I
nte
rna
t
io
na
l J
o
urna
l o
f
Adv
a
nces in Applie
d Science
s
(
I
J
AAS)
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
20
2
6
,
p
p
.
583
~
5
9
3
I
SS
N:
2252
-
8
8
1
4
,
DOI
:
1
0
.
1
1
5
9
1
/ijaas
.
v
1
5
.
i
2
.
pp
583
-
5
9
3
583
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
a
s
.
ia
esco
r
e.
co
m
Prog
ra
m defec
t
p
rediction
m
o
de
l b
a
sed o
n t
o
po
lo
g
y
awa
re node
ev
a
lua
tion po
o
l gra
ph
topo
lo
g
y
mo
del
Da
n L
i
1
,4
,
P
o
h So
o
n J
o
s
ephNg
1
,
P
eng
Yin Cho
o
2
,
K
o
o
Yuen P
ha
n
2
,
Wo
ng
See
Wa
n
3
1
I
n
st
i
t
u
t
e
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
D
i
g
i
t
a
l
I
n
n
o
v
a
t
i
o
n
,
U
C
S
I
U
n
i
v
e
r
si
t
y
,
K
u
a
l
a
L
u
mp
u
r
,
M
a
l
a
y
s
i
a
2
F
a
c
u
l
t
y
o
f
I
n
f
o
r
m
a
t
i
o
n
a
n
d
C
o
mm
u
n
i
c
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
s
i
t
i
T
u
n
k
u
A
b
d
u
l
R
a
h
ma
n
,
K
a
m
p
a
r
,
M
a
l
a
y
s
i
a
3
V
i
se
Tr
a
d
i
n
g
,
P
e
n
a
n
g
,
M
a
l
a
y
si
a
4
S
h
e
n
z
h
e
n
C
EP
R
EI
I
n
d
u
st
r
y
T
e
c
h
n
o
l
o
g
y
R
e
s
e
a
r
c
h
I
n
st
i
t
u
t
e
C
o
.
L
t
d
.
,
S
h
e
n
z
h
e
n
,
C
h
i
n
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Au
g
1
9
,
2
0
2
5
R
ev
is
ed
Ap
r
8
,
2
0
2
6
Acc
ep
ted
Ap
r
2
2
,
2
0
2
6
Trad
it
io
n
a
l
f
u
z
z
in
g
str
u
g
g
les
wit
h
e
fficie
n
c
y
,
a
s
m
a
x
imiz
in
g
c
o
d
e
c
o
v
e
ra
g
e
d
o
e
s
n
o
t
g
u
a
ra
n
tee
th
e
d
isc
o
v
e
ry
o
f
a
d
d
i
ti
o
n
a
l
v
u
ln
e
ra
b
i
li
ti
e
s.
T
o
so
lv
e
t
h
is
,
th
e
stu
d
y
i
n
tro
d
u
c
e
s
to
p
o
lo
g
y
-
a
wa
re
n
o
d
e
e
v
a
lu
a
ti
o
n
(
TANE
-
P
o
o
l
)
,
a
d
e
e
p
lea
rn
in
g
m
o
d
e
l
t
h
a
t
p
r
o
a
c
ti
v
e
ly
p
re
d
icts
d
e
fe
c
ts
to
g
u
i
d
e
th
e
fu
z
z
i
n
g
p
r
o
c
e
ss
.
Th
e
m
o
d
e
l
a
n
a
ly
z
e
s
th
e
stru
c
t
u
ra
l
p
ro
p
e
rti
e
s
o
f
a
p
ro
g
ra
m
’
s
a
tt
rib
u
ted
c
o
n
tro
l
flo
w
g
ra
ph
(
ACFG
)
v
i
a
a
d
iffu
si
o
n
a
tt
e
n
ti
o
n
m
e
c
h
a
n
ism
.
Th
is
p
ro
c
e
ss
id
e
n
ti
fies
fra
g
i
le
c
o
d
e
r
e
g
io
n
s
a
n
d
g
e
n
e
ra
tes
a
st
a
ti
c
v
u
ln
e
ra
b
il
it
y
sc
o
re
(S
VS)
fo
r
e
a
c
h
b
a
sic
b
lo
c
k
.
Th
e
fu
z
z
e
r
th
e
n
u
se
s th
is sc
o
re
to
p
rio
ri
ti
z
e
tes
t
c
a
se
s,
c
o
n
c
e
n
tratin
g
it
s
e
ffo
r
t
s
o
n
t
h
e
a
re
a
s
m
o
st
li
k
e
ly
t
o
c
o
n
t
a
in
flaw
s.
Ev
a
lu
a
ted
o
n
th
e
J
u
li
e
t
tes
t
su
it
e
a
n
d
a
se
t
o
f
re
a
l
-
wo
rld
p
ro
g
ra
m
s,
TANE
-
Po
o
l
d
e
m
o
n
stra
tes
su
p
e
rio
r
p
re
d
i
c
ti
o
n
a
c
c
u
ra
c
y
.
Its
in
te
g
ra
ti
o
n
i
n
t
o
a
f
u
z
z
e
r
sig
n
ifi
c
a
n
t
ly
e
n
h
a
n
c
e
s
th
e
ra
te
o
f
v
u
l
n
e
ra
b
il
it
y
d
isc
o
v
e
ry
,
p
r
o
v
in
g
th
a
t
a
d
e
fe
c
t
-
p
re
d
ictio
n
-
g
u
id
e
d
a
p
p
ro
a
c
h
is
a
m
o
re
e
fficie
n
t
a
n
d
e
ffe
c
ti
v
e
stra
teg
y
fo
r
so
ftwa
re
se
c
u
rit
y
tes
ti
n
g
i
n
su
sta
in
a
b
l
e
in
fra
str
u
c
tu
re
d
e
v
e
lo
p
m
e
n
t
.
K
ey
w
o
r
d
s
:
Ar
tific
ial
in
tellig
en
ce
Fu
zz
y
test
in
g
Gr
ap
h
n
e
u
r
al
n
etwo
r
k
Ma
ch
in
e
lear
n
in
g
Pro
g
r
am
d
e
f
ec
t p
r
e
d
ictio
n
Su
s
tain
ab
le
in
f
r
astru
ctu
r
e
Vu
ln
er
ab
ilit
y
d
is
co
v
er
y
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
:
Po
h
So
o
n
J
o
s
ep
h
N
g
I
n
s
titu
te
o
f
C
o
m
p
u
ter
Scien
ce
an
d
Dig
ital I
n
n
o
v
atio
n
,
UC
SI
Un
iv
er
s
ity
1
s
t.
UC
SI,
UC
SI
H
eig
h
ts
(
T
am
an
C
o
n
n
au
g
h
t)
,
C
h
er
as
-
5
6
0
0
0
,
Ku
ala
L
u
m
p
u
r
,
Ma
lay
s
ia
E
m
ail: jo
s
ep
h
.
n
g
@
u
csiu
n
iv
er
s
ity
.
ed
u
.
m
y
1.
I
NT
RO
D
UCT
I
O
N
So
f
twar
e
s
y
s
tem
s
h
av
e
b
ec
o
m
e
in
d
is
p
en
s
ab
le
to
m
o
d
er
n
s
o
c
iety
,
s
u
p
p
o
r
tin
g
cr
itical
d
o
m
ain
s
s
u
ch
as
ae
r
o
s
p
ac
e,
f
in
an
ce
,
an
d
p
u
b
lic
in
f
r
astru
ctu
r
e.
As s
o
f
twar
e
s
ca
le
an
d
co
m
p
lex
ity
c
o
n
tin
u
e
t
o
in
cr
ea
s
e,
en
s
u
r
in
g
s
y
s
tem
r
eliab
ilit
y
h
as
b
ec
o
m
e
a
m
ajo
r
ch
allen
g
e
,
p
a
r
ticu
l
ar
ly
in
s
ec
u
r
ity
-
s
en
s
itiv
e
en
v
ir
o
n
m
en
ts
.
Am
o
n
g
v
ar
io
u
s
r
eliab
ilit
y
co
n
ce
r
n
s
,
s
o
f
twar
e
v
u
ln
e
r
ab
ilit
ies
—
d
ef
ec
ts
in
tr
o
d
u
ce
d
d
u
r
in
g
d
esig
n
o
r
im
p
lem
en
tatio
n
—
p
o
s
e
s
ev
er
e
r
is
k
s
b
ec
au
s
e
atta
ck
er
s
m
ay
ex
p
lo
it
th
em
to
co
m
p
r
o
m
is
e
s
y
s
tem
s
o
r
o
b
tain
s
en
s
itiv
e
in
f
o
r
m
atio
n
[
1
]
,
[
2
]
.
T
h
e
r
ef
o
r
e
,
ea
r
ly
id
e
n
tific
atio
n
an
d
m
itig
atio
n
o
f
s
o
f
twar
e
v
u
ln
er
ab
ilit
ies
h
av
e
b
ec
o
m
e
cr
u
cial
f
o
r
en
h
an
cin
g
s
o
f
twar
e
s
ec
u
r
ity
,
an
d
s
ec
u
r
ity
test
in
g
h
as
r
ec
ei
v
ed
g
r
o
win
g
atten
tio
n
f
r
o
m
b
o
th
ac
a
d
em
ia
an
d
in
d
u
s
tr
y
[
3
]
,
to
war
d
s
a
r
esil
ien
t in
f
r
astru
ctu
r
e
d
ev
elo
p
m
en
t.
Fu
zz
in
g
h
as
e
m
er
g
ed
as
o
n
e
o
f
th
e
m
o
s
t
ef
f
ec
tiv
e
au
to
m
ated
tec
h
n
iq
u
es
f
o
r
i
d
en
tify
in
g
v
u
ln
er
ab
ilit
ies
in
r
ea
l
-
wo
r
ld
p
r
o
g
r
am
s
.
B
y
co
n
tin
u
o
u
s
ly
m
u
tatin
g
ex
is
tin
g
in
p
u
ts
an
d
m
o
n
ito
r
in
g
ex
ec
u
tio
n
b
eh
av
io
r
s
,
f
u
zz
in
g
ca
n
d
etec
t
cr
ash
es,
ab
n
o
r
m
al
lo
g
ic,
an
d
an
o
m
alo
u
s
s
tates
th
at
in
d
icate
p
o
ten
tial
s
ec
u
r
it
y
f
laws
[
3
]
,
[
4
]
.
Mo
d
e
r
n
f
u
zz
er
s
,
s
u
ch
as
th
o
s
e
b
ased
o
n
e
v
o
lu
tio
n
ar
y
o
r
f
ee
d
b
ac
k
-
d
r
iv
e
n
s
tr
ateg
ies,
p
r
io
r
itize
test
ca
s
e
s
th
at
tr
ig
g
er
n
ew
ex
ec
u
tio
n
p
ath
s
,
e
n
ab
lin
g
d
ee
p
e
r
ex
p
lo
r
atio
n
o
f
a
p
r
o
g
r
am
’
s
b
eh
av
io
r
s
p
ac
e
[
5
]
.
Ho
wev
er
,
th
e
in
p
u
t
s
p
ac
es
o
f
co
m
p
lex
s
o
f
twar
e
s
y
s
tem
s
ar
e
o
f
ten
lar
g
e
an
d
s
tr
u
ctu
r
ally
c
o
n
s
tr
ain
ed
,
m
ak
in
g
r
an
d
o
m
m
u
tatio
n
alo
n
e
in
e
f
f
i
cien
t
f
o
r
v
u
ln
e
r
ab
ilit
y
d
is
co
v
e
r
y
.
As
a
r
esu
lt,
m
an
y
u
n
g
u
i
d
e
d
o
r
wea
k
ly
g
u
i
d
ed
f
u
zz
in
g
s
tr
ateg
ies
g
en
er
ate
a
lar
g
e
n
u
m
b
er
o
f
in
v
alid
in
p
u
ts
,
r
ed
u
cin
g
o
v
e
r
all
test
in
g
ef
f
ec
ti
v
en
ess
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
8
1
4
I
n
t J Ad
v
Ap
p
l Sci
,
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
2
0
2
6
:
5
8
3
-
593
584
T
o
ad
d
r
ess
th
is
is
s
u
e,
co
v
er
a
g
e
-
g
u
id
e
d
g
r
ey
-
b
o
x
f
u
zz
in
g
(
C
GF)
was
in
tr
o
d
u
ce
d
,
e
n
ab
li
n
g
f
u
zz
er
s
to
r
etain
o
n
ly
in
p
u
ts
th
at
ex
p
a
n
d
co
d
e
c
o
v
er
ag
e,
th
er
eb
y
im
p
r
o
v
in
g
te
s
t e
f
f
icien
cy
an
d
v
u
ln
e
r
ab
ilit
y
d
etec
tio
n
[
6
]
.
Desp
ite
th
e
s
u
cc
es
s
o
f
C
G
F
f
r
am
ewo
r
k
s
,
th
ey
s
till
ex
h
ib
it
im
p
o
r
tan
t
lim
itatio
n
s
.
Un
g
u
id
e
d
m
u
tatio
n
wastes
r
eso
u
r
ce
s
o
n
u
n
p
r
o
m
is
in
g
s
ee
d
s
,
s
ch
ed
u
lin
g
s
tr
ateg
ies
o
f
ten
r
ely
s
o
lely
o
n
h
eu
r
is
tics
with
o
u
t
p
r
ed
ictiv
e
ab
ilit
y
,
a
n
d
m
u
tatio
n
o
p
e
r
ato
r
s
r
ar
el
y
in
co
r
p
o
r
at
e
d
ee
p
er
s
em
an
tic
o
r
s
tr
u
ctu
r
al
ch
ar
ac
ter
is
tics
o
f
p
r
o
g
r
a
m
s
.
As
r
esear
ch
o
n
f
u
zz
in
g
h
as
p
r
o
g
r
ess
ed
,
s
ch
o
lar
s
h
av
e
attem
p
ted
to
o
v
er
c
o
m
e
th
ese
wea
k
n
ess
e
s
th
r
o
u
g
h
th
r
e
e
p
r
im
a
r
y
d
ir
e
ctio
n
s
:
f
o
r
m
at
-
c
o
n
s
tr
ain
t
-
b
as
ed
test
g
en
er
atio
n
,
in
f
o
r
m
a
tio
n
-
f
ee
d
b
ac
k
-
b
ased
f
u
zz
in
g
,
a
n
d
s
y
m
b
o
lic
-
ex
ec
u
tio
n
-
b
ased
a
p
p
r
o
ac
h
es
[
7
]
–
[
9
]
.
Fo
r
m
at
-
co
n
s
tr
ain
t
-
b
ased
m
e
th
o
d
s
(
e.
g
.
,
Peac
h
,
SP
I
KE
,
SNOOZ
E
)
g
en
er
ate
s
tr
u
ctu
r
ed
in
p
u
ts
th
at
a
d
h
e
r
e
to
s
tr
ict
s
y
n
tacti
c
s
p
ec
if
i
ca
tio
n
s
,
im
p
r
o
v
in
g
co
v
er
ag
e
f
o
r
p
r
o
to
co
ls
o
r
h
ig
h
ly
s
tr
u
ctu
r
ed
f
ile
f
o
r
m
ats.
Ho
wev
er
,
th
ese
ap
p
r
o
ac
h
es
r
eq
u
ir
e
s
u
b
s
tan
tial
m
an
u
al
ef
f
o
r
t
an
d
ex
p
er
t
d
o
m
ain
k
n
o
wled
g
e
,
lim
itin
g
th
eir
s
ca
lab
ilit
y
.
I
n
f
o
r
m
atio
n
-
f
ee
d
b
ac
k
-
b
ased
f
u
zz
in
g
m
eth
o
d
s
—
r
ep
r
esen
te
d
b
y
A
m
er
ican
f
u
zz
y
lo
p
(
AFL
)
,
A
FLFast,
an
d
Sy
zk
aller
—
u
s
e
r
u
n
tim
e
d
ata
s
u
ch
as
co
d
e
co
v
er
a
g
e
to
g
u
id
e
test
ca
s
e
m
u
tatio
n
an
d
s
elec
tio
n
,
s
ig
n
if
ican
tly
en
h
an
ci
n
g
s
ea
r
ch
ef
f
icien
cy
[
1
0
]
–
[
1
2
]
.
Yet,
f
o
r
p
r
o
g
r
am
s
with
co
m
p
lex
o
r
d
ee
p
ly
n
ested
in
p
u
t
s
tr
u
ctu
r
es,
th
ese
ap
p
r
o
ac
h
es
s
till
s
u
f
f
er
f
r
o
m
th
e
g
en
er
atio
n
o
f
in
v
alid
test
ca
s
es,
wh
ich
lo
wer
s
m
u
tatio
n
e
f
f
ic
ien
cy
an
d
s
lo
ws v
u
ln
er
ab
ilit
y
d
is
co
v
er
y
.
T
o
f
u
r
th
er
im
p
r
o
v
e
t
esti
n
g
p
r
e
cisi
o
n
an
d
co
v
er
ag
e
,
s
y
m
b
o
lic
ex
ec
u
tio
n
–
b
ased
ap
p
r
o
ac
h
es
h
av
e
b
ee
n
in
tr
o
d
u
ce
d
to
g
en
er
ate
h
ig
h
-
q
u
ality
,
p
ath
-
d
ir
ec
ted
test
ca
s
es.
T
o
o
ls
s
u
ch
as
SAGE
,
KL
E
E
,
an
d
T
ain
tSco
p
e
an
aly
ze
ex
ec
u
tio
n
p
ath
s
s
y
m
b
o
lically
an
d
s
o
lv
e
co
n
s
tr
ain
ts
to
cr
af
t
in
p
u
ts
th
at
f
o
r
ce
ex
p
l
o
r
atio
n
o
f
s
p
ec
if
ic
p
r
o
g
r
a
m
b
r
a
n
ch
es
[
1
3
]
–
[
1
5
]
.
Alth
o
u
g
h
h
ig
h
ly
ef
f
ec
tiv
e
in
th
eo
r
y
,
s
y
m
b
o
lic
ex
ec
u
tio
n
f
ac
es
cr
itical
ch
allen
g
es
s
u
ch
as
p
at
h
ex
p
lo
s
io
n
an
d
th
e
h
ig
h
c
o
m
p
u
tatio
n
al
co
s
t
o
f
co
n
s
tr
ain
t
s
o
lv
i
n
g
,
wh
ich
lim
it
its
ap
p
licab
ilit
y
in
lar
g
e
-
s
ca
le
r
ea
l
-
wo
r
ld
s
o
f
twar
e
s
y
s
tem
s
.
Giv
en
th
ese
lim
itatio
n
s
ac
r
o
s
s
ex
is
tin
g
f
u
zz
in
g
p
a
r
ad
ig
m
s
,
t
h
er
e
r
e
m
ain
s
a
s
u
b
s
tan
tial
n
ee
d
f
o
r
m
o
r
e
in
tellig
en
t,
s
tr
u
ctu
r
e
-
awa
r
e,
a
n
d
r
eso
u
r
ce
-
ef
f
icien
t
f
u
z
zin
g
g
u
id
an
ce
m
ec
h
an
is
m
s
.
Mo
tiv
ated
b
y
th
is
g
ap
,
th
is
s
tu
d
y
in
tr
o
d
u
ce
s
to
p
o
lo
g
y
-
a
war
e
n
o
d
e
ev
alu
ati
o
n
(
T
A
NE
-
Po
o
l)
,
a
d
ef
ec
t
-
p
r
ed
ictio
n
-
en
h
a
n
ce
d
f
u
zz
i
n
g
f
r
am
ewo
r
k
.
Un
lik
e
p
r
e
v
io
u
s
wo
r
k
s
th
at
r
ely
s
o
lely
o
n
co
d
e
m
etr
ics
o
r
f
lat
n
eu
r
al
n
etwo
r
k
s
,
th
e
c
o
r
e
n
o
v
elty
o
f
o
u
r
a
p
p
r
o
ac
h
lies
in
th
e
f
u
s
io
n
o
f
T
ANE
-
b
ased
d
ep
e
n
d
e
n
cy
ex
tr
ac
tio
n
with
a
h
ier
ar
c
h
ical
g
r
ap
h
p
o
o
lin
g
s
tr
ateg
y
.
Sp
ec
if
ically
,
p
r
o
g
r
a
m
d
ep
en
d
en
cies
ex
tr
ac
te
d
v
i
a
T
ANE
ca
p
tu
r
e
laten
t
s
tr
u
c
tu
r
al
r
elatio
n
s
h
ip
s
co
r
r
elate
d
with
d
ef
ec
t
-
p
r
o
n
en
ess
.
B
y
en
co
d
in
g
th
ese
d
ep
en
d
en
cies
in
to
g
r
ap
h
r
ep
r
esen
tatio
n
s
an
d
ap
p
ly
i
n
g
a
n
ad
ap
tiv
e
p
o
o
lin
g
m
ec
h
an
is
m
,
th
e
p
r
o
p
o
s
ed
m
o
d
el
ca
n
id
en
t
if
y
s
ee
d
s
with
a
h
ig
h
er
p
r
o
b
a
b
ilit
y
o
f
tr
ig
g
e
r
in
g
s
ec
u
r
ity
-
r
elev
an
t
b
e
h
av
io
r
s
.
T
h
e
p
r
ed
ictio
n
s
p
r
o
d
u
ce
d
b
y
t
h
e
T
ANE
-
Po
o
l
m
o
d
el
a
r
e
f
u
r
th
er
in
teg
r
ated
in
t
o
s
ch
ed
u
lin
g
an
d
m
u
tatio
n
s
tr
ateg
ies,
en
ab
lin
g
m
o
r
e
ef
f
ici
en
t
allo
ca
tio
n
o
f
f
u
zz
in
g
r
es
o
u
r
ce
s
an
d
d
ee
p
er
ex
p
lo
r
atio
n
o
f
p
r
ev
io
u
s
ly
u
n
r
e
ac
h
ab
le
ex
ec
u
tio
n
p
ath
s
.
T
h
e
m
ain
co
n
tr
ib
u
tio
n
s
o
f
th
i
s
s
tu
d
y
ar
e
th
r
ee
f
o
ld
.
First,
we
p
r
o
p
o
s
e
a
d
e
p
en
d
en
cy
-
d
r
i
v
en
d
ef
ec
t
p
r
ed
ictio
n
m
o
d
el
b
ased
o
n
T
ANE
an
d
g
r
a
p
h
p
o
o
lin
g
th
at
ca
p
tu
r
es
s
tr
u
ctu
r
al
cu
es
ass
o
c
iated
with
p
r
o
g
r
am
v
u
ln
er
ab
ilit
ies.
Seco
n
d
,
we
d
e
s
ig
n
a
p
r
ed
ictio
n
-
g
u
id
ed
s
ch
e
d
u
lin
g
s
tr
ateg
y
th
at
p
r
i
o
r
itizes
s
ee
d
s
u
s
in
g
d
ef
ec
t
lik
elih
o
o
d
a
n
d
f
itn
ess
s
co
r
in
g
,
im
p
r
o
v
in
g
t
h
e
ef
f
icien
c
y
o
f
r
eso
u
r
ce
allo
ca
tio
n
with
in
th
e
f
u
zz
in
g
lo
o
p
.
T
h
ir
d
,
we
in
tr
o
d
u
ce
a
m
u
tatio
n
m
ec
h
an
is
m
en
h
an
ce
d
b
y
lear
n
e
d
r
e
p
r
esen
tatio
n
s
,
en
ab
lin
g
d
ee
p
er
an
d
m
o
r
e
tar
g
eted
p
ath
ex
p
lo
r
atio
n
.
E
x
p
er
im
en
t
al
r
esu
lts
d
em
o
n
s
tr
ate
th
at
th
e
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
s
ig
n
if
ic
an
tly
im
p
r
o
v
es
b
o
th
co
d
e
c
o
v
er
a
g
e
a
n
d
c
r
ash
d
is
co
v
er
y
co
m
p
ar
ed
with
s
tate
-
of
-
th
e
-
ar
t
f
u
zz
er
s
.
T
h
is
wo
r
k
p
r
o
v
id
es
a
p
r
o
m
is
in
g
s
tep
to
war
d
th
e
d
e
v
elo
p
m
e
n
t o
f
in
tellig
en
t,
s
ca
lab
le,
an
d
s
ec
u
r
ity
-
en
h
an
cin
g
au
to
m
ated
test
in
g
s
y
s
tem
s
.
T
h
e
r
em
ai
n
d
er
o
f
th
is
p
ap
e
r
i
s
o
r
g
an
ize
d
as
f
o
llo
ws.
Sectio
n
2
r
e
v
iews
th
e
r
elate
d
wo
r
k
o
n
d
ee
p
lear
n
in
g
i
n
s
o
f
twar
e
test
in
g
an
d
g
r
ap
h
-
b
ased
v
u
l
n
er
ab
il
ity
d
etec
tio
n
.
Sectio
n
3
d
et
ails
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
l
o
g
y
,
in
clu
d
i
n
g
t
h
e
T
ANE
-
b
ased
d
ep
e
n
d
en
c
y
e
x
tr
ac
tio
n
,
g
r
ap
h
co
n
s
tr
u
ctio
n
,
a
n
d
th
e
T
ANE
-
Po
o
l
m
o
d
el
ar
ch
itectu
r
e
.
Sectio
n
4
p
r
esen
ts
th
e
ex
p
er
im
en
tal
s
etu
p
,
d
ataset
d
escr
ip
tio
n
,
an
d
a
co
m
p
r
e
h
en
s
iv
e
an
aly
s
is
o
f
th
e
r
esu
lts
co
m
p
ar
ed
to
s
tate
-
of
-
th
e
-
ar
t
b
aselin
es.
Fin
ally
,
s
ec
tio
n
5
c
o
n
cl
u
d
es
th
e
p
a
p
er
an
d
o
u
tl
in
es d
ir
ec
tio
n
s
f
o
r
f
u
t
u
r
e
r
esear
ch
.
2.
L
I
T
E
R
AT
U
RE
R
E
VI
E
W
I
n
r
ec
e
n
t
y
ea
r
s
,
th
e
r
a
p
id
a
d
v
an
ce
m
en
t
o
f
d
ee
p
lea
r
n
in
g
i
n
f
ield
s
s
u
ch
as
im
ag
e
r
ec
o
g
n
itio
n
an
d
n
atu
r
al
lan
g
u
ag
e
p
r
o
ce
s
s
in
g
[
1
6
]
,
[
1
7
]
h
as
s
p
u
r
r
ed
i
n
ter
est
in
it
s
p
o
ten
tial
ap
p
licatio
n
to
s
o
f
twar
e
test
in
g
[
1
8
]
.
As
an
ar
tific
ial
i
n
tellig
en
ce
te
ch
n
o
lo
g
y
with
p
o
wer
f
u
l
lear
n
in
g
an
d
g
en
er
aliza
tio
n
ca
p
a
b
ilit
ies,
d
ee
p
lear
n
in
g
o
f
f
er
s
n
ew
p
ar
a
d
ig
m
s
f
o
r
th
e
ev
o
lu
tio
n
o
f
f
u
zz
in
g
tech
n
iq
u
es.
C
u
r
r
e
n
tly
,
a
g
r
o
win
g
b
o
d
y
o
f
r
esear
ch
is
ex
p
lo
r
in
g
th
e
d
ee
p
in
teg
r
atio
n
o
f
th
ese
two
f
ield
s
to
en
h
a
n
c
e
th
e
ef
f
ec
tiv
e
n
ess
an
d
ef
f
icie
n
cy
o
f
f
u
zz
in
g
[
1
9
]
.
Giv
en
th
at
t
h
e
p
e
r
f
o
r
m
an
ce
o
f
f
u
zz
in
g
lar
g
ely
d
ep
e
n
d
s
o
n
th
e
q
u
ality
o
f
test
ca
s
es,
o
p
tim
izin
g
test
ca
s
e
g
en
er
atio
n
a
n
d
s
ch
e
d
u
lin
g
u
s
in
g
d
ee
p
lear
n
in
g
h
as e
m
er
g
ed
as a
k
ey
r
esear
ch
d
ir
ec
tio
n
[
2
0
]
.
I
n
f
u
zz
in
g
,
th
e
test
ca
s
e
s
ch
ed
u
lin
g
s
tr
ateg
y
is
a
k
ey
d
e
ter
m
in
an
t
o
f
test
in
g
ef
f
icac
y
,
d
ir
ec
tly
im
p
ac
tin
g
r
eso
u
r
ce
allo
ca
tio
n
ef
f
icien
cy
an
d
th
e
d
e
p
th
o
f
v
u
ln
er
a
b
ilit
y
d
is
co
v
e
r
y
.
T
r
a
d
itio
n
al
ap
p
r
o
ac
h
es
t
y
p
ically
ev
alu
ate
ea
ch
test
c
ase
in
th
e
co
r
p
u
s
b
ased
o
n
i
ts
h
is
to
r
ical
p
er
f
o
r
m
an
ce
,
ass
ig
n
in
g
weig
h
ts
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ad
v
Ap
p
l Sci
I
SS
N:
2252
-
8
8
1
4
P
r
o
g
r
a
m
d
efec
t p
r
ed
ictio
n
mo
d
el
b
a
s
ed
o
n
to
p
o
lo
g
y
a
w
a
r
e
n
o
d
e
ev
a
lu
a
tio
n
p
o
o
l g
r
a
p
h
…
(
Da
n
Li
)
585
d
eter
m
in
e
th
e
p
r
io
r
ity
an
d
f
r
eq
u
e
n
cy
o
f
s
u
b
s
eq
u
e
n
t
m
u
tatio
n
s
.
Ho
wev
er
,
th
ese
tr
a
d
itio
n
al
ev
alu
atio
n
s
tr
ateg
ies
ar
e
o
f
ten
lim
ited
in
th
eir
ab
ilit
y
to
ca
p
tu
r
e
co
m
p
lex
p
r
o
g
r
am
b
eh
a
v
io
r
s
.
T
h
e
a
d
v
an
ce
m
e
n
t
o
f
AI
,
p
ar
ticu
lar
ly
d
ee
p
lear
n
in
g
,
h
a
s
led
to
its
ap
p
licatio
n
in
s
o
f
twar
e
test
in
g
,
wh
er
e
it
h
as
b
ee
n
r
ec
o
g
n
ize
d
f
o
r
its
p
er
f
o
r
m
an
ce
o
n
c
o
m
p
lex
task
s
.
I
n
f
u
zz
in
g
,
ap
p
ly
i
n
g
d
ee
p
l
ea
r
n
in
g
t
o
k
e
y
asp
ec
ts
s
u
ch
as
test
ca
s
e
s
cr
ee
n
in
g
an
d
s
ch
ed
u
lin
g
h
as
y
ield
ed
r
em
ar
k
a
b
le
r
esu
lts
,
s
ig
n
if
i
ca
n
tly
im
p
r
o
v
in
g
b
o
t
h
test
in
g
ef
f
icien
cy
a
n
d
v
u
ln
er
ab
ilit
y
d
is
co
v
er
y
[
2
1
]
.
T
h
is
s
ec
tio
n
r
ev
iews th
e
r
esear
ch
p
r
o
g
r
ess
o
f
d
ee
p
lear
n
in
g
in
th
ese
two
ar
ea
s
.
I
n
th
e
d
o
m
ai
n
o
f
test
ca
s
e
s
ch
ed
u
lin
g
,
r
esear
c
h
er
s
b
u
ild
d
ee
p
n
eu
r
al
n
etwo
r
k
m
o
d
els to
p
ar
s
e
a
tar
g
et
p
r
o
g
r
a
m
’
s
s
o
u
r
ce
o
r
b
in
ar
y
c
o
d
e.
T
h
e
o
b
jectiv
e
i
s
t
o
e
x
tr
ac
t
f
ea
tu
r
es
clo
s
ely
r
elate
d
t
o
p
r
o
g
r
am
b
eh
av
i
o
r
a
n
d
id
en
tify
p
o
ten
tially
v
u
ln
e
r
ab
le
r
eg
io
n
s
.
B
y
tr
ain
in
g
m
o
d
els
to
lear
n
th
e
d
is
tr
ib
u
tio
n
al
ch
a
r
ac
ter
is
tics
o
f
th
ese
cr
itical
r
eg
io
n
s
,
th
e
test
ca
s
e
s
ch
ed
u
lin
g
s
tr
ateg
y
ca
n
b
e
in
t
ellig
en
tly
o
p
tim
ize
d
,
g
u
id
in
g
th
e
f
u
zz
e
r
to
m
o
r
e
ef
f
ec
tiv
ely
co
v
er
p
o
ten
tially
h
ig
h
-
r
is
k
ex
ec
u
tio
n
p
at
h
s
.
T
y
p
i
ca
l
r
esear
ch
ef
f
o
r
ts
in
th
is
ar
e
a
in
clu
d
e
Neu
Fu
zz
an
d
V
-
Fu
zz
.
Neu
Fu
zz
u
tili
ze
s
d
ee
p
r
ein
f
o
r
ce
m
e
n
t
lear
n
i
n
g
m
eth
o
d
s
to
d
y
n
am
ically
ad
ju
s
t
its
te
s
t
ca
s
e
s
elec
tio
n
s
tr
a
teg
y
,
im
p
r
o
v
i
n
g
t
h
e
v
alid
ity
o
f
test
s
am
p
les
as
p
ath
ex
p
lo
r
atio
n
d
ee
p
en
s
[
2
2
]
.
V
-
Fu
zz
co
n
s
tr
u
cts
a
m
o
d
el
b
ased
o
n
a
n
atten
tio
n
m
ec
h
an
is
m
to
m
ea
s
u
r
e
th
e
co
r
r
elatio
n
b
etwe
en
in
p
u
ts
an
d
co
d
e
co
v
e
r
ag
e,
th
er
eb
y
p
r
i
o
r
itizin
g
th
e
s
ch
ed
u
lin
g
o
f
h
ig
h
-
v
alu
e
test
ca
s
es
[
2
3
]
.
T
h
is
u
s
e
o
f
atten
tio
n
is
p
a
r
t
o
f
a
b
r
o
ad
e
r
tr
en
d
in
ap
p
ly
i
n
g
s
o
p
h
is
ticated
g
r
a
p
h
n
eu
r
al
n
etwo
r
k
(
GNN)
ar
ch
itectu
r
es
to
co
d
e.
T
o
h
a
n
d
le
th
e
c
o
m
p
lex
ity
o
f
lar
g
e
c
o
d
e
g
r
ap
h
s
,
h
ier
ar
ch
ical
p
o
o
lin
g
m
eth
o
d
s
h
av
e
b
ee
n
i
n
tr
o
d
u
ce
d
to
c
r
e
ate
co
ar
s
e
-
g
r
ain
e
d
r
ep
r
esen
tatio
n
s
o
f
th
e
g
r
a
p
h
.
Mo
d
els
lik
e
Dif
f
Po
o
l
lear
n
a
d
if
f
er
en
tiab
le
clu
s
ter
ass
ig
n
m
en
t
to
g
r
o
u
p
n
o
d
es
[
2
4
]
,
wh
ile
SAGPo
o
l
u
s
es
a
s
elf
-
atten
tio
n
m
ec
h
an
is
m
to
s
co
r
e
an
d
s
elec
t
t
h
e
m
o
s
t
im
p
o
r
tan
t
n
o
d
es
[
2
5
]
.
T
h
es
e
tech
n
iq
u
es
ar
e
v
ital
f
o
r
g
r
ap
h
class
if
icati
o
n
task
s
,
s
u
ch
as
d
eter
m
in
in
g
if
an
en
tire
f
u
n
ctio
n
is
v
u
ln
er
ab
le
,
b
u
t
th
eir
ef
f
ec
tiv
en
ess
d
ep
en
d
s
h
ea
v
ily
o
n
h
o
w
well
th
ey
p
r
eser
v
e
th
e
to
p
o
lo
g
ical
in
f
o
r
m
atio
n
r
elev
an
t to
s
ec
u
r
ity
d
ef
ec
ts
.
A
s
ig
n
if
ican
t
s
u
b
f
ie
ld
in
th
i
s
d
o
m
ain
lev
er
a
g
es
g
r
ap
h
-
b
a
s
ed
r
ep
r
esen
tatio
n
s
o
f
s
o
u
r
ce
co
d
e
f
o
r
v
u
ln
er
ab
ilit
y
p
r
ed
ictio
n
.
R
e
s
ea
r
ch
er
s
h
av
e
u
tili
ze
d
s
tr
u
ctu
r
es
s
u
ch
as
co
n
tr
o
l
f
lo
w
g
r
ap
h
s
(
C
FGs
)
,
ab
s
tr
ac
t
s
y
n
tax
tr
ee
s
(
ASTs)
,
an
d
co
d
e
p
r
o
p
er
ty
g
r
ap
h
s
(
C
PG
s
)
as
in
p
u
ts
f
o
r
d
ee
p
lear
n
in
g
m
o
d
els
[
2
6
]
.
Fo
r
in
s
tan
ce
,
s
o
m
e
m
o
d
els
a
p
p
ly
GNNs
d
ir
ec
tly
to
th
ese
g
r
ap
h
s
to
lear
n
s
tr
u
ctu
r
al
p
a
tter
n
s
in
d
icativ
e
o
f
v
u
ln
er
ab
ilit
ies
lik
e
b
u
f
f
er
o
v
e
r
f
lo
ws
o
r
u
s
e
-
af
ter
-
f
r
ee
er
r
o
r
s
.
T
h
e
o
u
tp
u
ts
o
f
th
ese
p
r
ed
icto
r
s
ar
e
th
en
u
s
e
d
to
cr
ea
te
a
“
h
ea
t
m
a
p
”
o
f
p
o
ten
t
ially
v
u
ln
e
r
ab
le
f
u
n
ctio
n
s
o
r
b
asic
b
lo
ck
s
,
allo
win
g
f
u
zz
er
s
to
p
r
io
r
itize
i
n
p
u
ts
th
at
ex
er
cise
th
ese
h
ig
h
-
r
is
k
co
d
e
r
eg
io
n
s
.
C
o
n
cu
r
r
en
tly
,
m
ain
s
tr
ea
m
f
u
zz
in
g
en
g
in
es
em
p
lo
y
ev
o
l
u
tio
n
ar
y
s
ea
r
ch
alg
o
r
ith
m
s
—
s
u
ch
as
co
v
er
ag
e
-
g
u
id
e
d
g
en
etic
s
tr
ateg
ies
o
r
v
u
ln
er
ab
ilit
y
-
p
r
io
r
it
ized
m
u
tatio
n
—
to
d
ir
ec
t
test
ca
s
e
g
en
er
atio
n
to
war
d
u
n
co
v
er
ed
a
n
d
h
i
g
h
-
r
is
k
p
r
o
g
r
am
p
ath
s
[
2
3
]
,
[
2
7
]
.
T
h
i
s
ap
p
r
o
ac
h
m
a
r
k
s
a
s
h
if
t
f
r
o
m
co
v
er
ag
e
-
g
u
id
a
n
ce
to
v
u
ln
er
a
b
ilit
y
-
g
u
i
d
an
ce
,
th
o
u
g
h
th
e
ac
c
u
r
ac
y
o
f
th
e
u
n
d
e
r
ly
in
g
g
r
a
p
h
m
o
d
el
is
p
ar
am
o
u
n
t t
o
its
s
u
cc
ess
.
I
n
ter
m
s
o
f
test
ca
s
e
s
cr
ee
n
in
g
,
ex
is
tin
g
r
esear
ch
o
f
ten
e
m
p
l
o
y
s
n
eu
r
al
n
etwo
r
k
m
o
d
els
as
p
r
ed
i
cto
r
s
to
d
eter
m
in
e
if
a
p
en
d
in
g
te
s
t
ca
s
e
is
lik
ely
to
tr
ig
g
er
a
n
ew
p
r
o
g
r
am
s
tate,
s
u
ch
as
an
ex
ce
p
tio
n
o
r
a
n
u
n
ex
p
l
o
r
ed
co
d
e
r
eg
io
n
.
So
m
e
s
tu
d
ies co
n
s
tr
u
ct
an
n
o
tated
d
atasets
b
y
co
llectin
g
h
is
to
r
ical
test
ca
s
e
s
an
d
th
eir
r
esu
ltin
g
o
u
tco
m
es,
wh
ic
h
a
r
e
th
e
n
u
s
ed
to
tr
ai
n
a
cla
s
s
if
icatio
n
m
o
d
el
[
2
8
]
,
[
2
9
]
.
T
h
is
m
o
d
el
ca
n
s
u
b
s
eq
u
en
tly
p
r
ed
ict
w
h
eth
er
n
ew
in
p
u
ts
ar
e
lik
ely
to
tr
ig
g
e
r
n
o
v
el
b
e
h
av
io
r
s
,
p
r
o
v
id
in
g
a
b
asis
f
o
r
test
ca
s
e
s
cr
ee
n
in
g
in
f
u
zz
i
n
g
.
Ho
we
v
er
,
a
s
y
s
tem
atic
ev
alu
atio
n
o
f
s
u
ch
m
eth
o
d
s
r
ev
ea
led
th
at
wh
ile
th
ey
ar
e
ef
f
ec
tiv
e
in
s
o
m
e
s
ce
n
ar
io
s
,
t
h
eir
o
v
er
all
co
n
tr
ib
u
tio
n
to
en
h
an
cin
g
f
u
zz
in
g
ef
f
icien
cy
is
lim
ited
.
I
n
co
n
tr
ast,
th
e
Fu
zz
Gu
ar
d
f
r
a
m
ewo
r
k
i
n
tr
o
d
u
ce
d
a
p
r
e
d
ictio
n
m
ec
h
an
is
m
co
m
b
in
e
d
with
a
d
e
ep
lear
n
in
g
m
o
d
el.
Fu
zz
Gu
ar
d
u
tili
ze
s
p
r
e
v
io
u
s
t
est
in
p
u
ts
an
d
th
eir
r
u
n
tim
e
f
ee
d
b
ac
k
as
tr
ain
in
g
d
ata
t
o
b
u
ild
a
m
o
d
el
th
at
s
cr
ee
n
s
o
u
t
in
ef
f
ec
tiv
e
o
r
r
ed
u
n
d
an
t
test
ca
s
es,
th
er
eb
y
en
h
an
cin
g
r
eso
u
r
ce
u
tili
za
tio
n
an
d
im
p
r
o
v
i
n
g
o
v
e
r
all
test
in
g
ef
f
icien
cy
[
3
0
]
–
[
3
2
]
.
Gan
et
a
l
.
[
3
3
]
p
r
o
p
o
s
ed
GR
E
YONE
,
a
d
ata
f
lo
w
s
en
s
itiv
e
f
u
zz
er
th
at
u
s
es
f
u
zz
in
g
-
d
r
iv
en
tain
t
in
f
e
r
en
ce
an
d
c
o
n
f
o
r
m
an
ce
-
g
u
i
d
ed
e
v
o
lu
tio
n
to
o
p
tim
ize
m
u
tatio
n
d
ir
ec
tio
n
,
en
h
an
cin
g
ex
ec
u
tio
n
ef
f
icien
cy
a
n
d
v
u
ln
er
ab
ilit
y
d
is
co
v
er
y
p
er
f
o
r
m
a
n
ce
3.
RE
S
E
ARCH
M
E
T
H
O
D
T
h
is
s
ec
tio
n
d
etails
th
e
p
r
o
p
o
s
ed
T
ANE
-
Po
o
l
f
r
am
ewo
r
k
,
d
esig
n
ed
to
ad
d
r
ess
th
e
lim
itatio
n
s
o
f
b
lin
d
m
u
tatio
n
in
tr
ad
itio
n
al
f
u
zz
in
g
[
3
4
]
–
[
3
6
]
.
T
h
e
m
et
h
o
d
o
lo
g
y
f
o
llo
ws
a
p
ip
elin
e
a
p
p
r
o
ac
h
:
i)
ex
tr
ac
tin
g
laten
t
p
r
o
g
r
am
d
e
p
en
d
e
n
cies
u
s
in
g
th
e
T
ANE
alg
o
r
ith
m
;
ii)
co
n
s
tr
u
ctin
g
attr
ib
u
ted
c
o
n
tr
o
l
f
l
o
w
g
r
a
p
h
s
(
AC
FGs
)
th
at
en
co
d
e
b
o
th
s
tr
u
ctu
r
al
an
d
s
em
an
tic
f
ea
t
u
r
es; iii)
ap
p
ly
in
g
a
n
o
v
el
g
r
ap
h
p
o
o
lin
g
m
ec
h
an
is
m
to
id
en
tify
h
ig
h
-
r
is
k
r
eg
i
o
n
s
;
iv
)
p
r
e
d
ictin
g
d
ef
ec
t
p
r
o
b
ab
ilit
ies;
an
d
v
)
in
teg
r
atin
g
t
h
ese
p
r
ed
ictio
n
s
in
to
a
f
u
zz
in
g
s
c
h
e
d
u
li
n
g
s
tr
ate
g
y
.
T
h
is
s
te
p
-
by
-
s
te
p
p
r
o
c
e
d
u
r
e
e
n
s
u
r
es
th
e
r
e
p
r
o
d
u
ci
b
ili
ty
o
f
e
x
p
er
i
m
e
n
ts
[
3
7
]
–
[
3
9
]
.
3
.
1
.
T
ANE
-
ba
s
ed
depe
nd
en
cy
ex
t
ra
c
t
i
on
T
o
o
v
er
c
o
m
e
th
e
lim
itatio
n
s
o
f
s
tatic
an
aly
s
is
,
wh
ich
o
f
ten
m
is
s
es
co
m
p
lex
d
ata
r
elatio
n
s
h
ip
s
,
th
e
T
ANE
Po
o
l,
a
d
ep
e
n
d
en
c
y
d
is
co
v
er
y
al
g
o
r
ith
m
,
is
em
p
l
o
y
ed
to
ex
tr
ac
t
laten
t
r
elatio
n
s
h
i
p
s
with
in
th
e
tar
g
et
b
in
ar
y
.
W
h
ile
s
tan
d
ar
d
C
FGs
o
n
ly
ca
p
tu
r
e
ex
ec
u
tio
n
p
at
h
s
,
th
ey
f
ail
to
r
e
p
r
esen
t
th
e
f
u
n
ct
io
n
al
d
ep
en
d
en
cies
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
8
1
4
I
n
t J Ad
v
Ap
p
l Sci
,
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
2
0
2
6
:
5
8
3
-
593
586
b
etwe
en
v
ar
ia
b
les
ac
r
o
s
s
d
if
f
er
en
t
b
asic
b
lo
ck
s
.
J
u
s
tific
atio
n
:
T
ANE
was
s
elec
ted
b
ec
au
s
e
it
ef
f
ec
tiv
ely
m
in
es
f
u
n
ctio
n
al
d
ep
e
n
d
en
ci
es
an
d
ap
p
r
o
x
im
ates
f
u
n
ctio
n
al
d
ep
en
d
e
n
cies
f
r
o
m
d
ata,
p
r
o
v
id
in
g
a
r
ich
er
s
em
an
tic
lay
er
th
at
co
m
p
lem
e
n
ts
th
e
s
tr
u
ctu
r
al
in
f
o
r
m
atio
n
o
f
th
e
C
FGs
.
T
h
is
s
tep
is
cr
u
cial
f
o
r
id
en
tify
in
g
“d
ee
p
”
v
u
ln
er
ab
ilit
ies th
at
ar
is
e
f
r
o
m
c
o
m
p
lex
d
ata
in
ter
ac
tio
n
s
r
ath
er
th
an
s
im
p
le
co
n
tr
o
l f
lo
w
er
r
o
r
s
.
3
.
2
.
G
ra
ph
co
ns
t
ruct
io
n
T
h
e
r
aw
b
in
ar
y
p
r
o
g
r
a
m
is
t
r
an
s
f
o
r
m
ed
in
to
a
g
r
ap
h
-
s
tr
u
ctu
r
ed
r
e
p
r
esen
tatio
n
s
u
itab
le
f
o
r
d
ee
p
lear
n
in
g
.
D
is
ass
em
b
ly
to
o
ls
(
e
.
g
.
,
I
DA
Pro
)
ar
e
u
tili
ze
d
to
p
ar
s
e
th
e
b
in
ar
y
f
u
n
ct
io
n
s
.
T
h
e
g
r
a
p
h
co
n
s
tr
u
ctio
n
p
r
o
ce
ed
s
as
f
o
llo
ws
.
No
d
e
r
ep
r
esen
tatio
n
:
e
ac
h
b
asic
b
l
o
ck
in
th
e
f
u
n
ctio
n
is
tr
ea
ted
as
a
n
o
d
e.
T
o
ca
p
tu
r
e
th
e
lo
ca
l
s
em
an
tics
,
a
m
u
lti
-
d
im
e
n
s
io
n
al
f
ea
tu
r
e
v
ec
t
o
r
is
e
x
tr
ac
ted
f
o
r
ea
ch
n
o
d
e,
co
m
p
r
is
in
g
th
e
n
u
m
b
er
o
f
in
s
tr
u
ctio
n
s
,
ar
ith
m
etic
o
p
er
at
io
n
s
,
m
em
o
r
y
o
p
er
atio
n
s
,
an
d
th
e
d
is
tr
ib
u
tio
n
o
f
o
p
co
d
es.
E
d
g
e
estab
lis
h
m
en
t:
ed
g
es
ar
e
i
n
itially
estab
lis
h
ed
b
ased
o
n
th
e
c
o
n
tr
o
l
f
l
o
w
ju
m
p
s
b
etwe
en
b
asic
b
lo
ck
s
.
Dep
en
d
en
cy
e
n
r
ich
m
en
t:
t
h
e
d
e
p
en
d
e
n
cies
ex
tr
ac
ted
b
y
T
ANE
in
s
u
b
-
s
ec
tio
n
3
.
1
ar
e
s
u
p
e
r
im
p
o
s
ed
o
n
to
th
e
g
r
a
p
h
.
I
f
a
f
u
n
ctio
n
al
d
ep
e
n
d
en
c
y
e
x
is
ts
b
etwe
en
b
asic
b
lo
ck
s
A
an
d
B
,
an
ad
d
itio
n
al
ed
g
e
(
o
r
weig
h
ted
ed
g
e)
is
a
d
d
ed
.
T
h
e
r
esu
ltin
g
s
tr
u
ctu
r
e
is
an
AC
FG,
wh
ich
s
er
v
es
as
a
co
m
p
r
eh
e
n
s
iv
e
r
e
p
r
esen
tatio
n
o
f
t
h
e
f
u
n
ctio
n
’
s
ex
ec
u
tio
n
lo
g
ic
an
d
s
em
an
tic
co
n
tex
t3
.
3
.
3
.
P
o
o
lin
g
m
e
cha
nis
m
:
diff
us
io
n a
t
t
ent
io
n str
a
t
eg
y
Pre
s
er
v
in
g
to
p
o
lo
g
ical
i
n
f
o
r
m
atio
n
r
ele
v
an
t
to
v
u
l
n
er
ab
ilit
ies.
T
r
ad
itio
n
al
g
lo
b
al
p
o
o
lin
g
m
eth
o
d
s
(
e.
g
.
,
s
u
m
m
in
g
all
n
o
d
e
f
ea
t
u
r
es)
o
f
ten
d
is
ca
r
d
cr
itical
lo
ca
l
s
tr
u
ctu
r
al
cu
es.
T
o
ad
d
r
ess
th
is
,
th
e
T
ANE
-
Po
o
l la
y
er
is
in
tr
o
d
u
ce
d
as sh
o
wn
in
Fig
u
r
e
1
,
wh
ich
em
p
lo
y
s
a
h
ier
ar
ch
ical
p
o
o
lin
g
s
tr
ateg
y
b
ased
o
n
d
if
f
u
s
io
n
atten
tio
n
.
i)
Atten
tio
n
s
co
r
e
ca
lcu
latio
n
:
f
ir
s
t,
th
e
r
elev
an
ce
b
etwe
en
c
o
n
n
ec
ted
n
o
d
es
is
co
m
p
u
ted
.
T
h
e
atten
tio
n
s
co
r
e
,
(
)
f
o
r
a
n
ed
g
e
b
etwe
en
t
h
e
n
o
d
es
an
d
at
lay
er
l
is
ca
lcu
la
ted
as (
1
)
.
,
(
)
=
(
(
)
,
(
(
)
ℎ
(
)
|
|
(
)
ℎ
(
)
)
)
(
1
)
W
h
er
e
|
|
d
en
o
tes
v
ec
to
r
co
n
ca
t
en
atio
n
,
an
d
(
∙
)
is
th
e
L
ea
k
y
R
eL
U
ac
tiv
atio
n
f
u
n
ctio
n
.
T
h
is
m
ec
h
an
is
m
allo
ws
th
e
m
o
d
el
to
d
y
n
am
ic
ally
weig
h
th
e
im
p
o
r
ta
n
ce
o
f
im
m
ed
iate
n
eig
h
b
o
r
s
b
ased
o
n
th
eir
f
ea
tu
r
e
co
m
p
atib
ilit
y
.
ii)
Dif
f
u
s
io
n
atten
tio
n
f
o
r
lo
n
g
-
r
an
g
e
d
ep
e
n
d
en
cies:
ju
s
tific
atio
n
:
v
u
ln
er
a
b
ilit
ies
o
f
ten
in
v
o
l
v
e
in
ter
ac
tio
n
s
b
etwe
en
b
asic
b
lo
c
k
s
th
at
ar
e
d
is
tan
t
in
th
e
co
n
tr
o
l
f
lo
w
(
e
.
g
.
,
a
m
em
o
r
y
allo
ca
tio
n
an
d
a
s
u
b
s
eq
u
en
t
d
is
tan
t
f
r
ee
)
.
Stan
d
ar
d
g
r
ap
h
co
n
v
o
l
u
tio
n
o
n
l
y
ag
g
r
e
g
ates
lo
ca
l
n
eig
h
b
o
r
s
.
T
o
ca
p
tu
r
e
th
ese
lo
n
g
-
r
an
g
e
d
ep
en
d
e
n
cies,
we
co
m
p
u
te
a
d
if
f
u
s
io
n
atten
tio
n
s
co
r
e
m
atr
ix
(
)
is
co
m
p
u
ted
as in
(
2
)
.
=
∑
=
0
(
2
)
Her
e
r
ep
r
esen
ts
th
e
tr
an
s
itio
n
m
atr
ix
o
f
i
s
tep
s
,
an
d
θ
i
as
a
d
ec
a
y
f
ac
to
r
(
>
+
1
)
.
T
h
is
f
o
r
m
u
latio
n
ex
p
an
d
s
t
h
e
r
ec
e
p
tiv
e
f
ield
,
al
lo
win
g
th
e
m
o
d
el
to
"see"
r
elatio
n
s
h
ip
s
u
p
to
K
Ho
p
s
awa
y
,
e
n
s
u
r
in
g
t
h
at
d
is
tan
t b
u
t c
au
s
ally
r
elate
d
c
o
d
e
b
lo
ck
s
in
f
lu
en
ce
th
e
n
o
d
e
i
m
p
o
r
tan
ce
s
co
r
e.
iii)
Gr
ap
h
co
a
r
s
en
in
g
th
e
p
o
o
lin
g
lay
er
f
in
ally
a
g
g
r
e
g
ates
n
o
d
e
f
ea
tu
r
es
u
s
in
g
:
a
to
p
-
k
s
elec
tio
n
m
eth
o
d
b
ased
o
n
th
e
ag
g
r
eg
ated
im
p
o
r
tan
ce
s
co
r
es
is
th
en
em
p
lo
y
e
d
to
r
etain
o
n
l
y
th
e
m
o
s
t
cr
itical
n
o
d
es
(
b
asic
b
lo
ck
s
m
o
s
t
lik
ely
t
o
c
o
n
tai
n
d
e
f
ec
ts
)
,
f
o
r
m
in
g
a
co
ar
s
e
n
ed
s
u
b
g
r
ap
h
.
T
h
is
h
ier
a
r
ch
i
ca
l
ap
p
r
o
ac
h
p
r
ese
r
v
es th
e
m
o
s
t salien
t stru
ctu
r
al
f
ea
tu
r
es wh
ile
f
ilter
in
g
o
u
t n
o
is
e,
as d
ef
i
n
ed
in
(
3
)
.
(
,
(
)
;
)
=
(
)
(
3
)
3
.
4
.
P
re
dict
io
n
m
o
del a
rc
hit
ec
t
ure
T
h
e
o
v
er
all
ar
ch
itectu
r
e,
as sh
o
wn
in
Fig
u
r
e
2
,
ad
o
p
ts
a
s
tack
ed
g
r
ap
h
co
n
v
o
lu
tio
n
al
n
etw
o
r
k
(
GC
N)
b
ac
k
b
o
n
e
in
ter
leav
e
d
with
T
A
NE
-
Po
o
l la
y
er
s
.
i)
Featu
r
e
ex
tr
ac
tio
n
:
th
e
AC
FG
p
ass
es
th
r
o
u
g
h
th
r
ee
co
n
s
ec
u
tiv
e
b
lo
ck
s
.
E
ac
h
b
lo
ck
co
n
s
i
s
ts
o
f
a
GC
N
lay
er
f
o
llo
wed
b
y
a
T
ANE
-
P
o
o
l
lay
er
.
T
h
e
GC
N
r
ef
in
es
n
o
d
e
f
ea
tu
r
es,
wh
ile
T
ANE
-
Po
o
l
r
ed
u
ce
s
th
e
g
r
ap
h
s
ize.
ii)
R
ea
d
o
u
t
p
h
ase:
af
ter
ea
c
h
p
o
o
lin
g
s
tep
,
a
r
ea
d
o
u
t
f
u
n
ctio
n
ag
g
r
e
g
ates
th
e
f
ea
tu
r
es
o
f
th
e
co
ar
s
en
e
d
s
u
b
g
r
ap
h
(
u
s
in
g
s
u
m
p
o
o
lin
g
)
to
g
en
er
ate
a
g
r
ap
h
-
lev
el
v
ec
to
r
.
iii)
Fin
al
p
r
ed
ictio
n
:
th
e
g
r
a
p
h
-
le
v
el
v
ec
to
r
s
f
r
o
m
all
th
r
e
e
lay
er
s
ar
e
co
n
ca
ten
ated
to
f
o
r
m
a
m
u
lti
-
s
ca
le
r
ep
r
esen
tatio
n
.
T
h
is
v
ec
to
r
is
f
ed
in
to
a
m
u
lti
-
lay
er
p
er
ce
p
tr
o
n
(
ML
P)
class
if
ier
,
wh
ic
h
o
u
tp
u
ts
th
e
p
r
o
b
a
b
ilit
y
o
f
th
e
f
u
n
ctio
n
b
ei
n
g
v
u
l
n
er
ab
le.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ad
v
Ap
p
l Sci
I
SS
N:
2252
-
8
8
1
4
P
r
o
g
r
a
m
d
efec
t p
r
ed
ictio
n
mo
d
el
b
a
s
ed
o
n
to
p
o
lo
g
y
a
w
a
r
e
n
o
d
e
ev
a
lu
a
tio
n
p
o
o
l g
r
a
p
h
…
(
Da
n
Li
)
587
Fig
u
r
e
1
.
Deta
iled
s
tr
u
ctu
r
e
o
f
th
e
T
ANE
-
Po
o
l
lay
er
in
co
r
p
o
r
atin
g
d
if
f
u
s
io
n
atten
tio
n
m
ec
h
an
is
m
Fig
u
r
e
2
.
Ar
c
h
itectu
r
e
o
f
th
e
T
ANE
-
Po
o
l
p
r
o
g
r
am
d
e
f
ec
t p
r
e
d
ictio
n
m
o
d
el:
f
r
o
m
AC
FG to
p
r
o
b
a
b
ilit
y
o
u
tp
u
t
3
.
5
.
I
nte
g
ra
t
io
n into
f
uzzing
:
prio
rit
ized
s
chedu
lin
g
T
h
e
u
ltima
te
g
o
al
o
f
o
u
r
m
o
d
el
is
to
g
u
id
e
th
e
f
u
zz
er
.
T
h
e
m
o
d
el’
s
o
u
tp
u
t
is
tr
an
s
lated
i
n
to
a
s
tatic
v
u
ln
er
ab
ilit
y
s
co
r
e
(
SVS)
to
p
r
io
r
itize
test
ca
s
es
as
il
lu
s
tr
ated
in
Fig
u
r
e
3
.
J
u
s
tific
atio
n
:
s
tan
d
ar
d
f
u
zz
er
s
(
lik
e
AFL)
tr
ea
t
all
p
ath
s
eq
u
ally
o
r
f
o
cu
s
s
o
lely
o
n
co
v
er
a
g
e.
B
y
in
teg
r
atin
g
SVS,
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
(
en
er
g
y
)
ar
e
d
ir
ec
ted
to
war
d
p
ath
s
th
at
tr
av
e
r
s
e
h
ig
h
-
r
is
k
b
asic
b
lo
ck
s
.
Fo
r
a
b
in
ar
y
f
u
n
ctio
n
with
v
u
ln
er
ab
ilit
y
p
r
o
b
ab
ilit
y
,
th
e
SVS o
f
its
co
n
s
titu
en
t b
asic b
lo
ck
s
(
)
is
d
er
iv
ed
as (
4
)
.
(
)
=
∗
+
(
4
)
W
h
er
e
=
20
an
d
=
0
.
1
ar
e
s
ca
lin
g
co
n
s
ta
n
ts
d
eter
m
in
ed
em
p
ir
ically
to
d
if
f
er
en
tiate
r
is
k
lev
els.
Du
r
in
g
f
u
zz
in
g
,
f
o
r
a
test
ca
s
e
t
ex
ec
u
tin
g
a
p
ath
ℎ
,
it
i
s
Fi
tn
ess
Sco
r
e
is
th
e
s
u
m
o
f
th
e
SVS o
f
all
v
is
ited
b
lo
ck
s
as in
(
5
)
.
=
∑
∈
ℎ
(
)
(
5
)
Fin
ally
,
m
u
tatio
n
en
er
g
y
is
allo
ca
ted
d
y
n
am
ically
.
T
est
ca
s
es
with
ab
o
v
e
th
e
p
o
p
u
latio
n
av
er
ag
e
(
)
ar
e
ass
ig
n
ed
d
o
u
b
le
th
e
s
tan
d
ar
d
en
er
g
y
2
(
)
,
ef
f
ec
tiv
ely
f
o
cu
s
i
n
g
th
e
f
u
zz
er
’
s
m
u
tatio
n
ef
f
o
r
t
s
o
n
th
e
m
o
s
t
s
u
s
p
icio
u
s
co
d
e
r
e
g
io
n
s
.
I
n
o
r
d
er
t
o
f
u
r
th
er
o
p
tim
ize
t
h
e
m
u
tatio
n
en
er
g
y
s
ch
ed
u
lin
g
,
th
is
p
ap
e
r
ad
d
s
th
e
ad
a
p
tatio
n
s
co
r
e
in
d
icato
r
o
f
th
e
u
s
e
ca
s
es
to
th
e
o
r
ig
in
al
e
n
er
g
y
s
ch
ed
u
lin
g
o
f
AFL.
A
h
ig
h
ad
ap
tatio
n
s
co
r
e
in
d
icate
s
th
at
th
e
u
s
e
ca
s
es h
av
e
a
h
ig
h
p
r
io
r
ity
an
d
th
u
s
s
h
o
u
ld
b
e
g
iv
e
n
m
o
r
e
m
u
tatio
n
en
er
g
y
.
Mu
tatio
n
en
er
g
y
is
th
e
n
u
m
b
er
o
f
m
u
tatio
n
s
in
th
e
HAVO
C
p
h
ase.
HAVO
C
is
a
m
u
tatio
n
p
h
ase
o
f
th
e
f
u
zz
if
ier
r
e
p
r
e
s
en
ted
b
y
AFL,
in
wh
ich
lar
g
e
-
s
ca
le
ch
an
g
es a
r
e
m
ad
e
to
th
e
u
s
e
ca
s
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
8
1
4
I
n
t J Ad
v
Ap
p
l Sci
,
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
2
0
2
6
:
5
8
3
-
593
588
T
h
e
en
er
g
y
s
ch
ed
u
lin
g
f
o
r
m
u
l
a
is
s
h
o
wn
in
(
6
)
.
Fo
r
a
g
iv
en
u
s
e
ca
s
e
,
th
e
en
er
g
y
(
)
ass
ig
n
ed
to
it c
an
b
e
ca
lcu
lated
f
r
o
m
(
6
)
.
(
)
{
2
(
)
,
(
)
≥
(
)
≤
2
(
)
,
(
)
<
,
ℎ
(
6
)
W
h
er
e
(
)
is
th
e
m
u
tatio
n
e
n
er
g
y
ass
ig
n
ed
t
o
th
e
u
s
e
ca
s
e
b
y
tr
ad
itio
n
al
g
r
ay
-
b
o
x
f
u
zz
y
test
in
g
to
o
ls
(
in
p
a
r
ticu
lar
,
AFL)
,
an
d
is
th
e
m
ax
im
u
m
e
n
er
g
y
th
at
ca
n
b
e
ass
ig
n
ed
b
y
AFL.
(
)
is
th
e
ad
ap
tatio
n
s
co
r
e
co
r
r
esp
o
n
d
i
n
g
t
o
th
e
u
s
e
ca
s
e
,
an
d
is
th
e
av
er
a
g
e
o
f
th
e
a
d
ap
tatio
n
s
co
r
es
o
f
all
cu
r
r
en
t
u
s
e
ca
s
es.
W
h
en
th
e
f
itn
ess
s
co
r
e
(
)
o
f
th
e
u
s
e
ca
s
e
is
g
r
ea
ter
th
an
o
r
eq
u
al
to
an
d
th
e
r
aw
en
er
g
y
(
)
is
less
th
an
o
r
eq
u
al
to
h
alf
o
f
,
2
(
)
o
f
en
e
r
g
y
is
ass
ig
n
e
d
to
th
e
u
s
e
ca
s
e.
W
h
en
th
e
f
itn
es
s
s
co
r
e
(
)
o
f
th
e
u
s
e
ca
s
e
is
les
s
th
an
,
th
e
en
er
g
y
o
f
(
)
is
as
s
ig
n
ed
.
I
f
it
is
n
o
t
eith
er
o
f
th
e
ab
o
v
e,
th
en
allo
ca
te
th
e
en
er
g
y
f
o
r
th
e
u
s
e
ca
s
e
.
E
n
er
g
y
s
ch
ed
u
lin
g
b
ased
o
n
te
s
t
ca
s
e
p
r
io
r
itizatio
n
is
d
esig
n
ed
to
g
iv
e
m
o
r
e
en
er
g
y
to
th
e
ca
s
es
with
a
h
ig
h
er
p
r
io
r
ity
lev
el,
wh
e
r
e
th
e
p
r
io
r
ity
lev
el
is
r
ef
lecte
d
b
y
th
e
f
itn
ess
s
co
r
es
o
f
th
e
test
ca
s
es
.
B
y
u
s
in
g
th
e
av
er
ag
e
f
itn
ess
s
co
r
e
o
f
th
e
cu
r
r
en
t
test
ca
s
es
as
a
b
aselin
e,
th
e
ca
s
es
with
f
it
n
ess
s
co
r
es
h
ig
h
e
r
th
a
n
t
h
e
b
aselin
e
ar
e
g
iv
en
h
ig
h
er
p
r
io
r
ity
an
d
ar
e
g
iv
e
n
m
o
r
e
en
er
g
y
,
an
d
th
e
ca
s
es
lo
wer
th
an
th
e
b
aselin
e
u
s
e
th
e
AFL
’
s
o
r
ig
in
al
allo
ca
tio
n
o
f
e
n
er
g
y
.
Mo
r
e
m
u
tatio
n
e
n
er
g
y
is
ass
ig
n
ed
to
u
s
e
ca
s
es
with
h
ig
h
er
f
itn
ess
s
co
r
es
to
im
p
r
o
v
e
th
e
ef
f
icien
cy
o
f
m
u
tatio
n
f
o
r
f
u
zz
y
test
ca
s
es.
F
i
g
u
r
e
3
.
S
c
h
e
m
at
i
c
o
f
t
h
e
te
s
t
c
a
s
e
p
r
i
o
r
it
i
z
a
ti
o
n
a
n
d
e
n
e
r
g
y
s
c
h
e
d
u
l
i
n
g
f
r
a
m
e
w
o
r
k
g
u
i
d
e
d
b
y
d
e
f
e
c
t
p
r
e
d
i
c
t
i
o
n
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
4
.
1
.
E
x
perim
ent
a
l
s
et
up
a
nd
da
t
a
s
et
s
T
o
co
m
p
r
e
h
en
s
iv
ely
v
alid
ate
th
e
T
ANE
-
Po
o
l
f
r
am
ewo
r
k
,
two
d
is
tin
ct
d
ataset
s
wer
e
u
tili
ze
d
,
as
s
h
o
wn
in
T
a
b
le
1
:
th
e
J
u
liet
test
s
u
ite
v
1
.
3
an
d
a
cu
r
ated
r
ea
l
-
wo
r
l
d
p
r
o
g
r
am
s
d
ataset.
T
h
e
J
u
liet
s
u
ite
p
r
o
v
id
es
a
co
n
tr
o
lled
en
v
i
r
o
n
m
en
t
with
lab
eled
m
em
o
r
y
-
r
elate
d
wea
k
n
ess
es
(
e.
g
.
,
b
u
f
f
er
o
v
er
f
lo
ws),
allo
win
g
f
o
r
p
r
ec
is
e
b
aselin
e
co
m
p
ar
is
o
n
s
.
T
o
ev
alu
ate
g
en
er
aliza
tio
n
ca
p
ab
ilit
y
,
t
h
e
r
ea
l
-
wo
r
ld
d
ataset
was
co
n
s
tr
u
cted
f
r
o
m
o
p
e
n
-
s
o
u
r
ce
p
r
o
jects
(
v
ia
GitHu
b
)
an
d
k
n
o
wn
ex
p
lo
its
(
v
ia
E
x
p
lo
it
-
DB
)
,
p
r
o
ce
s
s
ed
u
s
in
g
I
DA
Pro
f
o
r
b
in
ar
y
lab
elin
g
.
T
ANE
-
Po
o
l
was
b
en
ch
m
ar
k
e
d
ag
ain
s
t
two
ca
teg
o
r
ies
o
f
s
ta
te
-
of
-
th
e
-
a
r
t
m
o
d
els:
i)
s
tan
d
ar
d
GNNs:
GC
N,
Gr
ap
h
SAGE
,
an
d
g
r
a
p
h
atten
ti
o
n
n
etwo
r
k
(
GAT
)
(
r
ep
r
esen
tin
g
g
r
ap
h
lear
n
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ad
v
Ap
p
l Sci
I
SS
N:
2252
-
8
8
1
4
P
r
o
g
r
a
m
d
efec
t p
r
ed
ictio
n
mo
d
el
b
a
s
ed
o
n
to
p
o
lo
g
y
a
w
a
r
e
n
o
d
e
ev
a
lu
a
tio
n
p
o
o
l g
r
a
p
h
…
(
Da
n
Li
)
589
with
o
u
t
h
ier
a
r
ch
ical
p
o
o
lin
g
)
an
d
ii)
p
o
o
lin
g
-
b
ased
m
eth
o
d
s
:
Dif
f
Po
o
l
an
d
SAGPo
o
l
(
r
e
p
r
esen
tin
g
ad
v
a
n
ce
d
to
p
o
lo
g
y
-
awa
r
e
r
e
d
u
ctio
n
)
.
T
ab
le
1
.
T
r
ai
n
in
g
a
n
d
test
in
g
d
ataset
s
eg
m
en
tatio
n
D
a
t
a
set
C
a
t
e
g
o
r
y
#
V
u
l
n
e
r
a
b
l
e
sam
p
l
e
s
#
N
o
r
m
a
l
s
a
m
p
l
e
s
T
o
t
a
l
Ju
l
i
e
t
T
e
st
S
u
i
t
e
v
1
.
3
Tr
a
i
n
i
n
g
2
0
,
0
0
0
2
0
,
0
0
0
4
0
,
0
0
0
Te
st
i
n
g
2
,
0
0
0
2
,
0
0
0
4
,
0
0
0
R
e
a
l
-
w
o
r
l
d
P
r
o
g
r
a
ms
Tr
a
i
n
i
n
g
2
0
,
0
0
0
2
0
,
0
0
0
4
0
,
0
0
0
Te
st
i
n
g
2
,
0
0
0
2
,
0
0
0
4
,
0
0
0
4
.2
.
P
er
f
o
r
m
a
nce
a
na
ly
s
is
a
nd
co
m
pa
riso
n wit
h lit
er
a
t
ure
T
h
e
co
m
p
ar
ativ
e
ac
cu
r
ac
y
r
esu
lts
ar
e
p
r
esen
ted
in
Fig
u
r
e
4
.
T
ANE
-
Po
o
l
co
n
s
is
ten
tly
o
u
tp
er
f
o
r
m
e
d
all
b
aselin
es
o
n
b
o
th
d
atasets
.
Sp
ec
if
ically
,
o
n
th
e
co
m
p
lex
R
ea
l
-
wo
r
ld
d
ataset,
o
u
r
m
o
d
el
ac
h
iev
ed
a
3
.
2
%
ac
cu
r
ac
y
g
ai
n
o
v
e
r
th
e
s
tr
o
n
g
e
s
t b
aselin
e,
SAGPo
o
l.
F
i
g
u
r
e
4
.
C
o
m
p
a
r
is
o
n
o
f
t
h
e
a
cc
u
r
a
c
y
o
f
d
i
f
f
e
r
e
n
t
m
o
d
e
l
s
o
n
J
u
l
i
e
t
t
es
t
s
u
i
te
a
n
d
r
e
a
l
-
w
o
r
l
d
p
r
o
g
r
a
m
s
d
a
t
a
s
et
s
C
r
itical
d
is
cu
s
s
io
n
o
n
g
r
ap
h
t
o
p
o
lo
g
y
:
th
e
r
esu
lts
h
ig
h
lig
h
t
a
s
ig
n
if
ican
t
p
er
f
o
r
m
an
ce
g
a
p
b
etwe
en
p
o
o
lin
g
-
b
ased
m
eth
o
d
s
(
Dif
f
P
o
o
l,
SAGPo
o
l,
a
n
d
T
ANE
-
Po
o
l)
an
d
f
lat
GNNs
(
GC
N
an
d
GAT
)
.
Ob
s
er
v
atio
n
:
GC
N
an
d
GAT
s
tr
u
g
g
led
to
g
en
er
alize
o
n
t
h
e
r
ea
l
-
wo
r
l
d
d
ataset.
L
iter
atu
r
e
v
alid
atio
n
:
th
is
f
in
d
in
g
alig
n
s
with
th
e
wo
r
k
o
f
Z
h
o
u
et
a
l
.
[
2
6
]
,
wh
o
d
em
o
n
s
tr
ated
t
h
at
GNN
m
o
d
els
ca
p
tu
r
in
g
p
r
o
g
r
a
m
s
em
an
tics
s
ig
n
if
ican
tly
o
u
tp
e
r
f
o
r
m
f
lat
(
n
o
n
-
g
r
ap
h
)
b
aselin
es
f
o
r
v
u
ln
er
ab
ilit
y
d
etec
tio
n
,
co
n
f
ir
m
i
n
g
th
at
h
ie
r
ar
ch
ical
s
tr
u
ctu
r
al
in
f
o
r
m
atio
n
is
ess
en
tial.
T
h
e
co
n
tr
i
b
u
tio
n
:
B
y
i
n
co
r
p
o
r
atin
g
th
e
T
ANE
-
Po
o
l
lay
er
,
o
u
r
m
o
d
el
ef
f
ec
tiv
ely
p
r
eser
v
es
th
e
lo
ca
l
s
u
b
s
tr
u
ctu
r
es
o
f
d
e
f
ec
ts
.
W
h
ile
Z
h
o
u
et
a
l
.
[
2
6
]
d
em
o
n
s
tr
ated
th
at
g
r
ap
h
-
b
ased
p
r
o
g
r
a
m
r
ep
r
esen
tatio
n
s
im
p
r
o
v
e
v
u
ln
er
ab
ilit
y
d
etec
tio
n
,
th
e
r
esu
lts
ex
ten
d
th
is
b
y
s
h
o
win
g
th
at
co
m
b
in
in
g
d
ep
en
d
e
n
cy
m
in
in
g
(
T
ANE
)
with
d
if
f
u
s
io
n
atten
tio
n
y
ie
ld
s
s
u
p
er
io
r
d
ef
ec
t
l
o
ca
lizat
io
n
co
m
p
ar
e
d
with
to
p
o
lo
g
y
lear
n
in
g
alo
n
e.
W
h
y
T
ANE
-
Po
o
l
b
ea
ts
SAG
Po
o
l:
alth
o
u
g
h
SAGPo
o
l
is
ef
f
ec
tiv
e,
it
r
elies
s
o
lely
o
n
s
elf
-
atten
tio
n
f
o
r
n
o
d
e
d
r
o
p
p
i
n
g
.
I
n
co
n
tr
ast,
T
ANE
-
Po
o
l
l
ev
er
ag
es
d
ata
d
ep
en
d
en
cies
e
x
tr
ac
ted
v
ia
T
ANE
.
J
u
s
tific
atio
n
:
a
v
u
ln
er
ab
ili
ty
o
f
ten
s
p
an
s
s
tr
u
ctu
r
ally
d
is
tan
t
b
u
t
d
ata
-
d
e
p
en
d
e
n
t
b
asic
b
lo
c
k
s
.
Stan
d
ar
d
g
r
ap
h
p
o
o
lin
g
m
ig
h
t
d
r
o
p
a
"c
o
n
n
e
cto
r
"
n
o
d
e
th
at
s
ee
m
s
to
p
o
lo
g
ically
u
n
im
p
o
r
tan
t
b
u
t
ca
r
r
ies
cr
itical
d
ata
f
lo
w.
T
ANE
-
Po
o
l
’
s
d
if
f
u
s
io
n
m
ec
h
a
n
is
m
en
s
u
r
es th
ese
laten
t d
ep
e
n
d
en
cies
ar
e
p
r
eser
v
ed
.
4
.
3
.
T
o
p
-
k
a
cc
ura
cy
a
nd
pra
ct
ica
l im
pli
ca
t
io
ns
Fo
r
p
r
ac
tical
d
ep
l
o
y
m
en
t
i
n
C
I
/C
D
p
ip
elin
es,
th
e
"
T
o
p
-
k
ac
cu
r
ac
y
"
(
th
e
p
r
o
p
o
r
tio
n
o
f
ac
tu
al
v
u
ln
er
ab
ilit
ies
f
o
u
n
d
with
in
t
h
e
to
p
-
k
r
an
k
e
d
ca
n
d
id
ates)
is
a
m
o
r
e
cr
itical
m
etr
ic
th
an
o
v
er
all
class
if
icatio
n
ac
cu
r
ac
y
.
Hig
h
to
p
-
k
p
e
r
f
o
r
m
an
ce
im
p
lies
th
at
f
u
zz
er
wastes
less
en
er
g
y
o
n
f
alse
p
o
s
itiv
es.
T
ab
les
2
an
d
3
p
r
esen
t
th
e
to
p
-
k
an
aly
s
is
.
T
ANE
-
Po
o
l
s
h
o
ws
a
r
em
ar
k
ab
le
ad
v
an
tag
e
in
t
h
e
"
ea
r
ly
d
is
co
v
er
y
"
p
h
ase.
R
esu
lt:
o
n
th
e
r
ea
l
-
wo
r
ld
d
ataset,
a
s
s
h
o
wn
in
T
ab
le
3
,
T
ANE
-
Po
o
l
im
p
r
o
v
e
d
ac
cu
r
ac
y
b
y
1
1
.
1
%
at
t
op
-
8
0
0
co
m
p
ar
ed
to
SAGPo
o
l.
0
.8
6
0
.
8
8
0
.
9
0
.
9
2
0
.
9
4
0
.
9
6
0
.
9
8
1
G
C
N
G
r
a
p
h
S
A
G
E
GAT
D
i
f
f
P
o
o
l
S
A
G
P
o
o
l
TA
N
E-
P
o
o
l
A
c
c
u
r
a
c
y
M
o
d
e
l
Ju
l
i
e
t
Te
st
S
u
i
t
e
R
e
a
l
-
w
o
r
l
d
P
r
o
g
r
a
ms
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
8
1
4
I
n
t J Ad
v
Ap
p
l Sci
,
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
2
0
2
6
:
5
8
3
-
593
590
T
ab
le
2
.
C
o
m
p
a
r
is
o
n
o
f
to
p
-
k
ac
cu
r
ac
y
o
f
d
if
f
er
en
t m
o
d
els o
n
J
u
liet
test
s
u
ite
d
ataset
M
o
d
e
l
To
p
-
2
0
0
a
c
c
u
r
a
c
y
To
p
-
4
0
0
a
c
c
u
r
a
c
y
To
p
-
6
0
0
a
c
c
u
r
a
c
y
To
p
-
8
0
0
a
c
c
u
r
a
c
y
To
p
-
1
,
0
0
0
a
c
c
u
r
a
c
y
To
p
-
1
,
2
0
0
a
c
c
u
r
a
c
y
G
C
N
0
.
9
1
4
0
.
8
7
2
0
.
8
4
2
0
.
8
3
1
0
.
7
8
1
0
.
7
8
6
G
r
a
p
h
S
A
G
E
0
.
9
2
5
0
.
9
1
8
0
.
9
0
5
0
.
8
7
4
0
.
8
5
2
0
.
7
9
4
GAT
0
.
9
2
7
0
.
9
1
9
0
.
9
1
2
0
.
8
9
6
0
.
8
7
9
0
.
8
0
7
D
i
f
f
P
o
o
l
0
.
9
3
4
0
.
9
2
8
0
.
9
2
4
0
.
9
0
9
0
.
8
9
3
0
.
8
5
6
S
A
G
P
o
o
l
0
.
9
4
8
0
.
9
3
6
0
.
9
3
4
0
.
9
2
4
0
.
9
0
8
0
.
8
7
4
TA
N
E
-
P
o
o
l
0
.
9
6
8
0
.
9
4
5
0
.
9
3
8
0
.
9
2
5
0
.
9
1
7
0
.
8
9
1
T
ab
le
3
.
C
o
m
p
a
r
is
o
n
o
f
to
p
-
k
ac
cu
r
ac
y
o
f
d
if
f
er
en
t m
o
d
els o
n
R
ea
l
-
wo
r
ld
p
r
o
g
r
am
s
d
atase
t
M
o
d
e
l
To
p
-
2
0
0
a
c
c
u
r
a
c
y
To
p
-
4
0
0
a
c
c
u
r
a
c
y
To
p
-
6
0
0
a
c
c
u
r
a
c
y
To
p
-
8
0
0
a
c
c
u
r
a
c
y
To
p
-
1
,
0
0
0
a
c
c
u
r
a
c
y
To
p
-
1
,
2
0
0
a
c
c
u
r
a
c
y
G
C
N
0
.
9
2
6
0
.
8
8
1
0
.
8
3
5
0
.
7
8
3
0
.
7
8
9
0
.
7
5
4
G
r
a
p
h
S
A
G
E
0
.
9
2
8
0
.
8
9
1
0
.
8
3
6
0
.
7
8
5
0
.
7
9
1
0
.
7
7
1
GAT
0
.
9
2
9
0
.
8
9
3
0
.
8
4
5
0
.
8
0
1
0
.
7
9
2
0
.
7
8
1
D
i
f
f
P
o
o
l
0
.
9
3
1
0
.
9
0
4
0
.
9
5
5
0
.
8
1
2
0
.
7
9
6
0
.
7
9
5
S
A
G
P
o
o
l
0
.
9
3
8
0
.
9
1
4
0
.
8
6
4
0
.
8
2
3
0
.
8
0
5
0
.
8
0
7
TA
N
E
-
P
o
o
l
0
.
9
6
7
0
.
9
3
4
0
.
9
2
9
0
.
9
1
5
0
.
8
5
4
0
.
8
2
8
Pra
ctica
l
r
elev
an
ce
:
as
h
ig
h
lig
h
ted
in
th
e
2
0
2
5
s
u
r
v
ey
b
y
Qiu
et
a
l.
[
4
]
,
th
e
p
r
im
ar
y
b
o
ttlen
ec
k
i
n
g
r
ey
b
o
x
f
u
zz
in
g
is
th
e
"sch
ed
u
lin
g
o
f
in
e
f
f
ec
tiv
e
s
ee
d
s
,
"
wh
ich
d
ilu
tes
th
e
test
in
g
b
u
d
g
et.
Valid
atio
n
:
th
e
h
ig
h
to
p
-
k
s
co
r
es
d
ir
ec
tly
ad
d
r
ess
th
is
b
o
ttlen
ec
k
.
B
y
ac
cu
r
at
ely
p
r
io
r
itizin
g
th
e
to
p
2
0
% o
f
s
u
s
p
icio
u
s
b
lo
ck
s
,
T
ANE
-
Po
o
l
allo
ws
th
e
in
teg
r
ated
f
u
zz
er
(
s
u
b
-
s
ec
tio
n
3
.
5
)
to
allo
ca
te
en
er
g
y
w
h
er
e
it
is
s
tatis
t
ically
m
o
s
t
lik
ely
to
y
ield
cr
ash
es.
T
h
is
co
n
f
ir
m
s
th
at
o
u
r
d
ef
ec
t
-
p
r
ed
ic
tio
n
-
g
u
id
e
d
s
tr
a
teg
y
is
n
o
t
ju
s
t
th
eo
r
etica
lly
s
o
u
n
d
b
u
t o
p
e
r
atio
n
ally
v
ia
b
le
f
o
r
lar
g
e
-
s
ca
le
s
o
f
twar
e
r
ep
o
s
ito
r
ies.
4
.
4
.
P
a
ra
m
et
er
s
ens
it
iv
it
y
a
na
ly
s
is
T
h
e
im
p
ac
t
o
f
t
h
e
d
if
f
u
s
io
n
d
is
tan
ce
(
)
o
n
m
o
d
el
p
e
r
f
o
r
m
an
c
e
was
f
u
r
th
er
an
aly
ze
d
as
s
h
o
wn
in
Fig
u
r
e
5
.
T
h
e
ac
cu
r
ac
y
p
ea
k
s
at
=
4
f
o
r
b
o
th
d
atasets
.
I
n
ter
p
r
etatio
n
:
a
d
is
tan
ce
o
f
<
4
ca
p
tu
r
es
in
s
u
f
f
icien
t
co
n
tex
t
(
m
is
s
in
g
lo
n
g
-
r
a
n
g
e
d
ep
e
n
d
en
cies)
,
wh
i
le
>
5
.
I
t
in
t
r
o
d
u
ce
s
n
o
is
e
f
r
o
m
ir
r
elev
an
t
co
d
e
r
eg
io
n
s
.
T
h
is
"sweet
s
p
o
t"
at
=
4
s
u
g
g
ests
th
at
m
em
o
r
y
-
r
elate
d
v
u
ln
er
ab
ilit
ies
in
C
/C
++
b
in
ar
ies
ty
p
ically
m
an
if
est
with
in
a
m
e
d
iu
m
-
r
a
n
g
e
lo
ca
lity
o
f
4
h
o
p
s
in
th
e
d
e
p
en
d
en
c
y
g
r
ap
h
.
T
h
is
em
p
i
r
ical
f
in
d
in
g
p
r
o
v
id
es
a
r
ef
er
en
ce
f
o
r
f
u
tu
r
e
r
esear
ch
er
s
o
p
tim
izin
g
GNN
d
ep
th
f
o
r
b
in
ar
y
a
n
aly
s
is
.
Fig
u
r
e
5
.
E
f
f
ec
t o
f
d
if
f
e
r
en
t d
i
f
f
u
s
io
n
d
is
tan
ce
s
o
n
th
e
ac
cu
r
ac
y
o
f
p
r
ed
ictio
n
m
o
d
els
5.
CO
NCLU
SI
O
N
T
o
ad
d
r
ess
th
e
in
ef
f
icie
n
cy
o
f
tr
ad
itio
n
al
c
o
v
er
ag
e
-
g
u
i
d
ed
f
u
zz
in
g
,
w
h
ich
o
f
ten
e
x
p
en
d
s
s
ig
n
if
ican
t
r
eso
u
r
ce
s
o
n
lo
w
-
r
is
k
p
ath
s
w
ith
o
u
t
g
u
ar
an
teein
g
v
u
ln
er
ab
il
ity
d
is
co
v
er
y
,
th
is
s
tu
d
y
s
u
cc
ess
f
u
lly
d
ev
elo
p
e
d
an
d
v
alid
ated
T
ANE
-
Po
o
l,
an
in
tellig
en
t
d
ef
ec
t
-
p
r
ed
ictio
n
-
g
u
id
e
d
f
u
zz
in
g
f
r
am
ewo
r
k
.
B
y
f
u
s
in
g
T
ANE
-
b
ased
d
e
p
en
d
e
n
cy
e
x
tr
ac
tio
n
with
a
n
o
v
el
d
if
f
u
s
io
n
a
tten
tio
n
p
o
o
lin
g
m
ec
h
a
n
is
m
,
t
h
e
p
r
o
p
o
s
ed
m
o
d
el
0
.
8
4
0
.
8
6
0
.
8
8
0
.
9
0
.9
2
0
.
9
4
0
.
9
6
0
.
9
8
1
2
3
4
5
6
A
cc
u
r
ac
y
D
i
f
f
u
si
o
n
d
i
st
a
n
c
e
(
h
o
p
s)
Ju
l
i
et
Te
st
S
u
i
t
e
R
ea
l
W
o
r
l
d
Pr
o
g
r
am
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ad
v
Ap
p
l Sci
I
SS
N:
2252
-
8
8
1
4
P
r
o
g
r
a
m
d
efec
t p
r
ed
ictio
n
mo
d
el
b
a
s
ed
o
n
to
p
o
lo
g
y
a
w
a
r
e
n
o
d
e
ev
a
lu
a
tio
n
p
o
o
l g
r
a
p
h
…
(
Da
n
Li
)
591
ef
f
ec
tiv
ely
ca
p
tu
r
es
laten
t
s
tr
u
ctu
r
al
d
ep
en
d
en
cies
with
in
b
in
ar
y
p
r
o
g
r
am
s
th
at
a
r
e
ty
p
i
ca
lly
m
is
s
ed
b
y
f
lat
GNNs.
E
x
p
er
im
en
tal
r
esu
lts
o
n
b
o
th
th
e
J
u
liet
test
s
u
ite
an
d
r
ea
l
-
wo
r
ld
r
e
p
o
s
ito
r
ies
d
em
o
n
s
tr
ated
th
at
T
ANE
-
Po
o
l
ac
h
iev
es
s
u
p
er
i
o
r
ac
cu
r
ac
y
co
m
p
ar
ed
to
s
tate
-
of
-
th
e
-
ar
t
b
aselin
es
(
in
clu
d
in
g
SAGPo
o
l
an
d
Dif
f
Po
o
l)
,
co
n
f
ir
m
in
g
th
at
h
i
er
ar
ch
ical
to
p
o
lo
g
y
p
r
eser
v
at
io
n
is
ess
en
tial
f
o
r
p
r
ec
is
e
d
ef
ec
t
lo
ca
lizatio
n
.
Fu
r
th
er
m
o
r
e
,
th
e
in
teg
r
atio
n
o
f
SVS
in
to
th
e
f
u
zz
er
'
s
en
e
r
g
y
s
ch
ed
u
lin
g
s
tr
ateg
y
p
r
o
v
ed
to
b
e
a
cr
itical
p
r
ac
tical
in
n
o
v
atio
n
;
it
en
ab
l
es
th
e
s
y
s
tem
to
p
r
io
r
itize
h
ig
h
-
r
is
k
s
ee
d
s
,
th
er
e
b
y
d
ir
ec
t
ly
ad
d
r
ess
in
g
th
e
b
o
ttlen
ec
k
o
f
in
ef
f
ec
tiv
e
m
u
ta
tio
n
in
co
n
tin
u
o
u
s
in
teg
r
atio
n
en
v
ir
o
n
m
en
ts
.
W
h
ile
th
e
cu
r
r
e
n
t
im
p
lem
en
tatio
n
f
o
cu
s
es
o
n
C
/C
++
b
in
ar
ies,
th
e
f
r
am
ewo
r
k
'
s
m
o
d
u
lar
d
esig
n
estab
lis
h
es
a
s
o
lid
f
o
u
n
d
atio
n
f
o
r
f
u
tu
r
e
s
u
s
tain
ab
le
in
d
u
s
tr
ializatio
n
r
es
ea
r
ch
,
wh
ich
will
aim
to
e
n
h
an
ce
a
d
ap
tab
ilit
y
ac
r
o
s
s
d
iv
er
s
e
p
r
o
g
r
am
m
in
g
lan
g
u
ag
es th
r
o
u
g
h
tr
an
s
f
er
lear
n
in
g
a
n
d
ad
v
an
ce
d
p
r
o
g
r
am
a
n
aly
s
is
tech
n
iq
u
es.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Dan
L
i
✓
✓
✓
✓
✓
✓
✓
✓
✓
Po
h
So
o
n
J
o
s
ep
h
N
g
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Pen
g
Yin
C
h
o
o
✓
✓
✓
✓
✓
✓
Ko
o
Yu
en
Ph
a
n
✓
✓
✓
✓
✓
✓
W
o
n
g
See
W
an
✓
✓
✓
✓
✓
✓
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
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
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
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
I
NF
O
RM
E
D
CO
NS
E
N
T
W
e
h
av
e
o
b
tain
ed
in
f
o
r
m
ed
c
o
n
s
en
t f
r
o
m
all
in
d
iv
id
u
als in
c
lu
d
ed
in
t
h
is
s
tu
d
y
.
DATA AV
AI
L
AB
I
L
I
T
Y
Data
av
ailab
ilit
y
d
o
es n
o
t a
p
p
l
y
to
th
is
p
ap
e
r
as n
o
n
ew
d
ata
wer
e
cr
ea
ted
o
r
an
aly
ze
d
in
th
is
s
tu
d
y
.
RE
F
E
R
E
NC
E
S
[
1
]
F
.
Y
.
A
ssi
r
i
a
n
d
A
.
O
.
A
l
j
a
h
d
a
l
i
,
“
S
o
f
t
w
a
r
e
v
u
l
n
e
r
a
b
i
l
i
t
y
f
u
z
z
t
e
s
t
i
n
g
:
a
mu
t
a
t
i
o
n
-
se
l
e
c
t
i
o
n
o
p
t
i
m
i
z
a
t
i
o
n
sy
st
e
ma
t
i
c
r
e
v
i
e
w
,
”
En
g
i
n
e
e
ri
n
g
,
T
e
c
h
n
o
l
o
g
y
&
A
p
p
l
i
e
d
S
c
i
e
n
c
e
R
e
se
a
rc
h
,
v
o
l
.
1
4
,
n
o
.
4
,
p
p
.
1
4
9
6
1
–
149
6
9
,
A
u
g
.
2
0
2
4
,
d
o
i
:
1
0
.
4
8
0
8
4
/
e
t
a
sr
.
6
9
7
1
.
[
2
]
B
.
A
.
D
a
p
s
h
i
m
a
a
n
d
S
.
K
.
A
h
m
a
d
,
“
Ev
a
l
u
a
t
i
o
n
a
n
d
a
ss
e
ssm
e
n
t
o
f
s
o
f
t
w
a
r
e
se
c
u
r
i
t
y
r
i
sk
s
a
n
d
v
u
l
n
e
r
a
b
i
l
i
t
i
e
s
w
i
t
h
i
n
t
h
e
r
e
a
l
m
o
f
sec
u
r
e
D
e
v
O
p
s,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
F
o
r
Mu
l
t
i
d
i
sc
i
p
l
i
n
a
ry
R
e
se
a
rc
h
,
v
o
l
.
6
,
n
o
.
4
,
J
u
l
.
2
0
2
4
,
d
o
i
:
1
0
.
3
6
9
4
8
/
i
j
f
mr
.
2
0
2
4
.
v
0
6
i
0
4
.
2
5
0
2
6
.
[
3
]
V
.
C
a
s
o
l
a
,
A
.
D
e
B
e
n
e
d
i
c
t
i
s
,
C
.
M
a
z
z
o
c
c
a
,
a
n
d
V
.
O
r
b
i
n
a
t
o
,
“
S
e
c
u
r
e
so
f
t
w
a
r
e
d
e
v
e
l
o
p
m
e
n
t
a
n
d
t
e
s
t
i
n
g
:
a
m
o
d
e
l
-
b
a
s
e
d
met
h
o
d
o
l
o
g
y
,
”
C
o
m
p
u
t
e
rs &
S
e
c
u
ri
t
y
,
v
o
l
.
1
3
7
,
F
e
b
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
se.
2
0
2
3
.
1
0
3
6
3
9
.
[
4
]
J.
Q
i
u
,
Y
.
J
i
a
n
g
,
Y
.
M
i
a
o
,
W
.
Lu
o
,
L.
P
a
n
,
a
n
d
X
.
Z
h
e
n
g
,
“
A
s
u
r
v
e
y
o
f
c
o
v
e
r
a
g
e
-
g
u
i
d
e
d
g
r
e
y
b
o
x
f
u
z
z
i
n
g
w
i
t
h
d
e
e
p
n
e
u
r
a
l
mo
d
e
l
s
,
”
I
n
f
o
rm
a
t
i
o
n
a
n
d
S
o
f
t
w
a
r
e
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
1
8
6
,
O
c
t
.
2
0
2
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
n
f
s
o
f
.
2
0
2
5
.
1
0
7
7
9
7
.
[
5
]
R
.
Q
i
a
n
,
Q
.
Z
h
a
n
g
,
C
.
F
a
n
g
,
a
n
d
L.
G
u
o
,
“
I
n
v
e
st
i
g
a
t
i
n
g
c
o
v
e
r
a
g
e
g
u
i
d
e
d
f
u
z
z
i
n
g
w
i
t
h
m
u
t
a
t
i
o
n
t
e
s
t
i
n
g
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
1
3
t
h
As
i
a
-
P
a
c
i
f
i
c
S
y
m
p
o
si
u
m
o
n
I
n
t
e
r
n
e
t
w
a
re
,
J
u
n
.
2
0
2
2
,
p
p
.
2
7
2
–
2
8
1
,
d
o
i
:
1
0
.
1
1
4
5
/
3
5
4
5
2
5
8
.
3
5
4
5
2
8
5
.
[
6
]
R
.
L
i
a
n
g
e
t
a
l
.
,
“
V
u
l
s
e
y
e
:
d
e
t
e
c
t
sm
a
r
t
c
o
n
t
r
a
c
t
v
u
l
n
e
r
a
b
i
l
i
t
i
e
s
v
i
a
s
t
a
t
e
f
u
l
d
i
r
e
c
t
e
d
g
r
a
y
b
o
x
f
u
z
z
i
n
g
,
”
I
EEE
T
ra
n
s
a
c
t
i
o
n
s
o
n
I
n
f
o
rm
a
t
i
o
n
F
o
re
n
si
c
s
a
n
d
S
e
c
u
ri
t
y
,
v
o
l
.
2
0
,
p
p
.
2
1
5
7
–
2
1
7
0
,
2
0
2
5
,
d
o
i
:
1
0
.
1
1
0
9
/
TI
F
S
.
2
0
2
5
.
3
5
3
7
8
2
7
.
[
7
]
G
.
La
c
o
m
b
e
,
D
.
F
e
l
i
o
t
,
E
.
B
o
e
sp
f
l
u
g
,
a
n
d
M
.
-
L.
P
o
t
e
t
,
“
C
o
mb
i
n
i
n
g
st
a
t
i
c
a
n
a
l
y
s
i
s
a
n
d
d
y
n
a
m
i
c
sy
m
b
o
l
i
c
e
x
e
c
u
t
i
o
n
i
n
a
t
o
o
l
c
h
a
i
n
t
o
d
e
t
e
c
t
f
a
u
l
t
i
n
j
e
c
t
i
o
n
v
u
l
n
e
r
a
b
i
l
i
t
i
e
s,
”
J
o
u
rn
a
l
o
f
C
ry
p
t
o
g
r
a
p
h
i
c
E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
4
,
n
o
.
1
,
p
p
.
1
4
7
–
1
6
4
,
A
p
r
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
3
3
8
9
-
023
-
0
0
3
1
0
-
8.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
8
1
4
I
n
t J Ad
v
Ap
p
l Sci
,
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
2
0
2
6
:
5
8
3
-
593
592
[
8
]
T.
S
u
t
t
e
r
,
T.
K
e
h
r
e
r
,
M
.
R
e
n
n
h
a
r
d
,
B
.
T
e
l
l
e
n
b
a
c
h
,
a
n
d
J.
K
l
e
i
n
,
“
D
y
n
a
mi
c
sec
u
r
i
t
y
a
n
a
l
y
si
s
o
n
A
n
d
r
o
i
d
:
a
sy
s
t
e
mat
i
c
l
i
t
e
r
a
t
u
r
e
r
e
v
i
e
w
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
1
2
,
p
p
.
5
7
2
6
1
–
5
7
2
8
7
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
4
.
3
3
9
0
6
1
2
.
[
9
]
C
.
Z
i
mm
e
r
m
a
n
,
J.
D
i
V
i
n
c
e
n
z
o
,
a
n
d
J.
A
l
d
r
i
c
h
,
“
S
o
u
n
d
g
r
a
d
u
a
l
v
e
r
i
f
i
c
a
t
i
o
n
w
i
t
h
sy
m
b
o
l
i
c
e
x
e
c
u
t
i
o
n
,
”
Pr
o
c
e
e
d
i
n
g
s o
f
t
h
e
AC
M
o
n
Pro
g
ra
m
m
i
n
g
L
a
n
g
u
a
g
e
s
,
v
o
l
.
8
,
p
p
.
2
5
4
7
–
2
5
7
6
,
Ja
n
.
2
0
2
4
,
d
o
i
:
1
0
.
1
1
4
5
/
3
6
3
2
9
2
7
.
[
1
0
]
H
.
Y
u
a
n
,
X
.
B
o
,
Z
.
Li
n
a
,
a
n
d
L.
C
h
e
n
g
y
a
n
g
,
“
A
c
o
n
t
r
o
l
f
l
o
w
g
r
a
p
h
o
p
t
i
m
i
z
a
t
i
o
n
met
h
o
d
f
o
r
e
n
h
a
n
c
i
n
g
f
u
z
z
t
e
s
t
i
n
g
,
”
I
EE
E
Ac
c
e
ss
,
v
o
l
.
1
2
,
p
p
.
1
6
9
3
7
0
–
1
6
9
3
7
8
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
4
.
3
4
5
2
9
3
8
.
[
1
1
]
Z.
Z
h
a
n
g
,
“
S
y
st
e
ma
t
i
c
a
p
p
r
o
a
c
h
e
s
f
o
r
i
mp
r
o
v
i
n
g
f
u
z
z
i
n
g
e
v
a
l
u
a
t
i
o
n
a
n
d
f
u
z
z
i
n
g
p
r
o
g
r
a
m
c
o
n
f
i
g
u
r
a
t
i
o
n
s.”
P
h
.
D
.
d
i
ss
e
r
t
a
t
i
o
n
,
T
h
e
U
n
i
v
e
r
si
t
y
o
f
Te
x
a
s
a
t
D
a
l
l
a
s,
R
i
c
h
a
r
d
so
n
,
Te
x
a
s,
2
0
2
4
.
[
1
2
]
W
.
C
h
e
n
e
t
a
l
.
,
“
S
y
z
G
e
n
+
+
:
d
e
p
e
n
d
e
n
c
y
i
n
f
e
r
e
n
c
e
f
o
r
a
u
g
me
n
t
i
n
g
k
e
r
n
e
l
d
r
i
v
e
r
f
u
z
z
i
n
g
,
”
i
n
2
0
2
4
I
EEE
S
y
m
p
o
si
u
m
o
n
S
e
c
u
ri
t
y
a
n
d
Pri
v
a
c
y
(
S
P)
,
M
a
y
2
0
2
4
,
p
p
.
4
6
6
1
–
4
6
7
7
,
d
o
i
:
1
0
.
1
1
0
9
/
S
P
5
4
2
6
3
.
2
0
2
4
.
0
0
2
6
9
.
[
1
3
]
A
.
P
o
k
h
a
r
e
l
,
“
D
e
t
e
c
t
i
n
g
s
o
f
t
w
a
r
e
v
u
l
n
e
r
a
b
i
l
i
t
y
w
i
t
h
s
y
mb
o
l
i
c
e
x
e
c
u
t
i
o
n
a
n
d
f
u
z
z
i
n
g
.
”
M
.
S
.
t
h
e
si
s
,
U
n
i
v
e
r
s
i
t
y
o
f
I
d
a
h
o
,
M
o
sc
o
w
,
R
u
ss
i
a
,
2
0
2
4
.
[
1
4
]
H
.
Tu
,
“
B
o
o
st
i
n
g
s
y
mb
o
l
i
c
e
x
e
c
u
t
i
o
n
f
o
r
h
e
a
p
-
b
a
se
d
v
u
l
n
e
r
a
b
i
l
i
t
y
d
e
t
e
c
t
i
o
n
a
n
d
e
x
p
l
o
i
t
g
e
n
e
r
a
t
i
o
n
,
”
i
n
2
0
2
3
I
EEE/
A
C
M
4
5
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
S
o
f
t
w
a
re
En
g
i
n
e
e
r
i
n
g
:
C
o
m
p
a
n
i
o
n
Pro
c
e
e
d
i
n
g
s
(
I
C
S
E
-
C
o
m
p
a
n
i
o
n
)
,
M
a
y
2
0
2
3
,
p
p
.
2
1
8
–
2
2
0
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
S
E
-
C
o
m
p
a
n
i
o
n
5
8
6
8
8
.
2
0
2
3
.
0
0
0
5
9
.
[
1
5
]
N
.
N
a
sc
i
me
n
t
o
,
“
S
a
f
e
g
u
a
r
d
a
n
a
l
y
z
e
r
:
a
f
u
z
z
t
e
s
t
i
n
g
t
o
o
l
.
”
M
.
S
.
t
h
e
s
i
s,
U
n
i
v
e
r
si
d
a
d
e
d
o
P
o
r
t
o
,
P
o
r
t
o
,
P
o
r
t
u
g
a
l
,
2
0
2
4
.
[
1
6
]
R
.
A
r
c
h
a
n
a
a
n
d
P
.
S
.
E.
Je
e
v
a
r
a
j
,
“
D
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
s
f
o
r
d
i
g
i
t
a
l
i
m
a
g
e
p
r
o
c
e
ssi
n
g
:
a
r
e
v
i
e
w
,
”
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
Re
v
i
e
w
,
v
o
l
.
5
7
,
n
o
.
1
,
Ja
n
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
0
4
6
2
-
023
-
1
0
6
3
1
-
z.
[
1
7
]
J.
L
i
u
e
t
a
l
.
,
“
A
p
p
l
i
c
a
t
i
o
n
o
f
d
e
e
p
l
e
a
r
n
i
n
g
-
b
a
s
e
d
n
a
t
u
r
a
l
l
a
n
g
u
a
g
e
p
r
o
c
e
ssi
n
g
i
n
m
u
l
t
i
l
i
n
g
u
a
l
se
n
t
i
me
n
t
a
n
a
l
y
si
s
,
”
Me
d
i
t
e
rr
a
n
e
a
n
J
o
u
rn
a
l
o
f
Ba
s
i
c
a
n
d
A
p
p
l
i
e
d
S
c
i
e
n
c
e
s
,
v
o
l
.
8
,
n
o
.
2
,
p
p
.
2
4
3
–
2
6
0
,
2
0
2
4
,
d
o
i
:
1
0
.
4
6
3
8
2
/
M
JB
A
S
.
2
0
2
4
.
8
2
1
9
.
[
1
8
]
J.
W
a
n
g
,
Y
.
H
u
a
n
g
,
C
.
C
h
e
n
,
Z.
L
i
u
,
S
.
W
a
n
g
,
a
n
d
Q
.
W
a
n
g
,
“
S
o
f
t
w
a
r
e
t
e
st
i
n
g
w
i
t
h
l
a
r
g
e
l
a
n
g
u
a
g
e
m
o
d
e
l
s:
s
u
r
v
e
y
,
l
a
n
d
s
c
a
p
e
,
a
n
d
v
i
s
i
o
n
,
”
I
E
EE
T
r
a
n
sa
c
t
i
o
n
s
o
n
S
o
f
t
w
a
re
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
5
0
,
n
o
.
4
,
p
p
.
9
1
1
–
9
3
6
,
A
p
r
.
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
TSE.
2
0
2
4
.
3
3
6
8
2
0
8
.
[
1
9
]
N
.
S
.
H
a
r
z
e
v
i
l
i
,
M
.
M
.
M
o
h
a
j
e
r
,
M
.
W
e
i
,
H
.
V
.
P
h
a
m
,
a
n
d
S
.
W
a
n
g
,
“
H
i
st
o
r
y
-
d
r
i
v
e
n
f
u
z
z
i
n
g
f
o
r
d
e
e
p
l
e
a
r
n
i
n
g
l
i
b
r
a
r
i
e
s,”
A
C
M
T
ra
n
s
a
c
t
i
o
n
s
o
n
S
o
f
t
w
a
r
e
E
n
g
i
n
e
e
ri
n
g
a
n
d
Me
t
h
o
d
o
l
o
g
y
,
v
o
l
.
3
4
,
n
o
.
1
,
p
p
.
1
–
2
9
,
Ja
n
.
2
0
2
5
,
d
o
i
:
1
0
.
1
1
4
5
/
3
6
8
8
8
3
8
.
[
2
0
]
A
.
M
e
h
m
o
o
d
,
Q
.
M
.
I
l
y
a
s,
M
.
A
h
ma
d
,
a
n
d
Z.
S
h
i
,
“
T
e
st
s
u
i
t
e
o
p
t
i
m
i
z
a
t
i
o
n
u
s
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s:
a
c
o
m
p
r
e
h
e
n
s
i
v
e
st
u
d
y
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
1
2
,
p
p
.
1
6
8
6
4
5
–
1
6
8
6
7
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
4
.
3
4
9
0
4
5
3
.
[
2
1
]
J.
C
h
e
n
,
L
.
Y
a
n
,
S
.
W
a
n
g
,
a
n
d
W
.
Zh
e
n
g
,
“
D
e
e
p
r
e
i
n
f
o
r
c
e
m
e
n
t
l
e
a
r
n
i
n
g
-
b
a
se
d
a
u
t
o
ma
t
i
c
t
e
st
c
a
se
g
e
n
e
r
a
t
i
o
n
f
o
r
h
a
r
d
w
a
r
e
v
e
r
i
f
i
c
a
t
i
o
n
,
”
J
o
u
r
n
a
l
o
f
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
G
e
n
e
r
a
l
S
c
i
e
n
c
e
,
v
o
l
.
6
,
n
o
.
1
,
p
p
.
4
0
9
–
4
2
9
,
N
o
v
.
2
0
2
4
,
d
o
i
:
1
0
.
6
0
0
8
7
/
j
a
i
g
s.
v
6
i
1
.
2
6
7
.
[
2
2
]
Y
.
W
a
n
g
,
Z.
W
u
,
Q
.
W
e
i
,
a
n
d
Q
.
W
a
n
g
,
“
N
e
u
F
u
z
z
:
e
f
f
i
c
i
e
n
t
f
u
z
z
i
n
g
w
i
t
h
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
I
E
E
E
Ac
c
e
ss
,
v
o
l
.
7
,
p
p
.
3
6
3
4
0
–
3
6
3
5
2
,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
1
9
.
2
9
0
3
2
9
1
.
[
2
3
]
Y
.
Li
e
t
a
l
.
,
“
V
-
F
u
z
z
:
v
u
l
n
e
r
a
b
i
l
i
t
y
p
r
e
d
i
c
t
i
o
n
-
a
ssi
s
t
e
d
e
v
o
l
u
t
i
o
n
a
r
y
f
u
z
z
i
n
g
f
o
r
b
i
n
a
r
y
p
r
o
g
r
a
ms
,
”
I
EEE
T
ra
n
s
a
c
t
i
o
n
s
o
n
C
y
b
e
r
n
e
t
i
c
s
,
v
o
l
.
5
2
,
n
o
.
5
,
p
p
.
3
7
4
5
–
3
7
5
6
,
M
a
y
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
T
C
Y
B
.
2
0
2
0
.
3
0
1
3
6
7
5
.
[
2
4
]
R
.
Y
i
n
g
,
J
.
Y
o
u
,
C
.
M
o
r
r
i
s,
X
.
R
e
n
,
W
.
L.
H
a
mi
l
t
o
n
,
a
n
d
J.
Le
s
k
o
v
e
c
,
“
H
i
e
r
a
r
c
h
i
c
a
l
g
r
a
p
h
r
e
p
r
e
s
e
n
t
a
t
i
o
n
l
e
a
r
n
i
n
g
w
i
t
h
d
i
f
f
e
r
e
n
t
i
a
b
l
e
p
o
o
l
i
n
g
,
”
i
n
3
2
n
d
C
o
n
f
e
ren
c
e
o
n
N
e
u
ra
l
I
n
f
o
rm
a
t
i
o
n
Pr
o
c
e
ssi
n
g
S
y
st
e
m
s (N
e
u
rI
P
S
2
0
1
8
)
,
2
0
1
8
,
p
p
.
4
8
0
5
–
4
8
1
5
.
[
2
5
]
J.
Le
e
,
I
.
Le
e
,
a
n
d
J.
K
a
n
g
,
“
S
e
l
f
-
a
t
t
e
n
t
i
o
n
g
r
a
p
h
p
o
o
l
i
n
g
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
3
6
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
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
,
L
o
n
g
B
e
a
c
h
,
C
a
l
i
f
o
r
n
i
a
,
2
0
1
9
,
p
p
.
1
-
10
.
[
2
6
]
Y
.
Z
h
o
u
,
S
.
Li
u
,
J.
S
i
o
w
,
X
.
D
u
,
a
n
d
Y
.
L
i
u
,
“
D
e
v
i
g
n
:
e
f
f
e
c
t
i
v
e
v
u
l
n
e
r
a
b
i
l
i
t
y
i
d
e
n
t
i
f
i
c
a
t
i
o
n
b
y
l
e
a
r
n
i
n
g
c
o
m
p
r
e
h
e
n
si
v
e
p
r
o
g
r
a
m
sema
n
t
i
c
s
v
i
a
g
r
a
p
h
n
e
u
r
a
l
n
e
t
w
o
r
k
s
,
”
i
n
3
3
rd
C
o
n
f
e
re
n
c
e
o
n
N
e
u
r
a
l
I
n
f
o
rm
a
t
i
o
n
P
ro
c
e
ss
i
n
g
S
y
st
e
m
s
(
N
e
u
rI
PS
2
0
1
9
)
,
2
0
1
9
,
p
p
.
1
0
1
9
7
–
1
0
2
0
7
.
[
2
7
]
M
.
W
u
e
t
a
l
.
,
“
O
n
e
f
u
z
z
i
n
g
s
t
r
a
t
e
g
y
t
o
r
u
l
e
t
h
e
m
a
l
l
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
4
4
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
S
o
f
t
w
a
re
En
g
i
n
e
e
ri
n
g
,
M
a
y
2
0
2
2
,
p
p
.
1
6
3
4
–
1
6
4
5
,
d
o
i
:
1
0
.
1
1
4
5
/
3
5
1
0
0
0
3
.
3
5
1
0
1
7
4
.
[
2
8
]
Y
.
F
e
n
g
,
Z
.
Lu
,
a
n
d
Q
.
C
a
o
,
“
S
e
c
u
r
e
sh
a
r
i
n
g
o
f
p
r
i
v
a
t
e
l
o
c
a
t
i
o
n
s
t
h
r
o
u
g
h
h
o
m
o
m
o
r
p
h
i
c
b
l
o
o
m
f
i
l
t
e
r
s,
”
i
n
2
0
1
8
I
EE
E
4
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
Bi
g
D
a
t
a
S
e
c
u
ri
t
y
o
n
C
l
o
u
d
(
B
i
g
D
a
t
a
S
e
c
u
r
i
t
y
)
,
I
E
EE
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
H
i
g
h
Pe
rf
o
rm
a
n
c
e
a
n
d
S
m
a
rt
C
o
m
p
u
t
i
n
g
(
H
PS
C
)
,
a
n
d
I
EEE
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
I
n
t
e
l
l
i
g
e
n
t
D
a
t
a
a
n
d
S
e
c
u
ri
t
y
(
I
D
S
)
,
M
a
y
2
0
1
8
,
p
p
.
1
0
7
–
1
1
3
,
d
o
i
:
1
0
.
1
1
0
9
/
B
D
S
/
H
P
S
C
/
I
D
S
1
8
.
2
0
1
8
.
0
0
0
3
4
.
[
2
9
]
G
.
K
l
e
e
s,
A
.
R
u
e
f
,
B
.
C
o
o
p
e
r
,
S
.
W
e
i
,
a
n
d
M
.
H
i
c
k
s
,
“
E
v
a
l
u
a
t
i
n
g
f
u
z
z
t
e
st
i
n
g
,
”
i
n
Pro
c
e
e
d
i
n
g
s
o
f
t
h
e
2
0
1
8
A
C
M
S
I
G
S
AC
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
e
r
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
s
S
e
c
u
r
i
t
y
,
O
c
t
.
2
0
1
8
,
p
p
.
2
1
2
3
–
2
1
3
8
,
d
o
i
:
1
0
.
1
1
4
5
/
3
2
4
3
7
3
4
.
3
2
4
3
8
0
4
.
[
3
0
]
P
.
Zo
n
g
,
T
.
L
v
,
D
.
W
a
n
g
,
Z.
D
e
n
g
,
R
.
L
i
a
n
g
,
a
n
d
K
.
C
h
e
n
,
“
F
u
z
z
G
u
a
r
d
:
f
i
l
t
e
r
i
n
g
o
u
t
u
n
r
e
a
c
h
a
b
l
e
i
n
p
u
t
s
i
n
d
i
r
e
c
t
e
d
g
r
e
y
-
b
o
x
f
u
z
z
i
n
g
t
h
r
o
u
g
h
d
e
e
p
l
e
a
r
n
i
n
g
,
”
i
n
P
ro
c
e
e
d
i
n
g
s
o
f
t
h
e
2
9
t
h
U
S
EN
I
X
C
o
n
f
e
re
n
c
e
o
n
S
e
c
u
r
i
t
y
S
y
m
p
o
s
i
u
m
,
2
0
2
0
,
p
p
.
2
2
5
5
–
2
2
6
9
,
d
o
i
:
1
0
.
5
5
5
5
/
3
4
8
9
2
1
2
.
3
4
8
9
3
3
9
.
[
3
1
]
D
.
S
h
e
,
K
.
P
e
i
,
D
.
Ep
s
t
e
i
n
,
J.
Y
a
n
g
,
B
.
R
a
y
,
a
n
d
S
.
Ja
n
a
,
“
N
EU
ZZ:
e
f
f
i
c
i
e
n
t
f
u
z
z
i
n
g
w
i
t
h
n
e
u
r
a
l
p
r
o
g
r
a
m
sm
o
o
t
h
i
n
g
,
”
i
n
2
0
1
9
I
EEE
S
y
m
p
o
s
i
u
m
o
n
S
e
c
u
r
i
t
y
a
n
d
Pri
v
a
c
y
(
S
P)
,
M
a
y
2
0
1
9
,
p
p
.
8
0
3
–
817
,
d
o
i
:
1
0
.
1
1
0
9
/
S
P
.
2
0
1
9
.
0
0
0
5
2
.
[
3
2
]
D
.
S
h
e
,
R
.
K
r
i
s
h
n
a
,
L.
Y
a
n
,
S
.
Ja
n
a
,
a
n
d
B
.
R
a
y
,
“
M
TF
u
z
z
:
f
u
z
z
i
n
g
w
i
t
h
a
mu
l
t
i
-
t
a
s
k
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
i
n
2
8
t
h
A
C
M
J
o
i
n
t
Me
e
t
i
n
g
o
n
Eu
r
o
p
e
a
n
S
o
f
t
w
a
re
En
g
i
n
e
e
ri
n
g
C
o
n
f
e
re
n
c
e
a
n
d
S
y
m
p
o
si
u
m
o
n
t
h
e
Fo
u
n
d
a
t
i
o
n
s
o
f
S
o
f
t
w
a
re
En
g
i
n
e
e
ri
n
g
,
N
o
v
.
2
0
2
0
,
p
p
.
7
3
7
–
749
,
d
o
i
:
1
0
.
1
1
4
5
/
3
3
6
8
0
8
9
.
3
4
0
9
7
2
3
.
[
3
3
]
S
.
G
a
n
e
t
a
l
.
,
“
G
R
EY
O
N
E:
d
a
t
a
f
l
o
w
se
n
si
t
i
v
e
f
u
z
z
i
n
g
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
2
9
t
h
U
S
E
N
I
X
C
o
n
f
e
r
e
n
c
e
o
n
S
e
c
u
ri
t
y
S
y
m
p
o
s
i
u
m
,
2
0
2
0
,
p
p
.
2
5
7
7
–
2
5
9
4
,
d
o
i
:
1
0
.
5
5
5
5
/
3
4
8
9
2
1
2
.
3
4
8
9
3
5
7
.
[
3
4
]
P
.
S
.
J
o
se
p
h
N
g
,
“
Ea
a
S
o
p
t
i
mi
z
a
t
i
o
n
:
a
v
a
i
l
a
b
l
e
y
e
t
h
i
d
d
e
n
i
n
f
o
r
m
a
t
i
o
n
t
e
c
h
n
o
l
o
g
y
i
n
f
r
a
st
r
u
c
t
u
r
e
i
n
s
i
d
e
me
d
i
u
m
si
z
e
e
n
t
e
r
p
r
i
se
,
”
T
e
c
h
n
o
l
o
g
i
c
a
l
F
o
re
c
a
s
t
i
n
g
a
n
d
S
o
c
i
a
l
C
h
a
n
g
e
,
v
o
l
.
1
3
2
,
p
p
.
1
6
5
–
1
7
3
,
J
u
l
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
t
e
c
h
f
o
r
e
.
2
0
1
8
.
0
1
.
0
3
0
.
[
3
5
]
P
.
S
.
Jo
s
e
p
h
N
g
,
Z.
F
u
,
R
.
Zh
a
n
g
,
a
n
d
K
.
Y
.
P
h
a
n
,
“
T
h
e
i
m
p
a
c
t
a
n
d
a
c
c
e
p
t
a
n
c
e
o
f
l
a
r
g
e
l
a
n
g
u
a
g
e
m
o
d
e
l
s
i
n
h
e
a
l
t
h
c
a
r
e
:
a
p
e
r
s
p
e
c
t
i
v
e
f
r
o
m
C
h
i
n
a
,
”
J
o
u
r
n
a
l
o
f
Ad
v
a
n
c
e
d
Re
s
e
a
rc
h
i
n
A
p
p
l
i
e
d
S
c
i
e
n
c
e
s
a
n
d
En
g
i
n
e
e
r
i
n
g
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
5
8
,
n
o
.
4
,
p
p
.
1
1
0
–
1
5
8
,
O
c
t
.
2
0
2
4
,
d
o
i
:
1
0
.
3
7
9
3
4
/
a
r
a
se
t
.
5
9
.
2
.
1
1
0
1
5
8
.
[
3
6
]
P
.
S
.
J
o
se
p
h
N
g
,
“
H
o
t
e
l
r
o
o
m
a
c
c
e
ss
c
o
n
t
r
o
l
:
a
n
N
F
C
a
p
p
r
o
a
c
h
e
c
o
t
o
u
r
i
sm f
r
a
mew
o
r
k
,
”
J
o
u
r
n
a
l
o
f
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
P
o
l
i
c
y
Ma
n
a
g
e
m
e
n
t
,
v
o
l
.
1
5
,
n
o
.
3
,
p
p
.
5
3
0
–
5
5
1
,
A
p
r
.
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
8
/
JS
TPM
-
10
-
2
0
2
1
-
0
1
5
3
.
[
3
7
]
M
.
A
l
i
a
s
e
t
a
l
.
,
“
A
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
-
p
o
w
e
r
e
d
i
ma
g
e
r
e
c
o
g
n
i
t
i
o
n
r
e
t
a
i
l
c
h
e
c
k
o
u
t
sy
st
e
ms,”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
A
d
v
a
n
c
e
s
i
n
Ap
p
l
i
e
d
S
c
i
e
n
c
e
s
,
v
o
l
.
1
5
,
n
o
.
1
,
p
p
.
1
8
7
–
1
9
7
,
M
a
r
.
2
0
2
6
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
a
a
s.v
1
5
.
i
1
.
p
p
1
8
7
-
1
9
7
.
[
3
8
]
L.
J
u
n
c
h
a
n
g
,
J
.
P
.
S
o
o
n
,
P
.
K
.
Y
u
e
n
,
a
n
d
W
.
S
.
W
a
n
,
“
Ev
e
r
y
se
c
o
n
d
c
o
u
n
t
s
f
o
r
s
e
a
r
c
h
a
n
d
r
e
s
c
u
e
:
a
sy
s
t
e
m
a
t
i
c
r
e
v
i
e
w
o
f
Ti
n
y
M
L
d
r
o
n
e
,
”
Ed
e
l
w
e
i
ss A
p
p
l
i
e
d
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
1
0
,
n
o
.
1
,
p
p
.
1
1
7
7
–
1
1
8
8
,
2
0
2
6
,
d
o
i
:
1
0
.
5
5
2
1
4
/
2
5
7
6
-
8
4
8
4
.
v
1
0
i
1
.
1
1
9
0
4
.
[
3
9
]
S
.
K
a
u
r
,
R
.
S
.
S
a
h
o
t
a
,
S
.
Y
.
Y
i
n
g
,
a
n
d
Y
.
H
a
o
t
i
a
n
,
“
Emer
g
i
n
g
t
r
e
n
d
s
i
n
I
n
d
u
st
r
y
4
.
0
a
n
d
p
r
e
d
i
c
t
i
v
e
ma
i
n
t
e
n
a
n
c
e
,
”
A
b
h
i
g
y
a
n
,
v
o
l
.
4
3
,
n
o
.
1
,
p
p
.
5
4
–
6
7
,
M
a
r
.
2
0
2
5
,
d
o
i
:
1
0
.
1
1
7
7
/
0
9
7
0
2
3
8
5
2
4
1
2
8
0
8
1
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.