
e
n
ts.
G
o
o
g
le
In
c
,
1
6
0
0
Am
p
h
it
h
e
a
tre P
a
rk
w
a
y
,
M
o
u
n
tain
V
iew
,
CA
9
4
0
4
3
.
[3
1
]
A
d
rian
S
a
n
b
o
rn
a
n
d
Ja
c
e
k
S
k
r
y
z
a
li
n
.
De
e
p
L
e
a
rn
in
g
f
o
r
S
e
m
a
n
ti
c
S
im
il
a
rit
y
.
M
IP
RO
2
0
1
7
,
M
a
y
2
2
-
2
6
,
2
0
1
7
,
Op
a
ti
ja,
Cro
a
ti
a
.
De
p
a
rtm
e
n
t
o
f
Co
m
p
u
ter S
c
ien
c
e
S
tan
f
o
rd
U
n
iv
e
rsity
.
[3
2
]
Hu
a
He
,
Ke
v
in
G
i
m
p
e
l,
a
n
d
Jim
m
y
L
in
.
M
u
lt
i
-
P
e
rs
p
e
c
ti
v
e
S
e
n
ten
c
e
S
im
il
a
rit
y
M
o
d
e
li
n
g
w
it
h
Co
n
v
o
lu
ti
o
n
a
l
Ne
u
ra
l
Ne
t
w
o
rk
s.
DO
I:
1
0
.
1
8
6
5
3
/v
1
/D1
5
-
1
1
8
1
Co
n
f
e
re
n
c
e
:
Co
n
f
e
r
e
n
c
e
:
P
ro
c
e
e
d
in
g
s
o
f
th
e
2
0
1
5
Co
n
f
e
re
n
c
e
o
n
Em
p
iri
c
a
l
M
e
th
o
d
s
i
n
Na
tu
ra
l
L
a
n
g
u
a
g
e
P
ro
c
e
ss
in
g
.
De
p
a
rtm
e
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
,
Un
iv
e
rsity
o
f
M
a
r
y
lan
d
,
Co
ll
e
g
e
P
a
rk
,
T
o
y
o
ta
T
e
c
h
n
o
l
o
g
ica
l
In
stit
u
te
a
t
Ch
ica
g
o
a
n
d
Da
v
id
R.
Ch
e
rit
o
n
S
c
h
o
o
l
o
f
Co
m
p
u
ter
S
c
ie
n
c
e
,
Un
iv
e
rsit
y
o
f
Wate
rlo
o
.
[3
3
]
T
o
m
Ke
n
ter
a
n
d
M
a
a
rten
d
e
Rij
k
e
.
S
h
o
rt
T
e
x
t
S
imil
a
rity
wit
h
W
o
rd
Emb
e
d
d
i
n
g
s
.
CIKM
'
1
5
P
ro
c
e
e
d
i
n
g
s
o
f
th
e
2
4
t
h
A
CM
In
tern
a
ti
o
n
a
l
o
n
C
o
n
f
e
re
n
c
e
o
n
In
f
o
rm
a
ti
o
n
a
n
d
K
n
o
w
led
g
e
M
a
n
a
g
e
m
e
n
t
P
a
g
e
s
1
4
1
1
-
1
4
2
0
.
Un
iv
e
rsity
o
f
Am
ste
rd
a
m
,
Am
st
e
rd
a
m
,
T
h
e
Ne
th
e
rlan
d
s.
[3
4
]
El
M
o
a
tez
Bil
lah
Na
g
o
u
d
i
a
n
d
Did
ier
S
c
h
w
a
b
.
S
e
ma
n
ti
c
S
imil
a
rity
o
f
Ara
b
ic
S
e
n
ten
c
e
s
wit
h
W
o
rd
Emb
e
d
d
in
g
s
.
P
r
o
c
e
e
d
in
g
s
o
f
T
h
e
T
h
ird
A
ra
b
i
c
Na
tu
ra
l
L
a
n
g
u
a
g
e
P
ro
c
e
ss
in
g
W
o
rk
sh
o
p
(W
A
N
L
P
),
p
a
g
e
s
1
8
–
2
4
,
V
a
len
c
ia
,
S
p
a
in
,
A
p
ril
3
,
2
0
1
7
.
©,
2
0
1
7
A
ss
o
c
iatio
n
f
o
r
Co
m
p
u
tatio
n
a
l
L
in
g
u
isti
c
.
L
IM
-
L
a
b
o
ra
to
ire
d
’In
f
o
rm
a
ti
q
u
e
e
t
d
e
M
a
th
é
m
a
ti
q
u
e
s,
Un
iv
e
rsité Am
a
r
T
e
li
d
ji
d
e
L
a
g
h
o
u
a
t,
A
lg
é
rie.
L
I
G
-
G
E
TAL
P
Un
iv
.
G
re
n
o
b
le A
lp
e
s
F
ra
n
c
e
.
[3
5
]
Ce
d
ric De
Bo
o
m
,
S
tev
e
n
V
a
n
Ca
n
n
e
y
t,
T
h
o
m
a
s
De
m
e
e
st
e
r
a
n
d
Ba
rt
Dh
o
e
d
t.
Re
p
re
se
n
tatio
n
lea
rn
in
g
f
o
r
v
e
r
y
sh
o
rt
tex
ts
u
sin
g
we
ig
h
ted
w
o
rd
e
m
b
e
d
d
i
n
g
a
g
g
re
g
a
ti
o
n
.
J
o
u
rn
a
l
P
a
tt
e
rn
Rec
o
g
n
it
io
n
L
e
tt
e
rs
a
rc
h
iv
e
Vo
lu
m
e
8
0
Iss
u
e
C,
S
e
p
tem
b
e
r
2
0
1
6
P
a
g
e
s
1
5
0
-
1
5
6
.
De
p
a
rtm
e
n
t
o
f
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
lo
g
y
,
T
e
c
h
n
o
lo
g
iep
a
rk
1
5
,
9
0
5
2
Zw
ij
n
a
a
rd
e
,
Be
lg
iu
m
.
[3
6
]
Ce
d
ric
De
Bo
o
m
,
S
tev
e
n
V
a
n
Ca
n
n
e
y
t,
S
tev
e
n
Bo
h
e
z
,
T
h
o
m
a
s
De
m
e
e
ste
r
a
n
d
Ba
rt
Dh
o
e
d
t.
L
e
a
rn
in
g
S
e
m
a
n
ti
c
S
im
il
a
rit
y
f
o
r
V
e
ry
S
h
o
rt
T
e
x
ts.
2
0
1
5
IEE
E
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Da
ta
M
in
i
n
g
W
o
rk
sh
o
p
(
ICDM
W
).
G
h
e
n
t
Un
iv
e
rsit
y
–
iM
in
d
s G
a
sto
n
Cro
m
m
e
n
laa
n
8
-
2
0
1
,
9
0
5
0
G
h
e
n
t,
Be
lg
iu
m
.
[3
7
]
X
in
Ye
,
Hu
i
S
h
e
n
,
X
iao
M
a
,
Ra
z
v
a
n
Bu
n
e
sc
u
,
a
n
d
Ch
a
n
g
L
iu
.
F
ro
m
W
o
rd
E
m
b
e
d
d
in
g
s
T
o
Do
c
u
m
e
n
t
S
im
i
lariti
e
s
f
o
r
I
m
p
ro
v
e
d
In
f
o
r
m
a
ti
o
n
Re
tri
e
v
a
l
in
S
o
f
tw
a
re
En
g
in
e
e
rin
g
.
CS
E
’1
6
,
M
a
y
1
4
-
2
2
,
2
0
1
6
,
A
u
stin
,
T
X
,
USA
.
S
c
h
o
o
l
o
f
El
e
c
tri
c
a
l
En
g
in
e
e
rin
g
a
n
d
C
o
m
p
u
ter S
c
ien
c
e
,
Oh
io
U
n
iv
e
rsity
A
th
e
n
s,
Oh
io
4
5
7
0
1
,
USA
.
[3
8
]
Ya
n
g
q
iu
S
o
n
g
a
n
d
Da
n
Ro
th
.
Un
su
p
e
rv
ise
d
S
p
a
rs
e
V
e
c
to
r
De
n
sifi
c
a
ti
o
n
f
o
r
S
h
o
rt
T
e
x
t
S
imil
a
rity
.
DO
I
:
1
0
.
3
1
1
5
/v
1
/N1
5
-
1
1
3
8
Co
n
f
e
re
n
c
e
:
Co
n
f
e
re
n
c
e
:
P
ro
c
e
e
d
in
g
s
o
f
th
e
2
0
1
5
C
o
n
f
e
re
n
c
e
o
f
th
e
N
o
rth
Am
e
ric
a
n
Ch
a
p
ter
o
f
th
e
A
ss
o
c
iatio
n
f
o
r
Co
m
p
u
tatio
n
a
l
L
in
g
u
isti
c
s:
H
u
m
a
n
Lan
g
u
a
g
e
Tec
h
n
o
lo
g
ies
.
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter S
c
ien
c
e
Un
iv
e
rsity
o
f
Ill
in
o
is at
Urb
a
n
a
-
C
h
a
m
p
a
ig
n
Urb
a
n
a
,
IL
6
1
8
0
1
,
USA
.
[3
9
]
A
d
n
e
n
M
a
h
m
o
u
d
a
n
d
M
o
u
n
ir
Zr
ig
u
i.
S
e
ma
n
ti
c
S
imil
a
rity
An
a
lys
i
s
fo
r
Pa
ra
p
h
ra
se
Id
e
n
ti
fi
c
a
ti
o
n
i
n
Ara
b
ic
T
e
x
ts
.
Co
n
f
e
re
n
c
e
:
Co
n
f
e
re
n
c
e
:
T
h
e
3
1
st
P
a
c
if
ic
A
sia
Co
n
f
e
re
n
c
e
o
n
L
a
n
g
u
a
g
e
,
In
f
o
r
m
a
ti
o
n
a
n
d
Co
m
p
u
tatio
n
P
A
CL
IC
3
1
(
2
0
1
7
),
A
t
Un
iv
e
rsit
y
o
f
th
e
P
h
il
ip
p
i
n
e
s
Ce
b
u
,
Ce
b
u
,
P
h
i
li
p
p
in
e
s.
LAT
ICE
Lab
o
ra
to
ry
Re
se
a
rc
h
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter S
c
ien
c
e
Un
iv
e
rsity
o
f
M
o
n
a
stir,
T
u
n
isia.
[4
0
]
Ba
sa
n
t
Ag
a
r
w
a
la,
He
ri
Ra
m
a
m
p
iaro
a
,
He
lg
e
L
a
n
g
se
th
a
,
M
a
ss
i
m
il
ian
o
Ru
o
c
c
o
.
A
De
e
p
Ne
two
rk
M
o
d
e
l
f
o
r
P
a
ra
p
h
ra
se
De
tec
ti
o
n
in
S
h
o
rt
T
e
x
t
M
e
ss
a
g
e
s.
a
rX
iv
:1
7
1
2
.
0
2
8
2
0
v
1
[
c
s.IR]
7
De
c
2
0
1
7
.
De
p
t.
o
f
Co
m
p
u
ter
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
1
,
Ma
r
ch
20
20
:
81
–
90
90
S
c
ien
c
e
,
No
rw
e
g
ian
Un
iv
e
rsit
y
o
f
S
c
ien
c
e
a
n
d
T
e
c
h
n
o
lo
g
y
,
No
rwa
y
S
wa
m
i
Ke
sh
v
a
n
a
n
d
In
stit
u
te
o
f
T
e
c
h
n
o
l
o
g
y
,
In
d
ia T
e
len
o
r
Re
se
a
rc
h
,
T
ro
n
d
h
e
i
m
,
No
r
w
a
y
.
[4
1
]
Hu
a
He
a
n
d
Jim
m
y
L
in
.
Pa
irwis
e
W
o
rd
In
ter
a
c
ti
o
n
M
o
d
e
li
n
g
wi
t
h
De
e
p
Ne
u
ra
l
Ne
two
rk
s
fo
r
S
e
m
a
n
ti
c
S
imil
a
rity
M
e
a
su
re
me
n
t
.
P
ro
c
e
e
d
i
n
g
s
o
f
NA
ACL
-
H
LT
2
0
1
6
,
p
a
g
e
s
9
3
7
–
9
4
8
,
S
a
n
Die
g
o
,
Ca
li
f
o
rn
ia,
Ju
n
e
1
2
-
1
7
,
2
0
1
6
.
c
©2
0
1
6
A
ss
o
c
ia
ti
o
n
f
o
r
Co
m
p
u
tatio
n
a
l
L
in
g
u
isti
c
s.
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
,
Un
iv
e
rsity
o
f
M
a
ry
l
a
n
d
,
Co
ll
e
g
e
P
a
rk
Da
v
id
R.
Ch
e
rit
o
n
S
c
h
o
o
l
o
f
Co
m
p
u
ter S
c
i
e
n
c
e
,
Un
iv
e
rsit
y
o
f
Wate
rlo
o
.
[4
2
]
Jo
n
a
s
M
u
e
ll
e
r
a
n
d
A
d
it
y
a
T
h
y
a
g
a
ra
jan
.
S
ia
me
se
Rec
u
rr
e
n
t
A
rc
h
it
e
c
tu
re
s
fo
r
L
e
a
r
n
in
g
S
e
n
te
n
c
e
S
imil
a
rity.
P
r
o
c
e
e
d
in
g
s
o
f
th
e
T
h
irt
ieth
AA
A
I
Co
n
fe
re
n
c
e
o
n
A
rti
f
icia
l
In
tell
ig
e
n
c
e
(
AA
A
I
-
1
6
).
Co
m
p
u
ter
S
c
ien
c
e
&
A
rti
f
icia
l
In
telli
g
e
n
c
e
L
a
b
o
ra
to
ry
M
a
ss
a
c
h
u
se
tt
s
In
stit
u
te
o
f
T
e
c
h
n
o
lo
g
y
.
De
p
a
rt
m
e
n
t
o
f
Co
m
p
u
t
e
r
S
c
ien
c
e
a
n
d
En
g
in
e
e
rin
g
M
.
S
.
Ra
m
a
iah
In
stit
u
te o
f
T
e
c
h
n
o
lo
g
y
.
[4
3
]
Hu
a
He
,
Ke
v
in
G
i
m
p
e
l,
a
n
d
Jim
m
y
L
in
.
M
u
lt
i
-
Per
sp
e
c
ti
v
e
S
e
n
ten
c
e
S
imil
a
rity
M
o
d
e
li
n
g
wit
h
Co
n
v
o
lu
ti
o
n
a
l
Ne
u
ra
l
Ne
two
rk
s.
P
r
o
c
e
e
d
in
g
s
o
f
th
e
2
0
1
5
Co
n
f
e
re
n
c
e
o
n
Em
p
iri
c
a
l
M
e
th
o
d
s
in
Na
tu
ra
l
L
a
n
g
u
a
g
e
P
ro
c
e
ss
in
g
,
p
a
g
e
s
1
5
7
6
–
1
5
8
6
,
L
isb
o
n
,
P
o
rt
u
g
a
l,
1
7
-
2
1
S
e
p
tem
b
e
r
2
0
1
5
.
c
©2
0
1
5
A
ss
o
c
iatio
n
f
o
r
Co
m
p
u
tatio
n
a
l
L
in
g
u
isti
c
s.
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
,
Un
iv
e
rsit
y
o
f
M
a
r
y
lan
d
,
Co
ll
e
g
e
P
a
rk
2
T
o
y
o
ta
Tec
h
n
o
lo
g
i
c
a
l
In
stit
u
te
a
t
Ch
ica
g
o
.
Da
v
id
R.
Ch
e
rit
o
n
S
c
h
o
o
l
o
f
Co
m
p
u
ter S
c
ien
c
e
,
U
n
iv
e
rsity
o
f
W
a
t
e
rlo
o
.
[4
4
]
A
li
a
k
se
i
S
e
v
e
r
y
n
,
A
les
s
a
n
d
ro
M
o
sc
h
it
ti
.
L
e
a
rn
in
g
to
Ra
n
k
S
h
o
rt
T
e
x
t
P
a
irs
w
it
h
Co
n
v
o
lu
t
i
o
n
a
l
De
e
p
Ne
u
ra
l
Ne
tw
o
rk
s.
G
o
o
g
le In
c
.
Qa
tar Co
m
p
u
ti
n
g
Re
se
a
rc
h
In
stit
u
te.
[4
5
]
El
v
y
s
L
in
h
a
re
s
P
o
n
tes
,
S
tép
h
a
n
e
Hu
e
t,
A
n
d
ré
a
Ca
rn
e
iro
L
in
h
a
re
s,
Ju
a
n
-
M
a
n
u
e
l
T
o
r
re
s
-
M
o
re
n
o
.
P
re
d
ictin
g
t
h
e
S
e
m
a
n
ti
c
Tex
tu
a
l
S
i
m
il
a
rit
y
w
it
h
S
ia
m
e
se
CNN
a
n
d
L
S
T
M
.
L
I
A
,
Un
iv
e
rsité
d
’
A
v
ig
n
o
n
e
t
d
e
s
P
a
y
s
d
e
V
a
u
c
lu
se
,
Av
ig
n
o
n
,
8
4
0
0
0
F
ra
n
c
e
Un
iv
e
rsid
a
d
e
F
e
d
e
ra
l
d
o
Ce
a
rá
,
S
o
b
ra
l,
Ce
a
rá
Bra
z
il
Éco
le
P
o
ly
tec
h
n
iq
u
e
d
e
M
o
n
tréa
l,
M
o
n
tréa
l,
Ca
n
a
d
a
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
El
M
o
sta
f
a
HA
M
BI
is a P
h
.
D
.
stu
d
e
n
t
o
f
Co
m
p
u
ter S
c
ien
c
e
.
His res
e
a
rc
h
a
re
a
s in
c
lu
d
e
d
a
ta
F
a
o
u
z
ia
Be
n
a
b
b
o
u
is
a
p
r
o
f
e
ss
o
r
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
m
e
m
b
e
r
o
f
Co
m
p
u
te
S
c
ien
c
e
a
n
d
In
f
o
rm
a
ti
o
n
P
ro
c
e
ss
in
g
lab
o
ra
t
o
r
y
.
S
h
e
is
H
e
a
d
o
f
th
e
tea
m
"
Clo
u
d
Co
m
p
u
ti
n
g
,
Ne
tw
o
rk
a
n
d
S
y
st
e
m
s
En
g
in
e
e
rin
g
(CCNSE
)"
.
S
h
e
re
c
e
iv
e
d
h
is
P
h
.
D.
in
Co
m
p
u
ter
S
c
ien
c
e
f
ro
m
th
e
F
a
c
u
lt
y
o
f
S
c
ie
n
c
e
s,
Un
iv
e
rsit
y
M
o
h
a
m
e
d
V
,
M
o
r
o
c
c
o
,
1
9
9
7
.
His
re
se
a
rc
h
a
r
e
a
s
in
c
lu
d
e
c
lo
u
d
Co
m
p
u
ti
n
g
,
d
a
ta m
in
in
g
,
m
a
c
h
in
e
lea
rn
in
g
,
a
n
d
Na
tu
ra
l
Lan
g
u
a
g
e
P
ro
c
e
ss
in
g
.
S
h
e
h
a
s p
u
b
li
sh
e
d
se
v
e
r
a
l
sc
ien
ti
f
ic article
s a
n
d
b
o
o
k
c
h
a
p
ters
in
t
h
e
se
a
re
a
s.
Evaluation Warning : The document was created with Spire.PDF for Python.