I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
25
,
p
p
.
3064
~
3
0
7
4
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
3
.
pp
3
0
6
4
-
3
0
7
4
3064
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Enha
ncing
ma
lw
a
re det
ec
tion wi
th
genetic
alg
o
rithms a
nd
g
enerativ
e adv
ersa
ria
l net
wo
rks
Abid
Dh
iy
a
E
dd
i
ne
1,
2
,
G
ha
zli
Abdelk
a
der
1,
2
,
B
o
ua
che
M
o
ura
d
1
,3
1
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
F
a
c
u
l
t
y
o
f
E
x
a
c
t
S
c
i
e
n
c
e
s
,
Ta
h
r
i
M
o
h
a
med
U
n
i
v
e
r
si
t
y
,
B
e
c
h
a
r
,
A
l
g
e
r
i
a
2
I
n
n
o
v
a
t
i
o
n
s
i
n
I
n
f
o
r
ma
t
i
c
s
a
n
d
En
g
i
n
e
e
r
i
n
g
La
b
o
r
a
t
o
r
y
(
I
N
I
E
LA
B
)
,
Ta
h
r
i
M
o
h
a
me
d
U
n
i
v
e
r
si
t
y
,
B
e
c
h
a
r
,
A
l
g
e
r
i
a
3
H
i
g
h
P
e
r
f
o
r
ma
n
c
e
C
o
m
p
u
t
i
n
g
C
e
n
t
e
r
,
U
n
i
v
e
r
si
t
y
o
f
S
t
a
n
f
o
r
d
,
C
a
l
i
f
o
r
n
i
a
,
U
n
i
t
e
d
S
t
a
t
e
s
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
l 2
0
,
2
0
2
4
R
ev
is
ed
J
an
8
,
2
0
2
5
Acc
ep
ted
Ma
r
3
,
2
0
2
5
M
a
lwa
re
d
e
tec
ti
o
n
is
a
c
rit
ica
l
tas
k
in
c
y
b
e
rse
c
u
rit
y
,
n
e
c
e
ss
it
a
ti
n
g
t
h
e
c
re
a
ti
o
n
o
f
r
o
b
u
st
a
n
d
a
c
c
u
ra
te
d
e
tec
ti
o
n
m
o
d
e
ls.
Ou
r
p
r
o
p
o
sa
l
e
m
p
lo
y
s
a
h
o
li
stic
m
e
th
o
d
o
lo
g
y
f
o
r
id
e
n
ti
f
y
in
g
a
n
d
m
it
i
g
a
ti
n
g
m
a
lwa
re
u
sin
g
d
e
e
p
lea
rn
in
g
tec
h
n
i
q
u
e
s.
I
n
it
iall
y
,
a
c
u
sto
m
ize
d
g
e
n
e
ti
c
a
l
g
o
ri
th
m
is
e
m
p
lo
y
e
d
fo
r
fe
a
tu
re
se
lec
ti
o
n
,
re
d
u
c
i
n
g
d
ime
n
si
o
n
a
li
ty
a
n
d
e
n
h
a
n
c
in
g
t
h
e
d
isc
rimin
a
to
r
y
p
o
we
r
o
f
th
e
d
a
tas
e
t.
S
u
b
se
q
u
e
n
tl
y
,
a
d
e
e
p
n
e
u
ra
l
n
e
two
rk
i
s
train
e
d
o
n
t
h
e
se
lec
ted
fe
a
tu
r
e
s,
a
c
h
iev
in
g
h
i
g
h
a
c
c
u
ra
c
y
a
n
d
r
o
b
u
st
p
e
rfo
rm
a
n
c
e
in
d
isti
n
g
u
ish
i
n
g
b
e
t
we
e
n
m
a
lwa
r
e
a
n
d
b
e
n
ig
n
d
a
ta.
G
e
n
e
r
a
ti
v
e
a
d
v
e
rsa
rial
n
e
two
rk
s
a
re
a
lso
u
t
il
ize
d
to
e
v
a
lu
a
te
m
o
d
e
l
e
ffe
c
ti
v
e
n
e
ss
o
n
u
n
se
e
n
d
a
ta
a
n
d
e
n
s
u
re
t
h
e
m
o
d
e
l'
s
ro
b
u
st
n
e
ss
a
n
d
g
e
n
e
ra
li
z
a
ti
o
n
c
a
p
a
b
il
it
ies
.
E
v
a
lu
a
ti
o
n
o
f
t
h
e
p
ro
p
o
se
d
m
o
d
e
l
d
e
m
o
n
stra
tes
a
c
c
u
ra
te
m
a
lwa
re
d
e
tec
ti
o
n
with
h
i
g
h
g
e
n
e
ra
li
z
a
ti
o
n
c
a
p
a
b
il
i
ti
e
s.
F
u
rth
e
rm
o
re
,
fu
t
u
re
re
se
a
rc
h
sh
o
u
ld
fo
c
u
s
o
n
d
e
v
e
lo
p
i
n
g
a
n
d
d
e
p
l
o
y
i
n
g
p
ra
c
ti
c
a
l
to
o
ls
o
r
sy
ste
m
s
th
a
t
imp
lem
e
n
t
th
e
p
r
o
p
o
se
d
m
o
d
e
l
fo
r
re
a
l
-
ti
m
e
m
a
lwa
re
d
e
tec
ti
o
n
in
o
p
e
ra
ti
o
n
a
l
e
n
v
iro
n
m
e
n
ts.
Th
i
s
re
se
a
rc
h
m
a
k
e
s
a
sig
n
ifi
c
a
n
t
c
o
n
tri
b
u
ti
o
n
to
th
e
f
ield
o
f
m
a
lwa
re
d
e
tec
ti
o
n
a
n
d
p
r
o
v
i
d
e
s
e
x
c
e
ll
e
n
t
o
p
p
o
rt
u
n
it
ies
fo
r
p
ra
c
ti
c
a
l
imp
lem
e
n
tatio
n
in
th
e
fi
e
ld
o
f
c
y
b
e
rse
c
u
rit
y
.
K
ey
w
o
r
d
s
:
C
y
b
er
s
ec
u
r
ity
Dee
p
lear
n
in
g
Gen
er
ativ
e
ar
tific
ial
in
tellig
en
ce
Gen
etic
alg
o
r
ith
m
Ma
lwar
e
d
etec
tio
n
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
:
Ab
id
Dh
iy
a
E
d
d
in
e
Dep
ar
tm
en
t o
f
Ma
th
em
atics a
n
d
C
o
m
p
u
te
r
Scien
ce
s
,
Facu
lty
o
f
E
x
ac
t Scie
n
ce
s
,
T
ah
r
i M
o
h
am
ed
Un
iv
er
s
ity
o
f
B
ec
h
ar
I
s
tik
lal
Stre
et,
0
8
0
0
0
,
B
ec
h
ar
,
Alg
er
ia
E
m
ail: a
b
id
.
d
h
iy
ae
d
d
in
e@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
T
ec
h
n
o
lo
g
y
is
ad
v
an
ci
n
g
d
ail
y
an
d
b
ei
n
g
u
s
ed
m
o
r
e
f
r
eq
u
en
tly
.
T
h
is
ad
v
an
ce
m
en
t
in
t
ec
h
n
o
lo
g
y
also
cr
ea
tes
a
co
n
d
u
civ
e
atm
o
s
p
h
er
e
f
o
r
th
e
s
p
r
ea
d
o
f
m
alw
ar
e,
[
1
]
lead
in
g
to
d
ata
b
r
ea
ch
es,
f
in
an
cial
lo
s
s
es,
an
d
d
is
r
u
p
tio
n
s
o
f
cr
itical
in
f
r
astru
ctu
r
e
[
2
]
.
Acc
o
r
d
in
g
to
th
e
s
tatis
tics
,
ev
er
y
d
ay
,
th
e
AV
-
T
E
ST
I
n
s
titu
te
d
etec
ts
o
v
er
4
5
0
,
0
0
0
n
o
v
el
p
o
ten
tially
u
n
wan
ted
ap
p
licatio
n
s
(
PUA)
an
d
m
alicio
u
s
p
r
o
g
r
a
m
s
(
m
alwa
r
e)
[
3
]
.
Fig
u
r
e
1
s
h
o
ws th
e
to
tal
am
o
u
n
t o
f
m
alwa
r
e
a
n
d
PUA
in
th
e
last
f
iv
e
y
ea
r
s
ac
co
r
d
in
g
to
A
V
-
T
E
ST
I
n
s
titu
te
.
Ma
lwar
e,
o
f
ten
k
n
o
wn
as
m
al
icio
u
s
p
r
o
g
r
am
,
is
wid
ely
ac
k
n
o
wled
g
ed
as
o
n
e
o
f
t
h
e
m
o
s
t
d
an
g
er
o
u
s
cy
b
er
th
r
ea
ts
a
n
d
r
is
k
s
to
c
o
n
t
em
p
o
r
ar
y
c
o
m
p
u
ter
s
y
s
tem
s
.
T
h
is
all
-
en
co
m
p
ass
in
g
ter
m
d
escr
ib
es
an
y
c
o
d
e
th
at
m
ig
h
t
h
av
e
n
eg
ativ
e
d
am
ag
in
g
ef
f
ec
ts
[
4
]
.
I
t
allu
d
es
to
a
b
r
o
ad
s
p
ec
tr
u
m
o
f
h
ar
m
f
u
l
s
o
f
twar
e
th
at
m
ig
h
t
m
an
if
est
in
v
ar
io
u
s
way
s
.
T
h
ese
in
clu
d
e
r
an
s
o
m
war
e
,
ad
w
ar
e,
s
p
y
war
e,
tr
o
jan
h
o
r
s
es,
wo
r
m
s
,
an
d
v
ir
u
s
es.
T
h
e
m
alwa
r
e
aim
s
to
co
m
p
r
o
m
is
e
th
e
s
ec
u
r
ity
an
d
p
r
iv
ac
y
o
f
a
u
s
er
'
s
co
m
p
u
ter
b
y
d
is
r
u
p
tin
g
its
o
p
er
atio
n
s
,
in
f
iltra
tin
g
u
n
a
u
th
o
r
ized
d
ata
o
r
s
y
s
tem
s
,
o
r
em
p
lo
y
in
g
o
t
h
er
m
eth
o
d
s
[
5
]
.
T
r
a
d
itio
n
al
m
alwa
r
e
d
etec
tio
n
tech
n
iq
u
es
f
r
eq
u
e
n
tly
d
e
p
en
d
o
n
m
eth
o
d
s
th
at
r
ely
o
n
s
ig
n
a
tu
r
es,
wh
ich
f
in
d
it
d
if
f
icu
lt
to
k
ee
p
u
p
with
th
e
q
u
ick
c
h
an
g
es
in
m
alwa
r
e
ty
p
es.
Mo
r
e
a
d
v
an
c
ed
m
eth
o
d
s
t
h
at
ca
n
ef
f
icien
tly
id
en
tif
y
a
n
d
ev
alu
ate
m
alwa
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
E
n
h
a
n
ci
n
g
ma
lw
a
r
e
d
etec
tio
n
w
ith
g
en
etic
a
lg
o
r
ith
ms a
n
d
g
en
era
tive
…
(
A
b
id
Dh
iy
a
E
d
d
i
n
e
)
3065
in
r
ea
l
-
tim
e
ar
e
th
er
ef
o
r
e
b
e
co
m
in
g
m
o
r
e
an
d
m
o
r
e
n
ec
ess
ar
y
,
p
ar
ticu
lar
ly
in
s
itu
atio
n
s
wh
er
e
h
ar
m
f
u
l
ac
tiv
ity
tak
es p
lace
in
v
o
latile
m
em
o
r
y
[
6
]
.
AI
ap
p
r
o
ac
h
es
h
av
e
b
ec
o
m
e
ef
f
ec
tiv
e
to
o
ls
in
co
m
b
atin
g
m
alwa
r
e.
B
y
u
tili
zin
g
m
ac
h
in
e
lear
n
in
g
tech
n
iq
u
es
an
d
d
ee
p
n
e
u
r
al
n
etwo
r
k
s
,
AI
p
o
wer
ed
m
alwa
r
e
d
etec
tio
n
s
y
s
tem
s
h
av
e
th
e
ab
ilit
y
to
e
x
am
in
e
ex
ten
s
iv
e
v
o
lu
m
es
o
f
d
ata
wh
er
e
th
e
aim
is
to
u
n
co
v
er
p
atter
n
s
an
d
an
o
m
alies
th
at
in
d
icate
p
o
ten
tially
d
an
g
er
o
u
s
ac
tiv
ity
[
7
]
.
T
h
ese
s
y
s
tem
s
h
av
e
s
h
o
wn
p
r
o
m
is
in
g
r
esu
lts
in
d
etec
tin
g
p
r
ev
io
u
s
ly
u
n
k
n
o
wn
m
alwa
r
es,
wh
er
e
co
n
v
en
tio
n
al
ap
p
r
o
ac
h
es
th
at
r
ely
o
n
s
ig
n
a
tu
r
es
m
ay
f
all
s
h
o
r
t
[
8
]
.
AI
p
o
wer
ed
s
o
lu
tio
n
s
in
cy
b
er
s
ec
u
r
ity
p
r
o
v
id
e
a
p
r
o
ac
t
iv
e
ap
p
r
o
ac
h
b
y
co
n
s
tan
tly
ac
q
u
ir
in
g
k
n
o
wled
g
e
f
r
o
m
f
r
esh
d
ata
an
d
ad
j
u
s
tin
g
to
ch
an
g
i
n
g
th
r
ea
ts
[
9
]
.
Fig
u
r
e
1
.
T
o
tal
am
o
u
n
t o
f
m
alwa
r
es a
n
d
PUA
in
th
e
last
5
y
ea
r
s
ac
co
r
d
in
g
to
AV
-
T
E
ST
I
n
s
titu
te
R
ec
en
t
r
esear
ch
h
as
m
ad
e
s
ig
n
if
ican
t
s
tr
id
es
in
th
e
d
o
m
ain
o
f
m
alwa
r
e
d
etec
tio
n
.
R
esear
ch
er
s
in
[
1
0
]
d
ev
el
o
p
ed
a
u
s
er
-
f
r
ie
n
d
l
y
web
s
ite
f
o
r
m
alwa
r
e
d
etec
tio
n
a
n
d
p
r
e
d
ictio
n
,
ca
p
ab
le
o
f
id
en
tify
in
g
th
e
ty
p
e
o
f
m
alwa
r
e
e
n
co
d
e
d
in
a
f
ile
.
Min
er
v
a
is
n
ew
ap
p
r
o
ac
h
p
r
esen
ted
b
y
[
1
1
]
f
o
r
t
h
e
r
a
n
s
o
m
war
e
d
etec
tio
n
.
Min
er
v
a
u
s
es
all
o
f
th
e
o
p
er
atio
n
s
th
at
f
iles
r
ec
eiv
e
d
u
r
in
g
a
g
iv
en
tim
e
in
ter
v
al
to
cr
ea
te
b
eh
av
io
r
al
p
r
o
f
iles
o
f
th
e
f
iles
in
o
r
d
er
to
d
etec
t
r
an
s
o
m
war
e.
Mic
r
o
s
o
f
t
Ma
lwa
r
es
d
ataset
was
u
s
ed
in
[
1
2
]
with
m
an
y
m
ac
h
in
e
lear
n
in
g
alg
o
r
ith
m
s
f
o
r
class
if
y
in
g
th
e
m
alicio
u
s
an
d
th
e
b
e
n
ig
n
s
o
f
twar
e
b
ased
o
n
th
e
an
aly
s
is
o
f
ex
ec
u
tab
le
f
ile
m
etad
ata.
Ma
lMe
m
-
2
0
2
2
it
is
a
d
ataset
to
ass
e
s
s
th
e
ef
f
ec
tiv
en
ess
o
f
m
em
o
r
y
-
b
ased
o
b
f
u
s
ca
ted
m
alwa
r
e
d
etec
tio
n
m
eth
o
d
s
[
1
3
]
,
th
is
d
ataset
u
s
ed
b
y
[
1
4
]
,
an
d
also
[
1
5
]
u
s
ed
it
b
u
t
with
a
cu
s
to
m
ized
K
-
Nea
r
est
Neig
h
b
o
r
s
alg
o
r
ith
m
,
a
n
d
it u
s
ed
b
y
[
1
]
f
o
r
d
etec
tin
g
m
alwa
r
es in
b
ig
d
ata
en
v
ir
o
n
m
en
t.
Ma
lwar
es
d
etec
tio
n
also
ca
n
b
e
ap
p
lied
to
m
o
b
ile
p
h
o
n
es
a
s
it
is
p
r
esen
ted
i
n
p
ap
er
[
1
6
]
,
wh
er
e
t
h
e
m
ain
id
ea
is
u
s
in
g
d
ee
p
lear
n
in
g
tech
n
iq
u
es
an
d
ex
p
lain
ab
le
ar
tific
ial
in
tellig
en
ce
(
XAI
)
f
o
r
th
e
d
etec
tio
n
a
n
d
th
e
class
if
icatio
n
o
f
an
d
r
o
id
m
o
b
ile
m
alwa
r
es
u
s
in
g
C
I
C
An
d
Ma
l2
0
1
7
d
ataset
[
1
7
]
.
it
is
an
o
th
er
w
o
r
k
i
n
th
e
f
ield
o
f
m
o
b
ile
p
h
o
n
es
wh
ich
is
b
ased
o
n
s
tack
in
g
,
p
r
esen
t
ed
a
m
ac
h
in
e
lear
n
i
n
g
(
ML
)
m
o
d
el
f
o
r
d
etec
tin
g
m
alwa
r
e
th
at
u
s
es
an
en
s
em
b
le
ap
p
r
o
ac
h
f
o
r
An
d
r
o
id
d
e
v
ices.
T
h
e
a
u
th
o
r
s
i
n
th
is
r
esear
ch
m
er
g
ed
b
etwe
e
n
th
e
C
I
C
-
Ma
lMe
m
2
0
2
2
a
n
d
C
I
C
-
Ma
lDr
o
id
2
0
2
0
d
atasets
in
th
e
ex
a
m
in
atio
n
o
f
th
e
e
f
f
ec
tiv
en
ess
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el.
T
h
e
I
n
ter
n
et
o
f
T
h
in
g
s
also
h
as
a
s
h
ar
e
in
p
r
ev
io
u
s
wo
r
k
s
d
e
v
o
ted
to
d
etec
tin
g
m
alwa
r
es,
an
d
th
is
is
wh
at
was
em
b
o
d
ie
d
in
th
e
r
esear
ch
[
1
8
]
an
d
[
1
9
]
.
T
h
ese
s
tu
d
ies
co
llectiv
ely
h
ig
h
lig
h
t
th
e
p
o
te
n
tial
o
f
m
ac
h
i
n
e
lear
n
i
n
g
a
n
d
b
eh
av
io
r
-
b
ased
d
etec
tio
n
in
en
h
a
n
cin
g
m
alwa
r
e
d
etec
tio
n
ca
p
ab
ilit
ies.
W
h
ile
all
p
r
ev
io
u
s
r
esear
ch
h
as
d
o
n
e
w
ell
in
cr
ea
tin
g
r
o
b
u
s
t
m
o
d
els
to
id
en
tif
y
m
alwa
r
e,
th
e
q
u
esti
o
n
a
r
is
es
ab
o
u
t
th
e
ef
f
ec
tiv
en
ess
o
f
th
ese
m
o
d
els
with
g
en
er
ativ
e
d
ata,
s
o
we
will
tr
y
to
an
s
wer
th
is
q
u
esti
o
n
in
th
is
r
esear
ch
p
ap
er
.
I
n
th
is
r
esear
ch
,
we
will
u
s
e
th
e
C
I
C
-
Ma
lMe
m
2
0
2
2
d
ataset
to
b
u
ild
a
m
o
d
el
to
i
d
en
tify
m
alwa
r
e.
W
e
will
ca
ll
g
en
etic
alg
o
r
ith
m
to
ch
o
o
s
e
th
e
ap
p
r
o
p
r
iate
f
ea
t
u
r
es,
th
en
we
will
tr
ain
th
e
f
in
al
d
ata
u
s
in
g
d
ee
p
n
eu
r
al
n
etwo
r
k
s
,
an
d
th
en
we
will
test
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
m
o
d
el
u
s
in
g
m
ac
h
in
e
lear
n
in
g
e
v
alu
atio
n
m
etr
ics.
T
h
e
n
e
x
t
s
tag
e
is
to
c
r
ea
te
a
g
e
n
er
ativ
e
a
d
v
er
s
ar
ial
n
etwo
r
k
,
t
h
en
g
en
er
ate
d
ata
s
i
m
ilar
to
th
e
o
r
ig
i
n
al
d
ata.
Fin
ally
,
we
will
test
th
e
ef
f
ec
tiv
en
ess
o
f
o
u
r
m
o
d
el
with
th
e
g
en
e
r
ativ
e
d
ata
in
o
r
d
er
to
d
is
co
v
er
t
h
e
ef
f
ec
tiv
en
ess
o
f
th
e
g
en
er
ativ
e
d
ata
in
th
ese
ca
s
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
25
:
3
0
6
4
-
3
0
7
4
3066
2.
P
RO
P
O
SE
D
M
E
T
H
O
DO
L
O
G
Y
T
o
en
h
a
n
ce
m
alwa
r
e
d
etec
tio
n
,
we
im
p
le
m
en
ted
a
m
u
lti
-
s
tag
e
ap
p
r
o
ac
h
s
tar
tin
g
with
th
e
p
r
ep
r
o
ce
s
s
in
g
o
f
th
e
C
I
C
-
Ma
lMe
m
2
0
2
2
d
ataset,
as
illu
s
tr
ated
in
Fig
u
r
e
2
.
A
cu
s
to
m
ized
g
en
etic
alg
o
r
ith
m
(
GA)
was
em
p
lo
y
e
d
to
s
elec
t
th
e
m
o
s
t
d
is
cr
im
in
ativ
e
f
ea
t
u
r
es,
en
s
u
r
in
g
th
e
m
o
d
el
was
b
u
ilt
o
n
th
e
m
o
s
t
r
elev
an
t
d
ata.
Usi
n
g
th
ese
s
elec
ted
f
ea
tu
r
es,
a
d
ee
p
n
e
u
r
al
n
etwo
r
k
(
DNN)
was
tr
ain
ed
t
o
class
if
y
s
am
p
les
as
b
en
ig
n
o
r
m
alicio
u
s
.
T
o
im
p
r
o
v
e
th
e
m
o
d
el'
s
r
o
b
u
s
tn
ess
an
d
g
e
n
er
aliza
tio
n
,
a
g
e
n
er
ativ
e
ad
v
er
s
ar
ial
n
etwo
r
k
(
GAN)
was
th
en
u
tili
ze
d
t
o
g
en
er
ate
s
y
n
t
h
etic
b
en
i
g
n
a
n
d
m
alicio
u
s
s
am
p
les.
T
h
e
G
AN,
co
n
s
is
tin
g
o
f
a
Gen
er
ato
r
an
d
a
Dis
cr
im
in
ato
r
,
cr
ea
ted
n
ew
d
ata
p
o
in
ts
th
a
t
clo
s
ely
r
esem
b
led
r
ea
l
-
wo
r
l
d
ex
am
p
les,
wh
ich
wer
e
u
s
ed
to
f
u
r
th
er
ev
alu
ate
th
e
DNN.
T
h
e
m
o
d
el's
p
er
f
o
r
m
an
ce
was
ass
es
s
ed
o
n
th
is
s
y
n
th
etic
d
ata
to
d
eter
m
in
e
its
ef
f
ec
tiv
e
n
ess
.
I
f
th
e
m
o
d
el
p
er
f
o
r
m
ed
well,
it
was
d
ee
m
ed
r
o
b
u
s
t;
o
th
er
wis
e
,
f
u
r
th
e
r
r
ef
i
n
em
en
t
was
n
ec
ess
ar
y
.
T
h
is
in
teg
r
ated
ap
p
r
o
ac
h
o
f
f
ea
tu
r
e
s
elec
tio
n
,
d
ee
p
lear
n
in
g
,
an
d
s
y
n
th
e
tic
d
ata
g
en
er
atio
n
en
s
u
r
ed
th
at
o
u
r
m
alwa
r
e
d
ete
ctio
n
m
o
d
el
was b
o
th
ac
cu
r
ate
an
d
r
esil
ien
t a
g
ain
s
t d
iv
e
r
s
e
th
r
ea
ts
.
Fig
u
r
e
2
.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
2
.
1
.
Da
t
a
prepro
ce
s
s
ing
As
m
en
tio
n
ed
p
r
ev
i
o
u
s
ly
,
w
e
u
s
ed
th
e
C
I
C
-
Ma
lMe
m
2
0
2
2
d
ataset
in
o
u
r
wo
r
k
.
Ou
r
d
atab
ase
co
n
tain
s
5
8
5
9
6
r
o
ws
an
d
5
7
co
lu
m
n
s
,
d
iv
id
ed
in
to
b
e
n
ig
n
(
2
9
2
9
8
)
a
n
d
m
alicio
u
s
(
2
9
2
9
8
)
.
T
h
e
m
alicio
u
s
class
co
n
tain
s
th
r
ee
ca
teg
o
r
ies
wh
ich
ar
e
Sp
y
war
e
(
1
0
0
2
0
)
,
R
an
s
o
m
war
e
(
9
7
9
1
)
,
an
d
T
r
o
jan
(
9
4
8
7
)
.
As
a
p
r
e
-
p
r
o
ce
s
s
in
g
o
f
th
e
d
ata,
we
co
d
ed
th
e
b
en
ig
n
class
as
0
a
n
d
th
e
m
alicio
u
s
class
as
1
.
As
f
o
r
th
e
ca
teg
o
r
y
co
lu
m
n
,
we
c
o
d
ed
th
e
b
e
n
ig
n
as 0
,
Sp
y
war
e
as 1
,
R
an
s
o
m
w
ar
e
as 2
,
an
d
T
r
o
ja
n
as 3
.
G
e
n
e
r
a
t
o
r
D
i
s
c
r
i
m
i
n
a
t
o
r
G
e
n
e
r
a
t
i
v
e
A
d
v
e
r
s
a
r
i
a
l
N
e
t
w
o
r
k
N
e
w
d
a
t
a
M
o
d
e
l
E
v
a
l
u
a
t
i
o
n
o
n
t
h
e
n
e
w
d
a
t
a
B
a
d
T
r
a
i
n
i
n
g
s
t
a
g
e
D
a
t
a
g
e
n
e
r
a
t
i
o
n
s
t
a
g
e
B
a
d
G
o
o
d
B
e
s
t
f
e
a
t
u
r
e
s
I
n
i
a
l
i
z
a
t
i
o
n
S
e
l
e
c
t
i
o
n
,
c
r
o
s
s
o
v
e
r
,
m
u
t
a
t
i
o
n
E
v
a
l
u
a
t
i
o
n
R
e
c
o
r
d
i
n
g
&
E
l
i
t
i
s
m
G
e
n
e
t
i
c
A
l
g
o
r
i
t
h
m
E
v
a
l
u
a
t
i
o
n
f
i
t
n
e
s
s
D
e
e
p
N
e
u
r
a
l
N
e
t
w
o
r
k
T
r
a
i
n
i
n
g
T
r
a
i
n
e
d
M
o
d
e
l
F
i
n
a
l
M
o
d
e
l
G
o
o
d
E
v
a
l
u
a
t
i
o
n
C
l
a
s
s
i
f
i
c
a
t
i
o
n
m
e
t
r
i
c
s
+
A
U
C
S
c
o
r
e
F
e
a
t
u
r
e
s
e
l
e
c
t
i
o
n
s
t
a
g
e
U
p
d
a
t
e
p
o
p
u
l
a
t
i
o
n
P
r
e
p
r
o
c
e
s
s
i
n
g
C
I
C
-
M
a
l
M
e
m
2
0
2
2
R
o
b
u
s
t
m
o
d
e
l
G
o
o
d
B
a
d
M
o
d
e
l
R
e
f
i
n
e
m
e
n
t
N
e
e
d
e
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
E
n
h
a
n
ci
n
g
ma
lw
a
r
e
d
etec
tio
n
w
ith
g
en
etic
a
lg
o
r
ith
ms a
n
d
g
en
era
tive
…
(
A
b
id
Dh
iy
a
E
d
d
i
n
e
)
3067
2
.
2
.
F
e
a
t
ure
s
elec
t
io
n us
ing
g
enet
ic
a
lg
o
rit
hm
T
h
e
s
elec
tio
n
o
f
b
est
f
ea
tu
r
es
is
a
tech
n
iq
u
e
aim
s
to
lis
t
th
e
s
ig
n
if
ican
t
f
ea
tu
r
es
th
at
h
elp
b
u
ild
b
etter
class
if
ier
s
an
d
r
ed
u
ce
co
m
p
u
t
atio
n
al
o
v
er
l
o
ad
[
2
0
]
.
Gen
etic
alg
o
r
ith
m
s
ar
e
o
p
tim
izatio
n
t
ec
h
n
iq
u
es
in
s
p
ir
ed
b
y
n
atu
r
al
s
elec
tio
n
.
T
h
ey
u
s
e
p
o
p
u
latio
n
s
o
f
ca
n
d
id
ate
s
o
lu
tio
n
s
,
f
av
o
r
in
g
th
e
f
ittes
t
in
d
iv
id
u
als
th
r
o
u
g
h
iter
ativ
e
s
elec
tio
n
,
cr
o
s
s
o
v
er
,
an
d
m
u
tatio
n
.
T
h
is
m
im
ics
th
e
p
r
o
ce
s
s
o
f
e
v
o
lu
tio
n
,
ai
d
in
g
in
ef
f
icien
tl
y
f
in
d
in
g
o
p
tim
al
s
o
lu
tio
n
s
to
co
m
p
lex
p
r
o
b
lem
s
[
2
1
]
.
I
n
o
u
r
r
esear
ch
we
cr
ea
ted
a
cu
s
to
m
ized
g
en
etic
alg
o
r
ith
m
f
o
r
f
ea
t
u
r
e
s
elec
tio
n
,
th
is
alg
o
r
ith
m
is
s
h
o
wn
in
th
e
alg
o
r
ith
m
1
.
Alg
o
r
ith
m
1
.
Featu
r
es selectio
n
g
en
etic
alg
o
r
ith
m
Input: data,
population_size, generations_number,
mutation_rate
.
Output: optimal_features
1
population
initialize_population_randomly()
2
score
evaluate(population)
3
foreach generation in generations_number do
4
parents
select_parents_for_crossover(population,
population_size, score)
5
next_generation
create_next_generation(parents)
6
next_generation
make_mutation(next_generation)
7
score
evaluate(next_generation)
8
add_to_results(score,
next_generation)
9
next_generation
elitism(next_generation)
10
population
next_generation
11
end
12
optimal_features
get_best_features_from_max_score()
T
h
r
o
u
g
h
th
e
iter
ativ
e
a
p
p
licatio
n
o
f
s
elec
tio
n
,
c
r
o
s
s
o
v
er
,
a
n
d
m
u
tatio
n
o
p
er
ato
r
s
,
g
en
etic
alg
o
r
ith
m
s
ex
p
lo
r
e
th
e
s
o
lu
tio
n
s
p
ac
e,
g
r
ad
u
ally
im
p
r
o
v
in
g
th
e
q
u
ality
o
f
s
o
lu
tio
n
s
o
v
er
s
u
cc
ess
iv
e
g
en
er
atio
n
s
.
B
y
p
r
o
m
o
tin
g
th
e
s
u
r
v
iv
al
a
n
d
r
ep
r
o
d
u
ctio
n
o
f
in
d
iv
id
u
als
with
h
ig
h
er
f
itn
ess
v
alu
es,
g
en
etic
alg
o
r
ith
m
s
ef
f
icien
tly
n
av
i
g
ate
co
m
p
lex
s
ea
r
ch
s
p
ac
es
an
d
d
is
co
v
e
r
h
ig
h
-
q
u
ality
s
o
lu
tio
n
s
to
o
p
tim
iz
atio
n
p
r
o
b
lem
s
in
a
m
an
n
er
in
s
p
ir
e
d
b
y
n
atu
r
al
ev
o
lu
tio
n
.
2
.
3
.
Da
t
a
t
r
a
ini
ng
Fo
r
th
e
d
ata
tr
ain
in
g
,
we
tr
ain
ed
o
u
r
f
in
al
d
ata
u
s
in
g
d
ee
p
n
eu
r
al
n
etwo
r
k
s
.
A
d
ee
p
n
e
u
r
al
n
etwo
r
k
is
s
am
e
as
an
ar
tific
ial
n
eu
r
al
n
etwo
r
k
b
u
t
with
m
an
y
h
id
d
en
lay
er
s
p
o
s
itio
n
ed
b
etwe
en
th
e
in
p
u
t
an
d
o
u
tp
u
t
lay
er
s
,
en
ab
lin
g
it
to
lear
n
c
o
m
p
lex
p
atter
n
s
in
d
ata.
I
t
i
s
ch
ar
ac
ter
ized
b
y
its
d
ep
th
,
wh
ich
allo
ws
it
t
o
r
ep
r
esen
t
h
ier
a
r
ch
ical
f
ea
tu
r
es
an
d
a
b
s
tr
ac
t
r
ep
r
esen
tatio
n
s
[
2
2
]
.
T
h
e
T
a
b
le
1
s
h
o
ws
th
e
s
tr
u
ctu
r
e
o
f
o
u
r
d
ee
p
n
eu
r
al
n
etwo
r
k
h
o
wev
er
T
a
b
le
2
s
h
o
ws th
e
u
s
ed
p
ar
am
eter
s
f
o
r
th
e
tr
ain
i
n
g
p
r
o
ce
s
s
.
T
ab
le
1
.
T
h
e
s
tr
u
ctu
r
e
o
f
th
e
d
ee
p
n
eu
r
al
n
etwo
r
k
La
y
e
r
(
t
y
p
e
)
O
u
t
p
u
t
s
h
a
p
e
D
e
n
se
D
r
o
p
o
u
t
D
e
n
se
D
r
o
p
o
u
t
D
e
n
se
A
c
t
i
v
a
t
i
o
n
(
si
g
m
o
i
d
)
(
n
o
n
e
,
1
2
8
)
(
n
o
n
e
,
1
2
8
)
(
n
o
n
e
,
6
4
)
(
n
o
n
e
,
6
4
)
(
n
o
n
e
,
1
)
(
n
o
n
e
,
1
)
T
ab
le
2
.
T
h
e
u
s
ed
p
a
r
am
eter
s
f
o
r
tr
ain
in
g
d
ee
p
n
eu
r
al
n
etwo
r
k
P
a
r
a
me
t
e
r
V
a
l
u
e
Le
a
r
n
i
n
g
r
a
t
e
0
.
0
0
1
O
p
t
i
mi
z
e
r
A
d
a
m
Lo
ss f
u
n
c
t
i
o
n
B
i
n
a
r
y
c
r
o
ss
e
n
t
r
o
p
y
Ep
o
c
h
s
30
I
n
o
r
d
er
to
ass
ess
th
e
ef
f
icac
ity
o
f
o
u
r
m
o
d
el,
we
h
av
e
u
s
ed
ass
ess
m
en
t
cr
iter
ia
in
clu
d
in
g
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F
1
-
S
co
r
e
.
Acc
u
r
ac
y
is
ca
lcu
lated
b
y
th
e
p
r
o
p
o
r
tio
n
o
f
p
r
o
p
er
ly
id
en
ti
f
ied
ca
s
es,
an
d
it
is
co
m
p
u
ted
u
s
in
g
(
1
)
:
=
(
+
)
+
+
+
(
1
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
25
:
3
0
6
4
-
3
0
7
4
3068
T
h
e
p
r
ec
is
io
n
m
ea
s
u
r
e,
wh
ich
q
u
an
tifie
s
th
e
ac
cu
r
a
cy
o
f
p
o
s
i
tiv
e
p
r
ed
ictio
n
s
,
is
co
m
p
u
ted
u
s
in
g
(
2
)
:
=
TP
+
(
2
)
R
ec
all
o
r
s
en
s
itiv
ity
,
it
is
d
e
f
i
n
ed
b
y
t
h
e
p
r
o
p
o
r
tio
n
o
f
tr
u
e
p
o
s
itiv
e
ca
s
es
th
at
wer
e
ac
cu
r
ately
p
r
e
d
icted
.
I
t
is
d
eter
m
in
ed
u
s
in
g
(
3
)
:
=
TP
+
(
3
)
T
h
e
F1
-
S
co
r
e
is
a
q
u
an
titativ
e
m
ea
s
u
r
e
th
at
co
m
b
in
es
r
ec
all
an
d
ac
c
u
r
ac
y
u
s
in
g
t
h
e
ir
h
ar
m
o
n
ic
m
ea
n
,
ac
h
iev
in
g
a
b
alan
ce
d
c
o
m
p
r
o
m
is
e
b
etwe
en
th
e
two
m
etr
ics
[
2
3
]
,
it c
alc
u
lated
b
y
(
4
)
:
1
−
=
2
x
x
+
(
4
)
An
o
th
er
m
et
h
o
d
f
o
r
ev
alu
at
in
g
th
e
p
er
f
o
r
m
an
ce
o
f
clas
s
if
icatio
n
m
o
d
els
is
ca
lled
r
ec
eiv
er
o
p
er
atin
g
ch
ar
ac
ter
is
tic
(
R
OC
)
,
an
d
it
i
s
a
g
r
ap
h
ical
r
ep
r
esen
tatio
n
th
at
s
h
o
ws
h
o
w
a
b
in
ar
y
class
if
ier
s
y
s
tem
m
ay
b
e
m
ad
e
to
f
u
n
ctio
n
as
a
d
iag
n
o
s
tic
ac
r
o
s
s
d
if
f
er
en
t
th
r
esh
o
ld
s
ettin
g
s
.
T
h
e
R
O
C
cu
r
v
e
illu
s
tr
ates
th
e
tr
ad
e
-
o
f
f
b
etwe
en
th
e
class
if
ier
's
s
en
s
it
iv
ity
an
d
s
p
ec
if
icity
b
y
p
l
o
ttin
g
th
e
tr
u
e
p
o
s
itiv
e
r
ate
(
T
PR
)
v
er
s
u
s
th
e
f
alse
p
o
s
itiv
e
r
ate
(
FP
R
)
at
d
if
f
er
e
n
t
th
r
esh
o
ld
lev
els.
ar
ea
u
n
d
e
r
th
e
cu
r
v
e
(
AUC)
is
a
s
ca
le
f
r
o
m
0
to
1
,
with
a
h
ig
h
er
n
u
m
b
e
r
d
e
n
o
tin
g
t
h
e
m
o
d
el'
s
s
u
p
er
io
r
d
is
cr
im
in
atin
g
p
o
wer
[
2
4
]
.
2
.
4
.
G
ener
a
t
iv
e
a
dv
er
s
a
ri
a
l
net
wo
rk
s
-
ba
s
ed
ev
a
lua
t
io
n
Gen
er
ativ
e
ad
v
er
s
ar
ial
n
etwo
r
k
s
(
GANs)
ar
e
a
ty
p
e
o
f
ar
tific
ial
in
tellig
en
ce
alg
o
r
ith
m
s
th
at
co
m
p
r
is
e
two
n
eu
r
al
n
etwo
r
k
s
,
wh
ich
ar
e
g
en
er
ato
r
an
d
d
is
cr
im
in
ato
r
.
T
h
ese
n
etwo
r
k
s
ar
e
tr
ain
e
d
c
o
n
cu
r
r
en
tly
u
s
in
g
a
co
m
p
etitiv
e
m
eth
o
d
.
T
h
e
g
e
n
er
ato
r
m
ak
es
a
u
th
en
tic
s
y
n
th
e
tic
d
ata
s
am
p
les,
wh
ile
th
e
d
is
cr
im
in
ato
r
ac
q
u
ir
es
th
e
ab
ilit
y
to
d
if
f
e
r
en
tiate
b
et
wee
n
g
en
u
in
e
a
n
d
c
o
u
n
ter
f
eit
d
ata.
B
y
em
p
lo
y
in
g
iter
ativ
e
t
r
ain
in
g
,
GANs
h
av
e
th
e
ab
ilit
y
to
p
r
o
d
u
ce
s
y
n
th
eti
c
d
ata
o
f
ex
ce
p
tio
n
al
q
u
ality
t
h
at
n
ea
r
ly
m
ir
r
o
r
s
th
e
d
is
tr
ib
u
tio
n
s
o
f
ac
tu
al
d
ata.
T
h
is
m
ak
es
GAN
s
v
er
y
ef
f
ec
tiv
e
to
o
ls
f
o
r
m
an
y
task
s
,
in
clu
d
in
g
p
ictu
r
e
s
y
n
th
esis
,
d
ata
au
g
m
en
tatio
n
,
an
d
u
n
s
u
p
er
v
is
ed
lear
n
in
g
[
2
5
]
,
[
2
6
]
.
I
n
o
u
r
ca
s
e,
we
u
s
ed
a
g
e
n
er
ativ
e
ad
v
er
s
ar
ial
n
etwo
r
k
i
n
o
r
d
er
t
o
g
e
n
er
ate
d
ata
s
im
ilar
to
t
h
e
o
r
ig
in
al
d
a
ta
an
d
test
th
e
ef
f
ec
tiv
en
ess
o
f
o
u
r
m
o
d
el
w
h
en
c
o
n
f
r
o
n
ted
with
th
e
g
e
n
er
ated
d
ata.
T
ab
le
3
s
h
o
ws
th
e
s
tr
u
c
tu
r
e
o
f
o
u
r
g
en
e
r
ato
r
n
eu
r
al
n
etwo
r
k
wh
ile
T
ab
le
4
s
h
o
ws
th
e
s
tr
u
ctu
r
e
o
f
o
u
r
d
is
cr
im
in
ato
r
n
eu
r
al
n
etwo
r
k
.
T
ab
le
3
.
T
h
e
g
en
e
r
ato
r
s
tr
u
ctu
r
e
G
e
n
e
r
a
t
o
r
(
I
n
p
u
t
:
1
0
0
-
d
i
me
n
si
o
n
a
l
n
o
i
se
v
e
c
t
o
r
)
La
y
e
r
(
t
y
p
e
)
O
u
t
p
u
t
S
h
a
p
e
D
e
n
se
Le
a
k
y
R
e
LU
B
a
t
c
h
N
o
r
ma
l
i
z
a
t
i
o
n
D
e
n
se
Le
a
k
y
R
e
LU
B
a
t
c
h
N
o
r
ma
l
i
z
a
t
i
o
n
D
e
n
se
Le
a
k
y
R
e
LU
B
a
t
c
h
N
o
r
ma
l
i
z
a
t
i
o
n
D
e
n
se
A
c
t
i
v
a
t
i
o
n
(
Ta
n
h
)
(
N
o
n
e
,
2
5
6
)
(
N
o
n
e
,
2
5
6
)
(
N
o
n
e
,
2
5
6
)
(
N
o
n
e
,
5
1
2
)
(
N
o
n
e
,
5
1
2
)
(
N
o
n
e
,
5
1
2
)
(
N
o
n
e
,
1
0
2
4
)
(
N
o
n
e
,
1
0
2
4
)
(
N
o
n
e
,
1
0
2
4
)
(
N
o
n
e
,
1
8
)
(
N
o
n
e
,
1
8
)
T
ab
le
4
.
T
h
e
d
is
cr
im
in
ato
r
s
tr
u
ctu
r
e
D
i
scri
m
i
n
a
t
o
r
(
I
n
p
u
t
:
D
a
t
a
s
a
m
p
l
e
s w
i
t
h
1
8
c
o
l
u
m
n
s)
La
y
e
r
(
t
y
p
e
)
O
u
t
p
u
t
S
h
a
p
e
D
e
n
se
Le
a
k
y
R
e
LU
D
r
o
p
o
u
t
D
e
n
se
Le
a
k
y
R
e
LU
D
r
o
p
o
u
t
D
e
n
se
A
c
t
i
v
a
t
i
o
n
(
S
i
g
m
o
i
d
)
(
N
o
n
e
,
5
1
2
)
(
N
o
n
e
,
5
1
2
)
(
N
o
n
e
,
5
1
2
)
(
N
o
n
e
,
2
5
6
)
(
N
o
n
e
,
2
5
6
)
(
N
o
n
e
,
2
5
6
)
(
N
o
n
e
,
1
)
(
N
o
n
e
,
1
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
E
n
h
a
n
ci
n
g
ma
lw
a
r
e
d
etec
tio
n
w
ith
g
en
etic
a
lg
o
r
ith
ms a
n
d
g
en
era
tive
…
(
A
b
id
Dh
iy
a
E
d
d
i
n
e
)
3069
T
h
e
g
en
er
ato
r
tak
es
a
1
0
0
-
d
i
m
en
s
io
n
al
n
o
is
e
v
ec
to
r
as
in
p
u
t.
I
t
co
n
s
is
ts
o
f
f
u
lly
co
n
n
ec
t
ed
(
Den
s
e)
lay
er
s
f
o
llo
wed
b
y
L
ea
k
y
R
eL
U
ac
tiv
atio
n
f
u
n
ctio
n
s
an
d
B
atch
No
r
m
aliza
tio
n
lay
er
s
to
s
tab
ilize
tr
ain
in
g
.
T
h
e
f
in
al
lay
er
u
s
es
th
e
T
an
h
in
o
r
d
er
to
en
s
u
r
e
th
at
th
e
o
u
t
p
u
t
v
alu
es
ar
e
with
in
th
e
r
an
g
e
[
-
1
,
1
]
,
s
u
itab
le
f
o
r
n
o
r
m
aliza
tio
n
if
n
ec
ess
ar
y
.
T
h
e
d
is
cr
im
in
ato
r
tak
es
d
ata
s
am
p
les
with
1
8
co
lu
m
n
s
as
in
p
u
t
(
r
ea
l
o
r
s
y
n
th
etic)
.
I
t
u
s
es
a
f
u
lly
co
n
n
ec
ted
(
Den
s
e)
lay
er
s
wit
h
L
ea
k
y
R
eL
U
ac
tiv
atio
n
f
u
n
ctio
n
s
an
d
Dr
o
p
o
u
t
lay
er
s
to
av
o
id
as
m
a
x
th
e
o
v
er
f
itti
n
g
.
T
h
e
f
in
al
lay
er
u
s
es
th
e
Sig
m
o
id
ac
tiv
atio
n
f
u
n
c
tio
n
to
o
u
tp
u
t
a
p
r
o
b
ab
ilit
y
(
0
to
1
)
in
d
icatin
g
th
e
lik
elih
o
o
d
o
f
th
e
in
p
u
t
b
ein
g
r
ea
l.
T
ab
le
5
s
h
o
ws
th
e
u
s
ed
p
ar
am
eter
s
f
o
r
tr
ain
in
g
th
e
g
e
n
er
ativ
e
ad
v
er
s
ar
ial
n
etwo
r
k
.
T
ab
le
5
.
T
h
e
u
s
ed
p
a
r
am
eter
s
f
o
r
tr
ain
in
g
th
e
g
en
er
ato
r
an
d
th
e
d
is
cr
im
in
ato
r
P
a
r
a
me
t
e
r
V
a
l
u
e
Le
a
r
n
i
n
g
r
a
t
e
0
.
0
0
0
1
O
p
t
i
mi
z
e
r
A
d
a
m
Lo
ss f
u
n
c
t
i
o
n
B
i
n
a
r
y
c
r
o
ss
e
n
t
r
o
p
y
Ep
o
c
h
s
2
0
0
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
I
n
th
is
r
esear
ch
a
d
ee
p
lear
n
in
g
m
o
d
el
b
ased
o
n
g
en
eti
c
alg
o
r
ith
m
s
an
d
g
e
n
er
ativ
e
ad
v
er
s
ar
ial
n
etwo
r
k
s
was
s
u
cc
ess
f
u
lly
cr
ea
ted
an
d
test
ed
.
I
n
th
is
s
ec
tio
n
we
will
p
r
esen
t
an
d
d
is
c
u
s
s
th
e
r
esu
lts
we
g
o
tten
.
W
h
ich
in
clu
d
e
f
ea
tu
r
e
s
elec
tio
n
b
ased
o
n
g
e
n
etic
alg
o
r
ith
m
s
,
d
ata
tr
ain
in
g
u
s
in
g
d
ee
p
n
eu
r
al
n
etwo
r
k
an
d
g
e
n
er
atio
n
o
f
n
ew
d
ata
u
s
in
g
g
en
er
ativ
e
a
d
v
er
s
ar
ial
n
e
two
r
k
an
d
th
e
ev
alu
atio
n
o
f
t
h
e
m
o
d
el
u
s
in
g
th
e
g
en
er
ated
d
ata.
3
.
1
.
F
e
a
t
ure
s
elec
t
io
n
I
n
th
e
s
tag
e
o
f
f
ea
tu
r
e
s
elec
t
io
n
,
we
em
p
lo
y
ed
a
cu
s
to
m
i
ze
d
g
en
etic
al
g
o
r
ith
m
with
a
n
ad
ap
tiv
e
m
u
tatio
n
r
ate,
b
i
n
ar
y
ch
r
o
m
o
s
o
m
e
r
ep
r
esen
tatio
n
,
to
u
r
n
a
m
en
t
s
elec
tio
n
,
s
in
g
le
-
p
o
in
t
cr
o
s
s
o
v
er
,
an
d
an
elitis
m
s
tr
ateg
y
to
r
etain
th
e
b
est
s
o
lu
tio
n
f
o
r
t
h
e
n
e
x
t
g
en
er
atio
n
.
T
h
e
f
itn
ess
f
u
n
ct
io
n
ev
al
u
ated
t
h
e
class
if
icatio
n
ac
cu
r
ac
y
u
s
in
g
a
r
an
d
o
m
f
o
r
est m
o
d
el,
r
esu
ltin
g
in
th
e
s
elec
tio
n
o
f
1
8
f
ea
tu
r
es f
r
o
m
th
e
o
r
ig
in
al
5
7
,
ac
h
ie
v
in
g
a
n
ac
cu
r
ac
y
o
f
9
9
.
9
3
%.
T
h
is
ap
p
r
o
ac
h
n
o
t
o
n
l
y
s
im
p
lifie
d
th
e
m
o
d
el
b
y
r
ed
u
cin
g
its
d
im
en
s
io
n
ality
b
u
t
also
p
o
te
n
tially
im
p
r
o
v
ed
its
in
ter
p
r
e
tab
ilit
y
an
d
g
en
e
r
aliza
tio
n
.
T
h
e
h
ig
h
ac
c
u
r
ac
y
in
d
icate
s
th
at
th
e
s
elec
ted
f
ea
t
u
r
es
co
n
tain
s
ig
n
i
f
ican
t
d
is
cr
i
m
in
ato
r
y
i
n
f
o
r
m
atio
n
f
o
r
d
is
tin
g
u
is
h
in
g
b
etwe
en
m
alwa
r
e
an
d
b
en
ig
n
d
ata,
d
e
m
o
n
s
tr
atin
g
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
g
en
etic
alg
o
r
ith
m
in
id
en
tify
in
g
th
e
m
o
s
t
r
elev
an
t f
ea
tu
r
es a
n
d
en
h
an
cin
g
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
in
tr
u
s
io
n
d
etec
tio
n
s
y
s
tem
.
3
.
2
.
T
ra
ini
ng
pro
ce
s
s
Af
ter
s
elec
tin
g
th
e
o
p
tim
al
s
et
o
f
1
8
f
ea
tu
r
es
u
s
in
g
th
e
cu
s
to
m
ized
g
e
n
etic
alg
o
r
ith
m
,
we
p
r
o
ce
ed
e
d
to
tr
ain
a
d
ee
p
n
eu
r
al
n
etwo
r
k
(
DNN)
o
n
th
e
r
ed
u
ce
d
f
e
atu
r
e
s
et.
T
h
e
m
o
d
el
was
tr
ain
ed
o
v
er
m
u
ltip
le
ep
o
ch
s
,
with
th
e
h
y
p
er
p
ar
am
eter
s
m
eticu
lo
u
s
ly
tu
n
ed
to
o
p
tim
ize
p
er
f
o
r
m
a
n
ce
.
T
ab
le
6
s
h
o
ws
o
u
r
tr
ain
in
g
r
esu
lts
.
T
ab
le
6
.
T
r
ain
in
g
r
esu
lts
Ev
a
l
u
a
t
i
o
n
m
e
t
r
i
c
V
a
l
u
e
P
r
e
c
i
s
i
o
n
0
.
9
9
5
0
R
e
c
a
l
l
0
.
9
9
6
1
F1
-
S
c
o
r
e
0
.
9
9
5
5
A
c
c
u
r
a
c
y
0
.
9
9
5
6
A
U
C
S
c
o
r
e
0
.
9
9
5
5
T
h
e
m
etr
ics
m
en
tio
n
ed
i
n
T
ab
le
6
in
d
icate
a
h
ig
h
l
y
ac
cu
r
ate
m
o
d
el.
T
h
e
p
r
ec
is
io
n
an
d
r
ec
all
v
alu
es,
b
o
th
ex
ce
ed
in
g
9
9
%,
s
u
g
g
es
t
th
at
th
e
m
o
d
el
is
h
ig
h
ly
ef
f
ec
tiv
e
in
d
is
tin
g
u
is
h
in
g
b
et
wee
n
m
alwa
r
e
an
d
b
en
ig
n
s
am
p
les.
T
h
e
F1
-
Sco
r
e
,
wh
ich
b
alan
ce
s
p
r
ec
is
io
n
a
n
d
r
ec
all,
f
u
r
th
er
c
o
n
f
ir
m
s
th
e
m
o
d
el'
s
ca
p
ab
ilit
y
to
m
in
im
ize
b
o
th
f
alse
p
o
s
itiv
es
an
d
f
alse
n
eg
ativ
es.
H
o
wev
er
,
th
e
AUC
s
co
r
e
o
f
0
.
9
9
5
5
d
e
m
o
n
s
tr
ates
th
at
th
e
m
o
d
el
p
er
f
o
r
m
s
ex
ce
p
tio
n
all
y
well
ac
r
o
s
s
d
if
f
er
en
t
class
if
icatio
n
th
r
esh
o
ld
s
,
in
d
icati
n
g
s
tr
o
n
g
o
v
er
all
d
is
cr
im
in
atio
n
ab
ilit
y
.
Fig
u
r
e
3
p
r
esen
ts
th
e
co
n
f
u
s
io
n
m
atr
ix
o
f
o
u
r
m
o
d
el
wh
ich
p
r
o
v
id
es
f
u
r
th
er
in
s
ig
h
t
in
to
t
h
e
m
o
d
el'
s
p
er
f
o
r
m
an
ce
.
Ou
t
o
f
th
e
to
tal
in
s
tan
ce
s
ev
alu
ated
,
th
e
DNN
m
ad
e
o
n
ly
5
2
m
is
class
if
icati
o
n
s
,
co
m
p
r
is
in
g
2
9
f
alse p
o
s
itiv
es a
n
d
2
3
f
alse n
e
g
ativ
es.
T
h
is
lo
w
er
r
o
r
r
ate
u
n
d
er
s
co
r
es th
e
m
o
d
el'
s
r
o
b
u
s
tn
e
s
s
an
d
r
eliab
ilit
y
in
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
25
:
3
0
6
4
-
3
0
7
4
3070
d
etec
tin
g
m
alwa
r
e.
I
n
s
u
m
m
a
r
y
,
th
e
tr
ain
in
g
r
esu
lts
illu
s
tr
ate
th
at
th
e
DNN,
tr
ain
ed
o
n
t
h
e
r
ed
u
ce
d
f
ea
t
u
r
e
s
et,
ac
h
iev
es
n
ea
r
p
er
f
ec
t
cla
s
s
if
icatio
n
p
er
f
o
r
m
an
ce
.
T
h
e
h
ig
h
p
r
ec
is
io
n
,
r
ec
all,
F1
-
Sc
o
r
e,
ac
cu
r
ac
y
,
an
d
AUC
s
co
r
e
co
llectiv
ely
d
em
o
n
s
tr
ate
th
e
m
o
d
el'
s
ef
f
ec
tiv
en
ess
in
ac
cu
r
ately
id
en
tify
in
g
m
alwa
r
e
wh
ile
m
ain
tain
in
g
a
lo
w
m
is
class
if
ic
atio
n
r
ate.
Fig
u
r
e
3
.
C
o
n
f
u
s
io
n
m
atr
i
x
o
f
o
u
r
m
o
d
el
3
.
3
.
M
o
del
ev
a
lua
t
i
o
n us
ing
t
he
g
ener
a
t
ed
da
t
a
T
h
is
p
h
ase
in
v
o
lv
ed
g
en
e
r
atin
g
n
ew
d
ata,
in
clu
d
in
g
b
o
th
b
en
ig
n
an
d
m
alicio
u
s
s
am
p
les,
to
f
u
r
th
er
ev
alu
ate
th
e
ef
f
ec
tiv
e
n
ess
o
f
o
u
r
m
o
d
el.
W
e
cr
ea
ted
a
GAN
an
d
u
s
ed
it
to
g
en
e
r
ate
1
0
0
b
e
n
ig
n
an
d
1
0
0
m
alicio
u
s
s
am
p
les,
wh
ich
we
r
e
th
en
u
s
ed
t
o
test
th
e
m
o
d
e
l's
p
er
f
o
r
m
a
n
ce
.
Fig
u
r
e
4
s
h
o
ws
th
e
lo
s
s
o
f
th
e
g
en
er
ato
r
an
d
d
is
cr
im
in
ato
r
with
th
e
m
alicio
u
s
g
en
er
ated
d
ata;
h
o
wev
er
,
Fig
u
r
e
5
s
h
o
ws
th
e
lo
s
s
o
f
th
e
g
en
er
ato
r
a
n
d
d
is
cr
im
in
ato
r
with
th
e
b
en
ig
n
g
en
er
ated
d
ata.
Fig
u
r
e
6
illu
s
tr
ate
s
o
m
e
g
en
er
ated
m
alicio
u
s
d
ata
wh
ile
th
e
Fig
u
r
e
7
illu
s
tr
ate
s
o
m
e
g
en
er
ated
b
en
ig
n
d
ata.
W
h
en
th
e
g
en
er
ate
d
s
am
p
les
wer
e
p
ass
ed
to
th
e
m
o
d
el,
it
r
ec
o
g
n
ized
t
h
e
m
ali
cio
u
s
s
am
p
les
with
1
0
0
%
ac
c
u
r
ac
y
,
wh
ile
th
e
b
en
ig
n
s
am
p
l
es
wer
e
r
ec
o
g
n
ized
with
9
6
% a
cc
u
r
ac
y
.
T
h
is
m
ea
n
s
th
at
4
b
en
i
g
n
s
am
p
les we
r
e
in
co
r
r
ec
tly
class
if
ied
as m
alicio
u
s
.
Utilizin
g
a
GAN
to
g
en
er
ate
n
ew
d
ata
h
elp
s
ass
e
s
s
th
e
m
o
d
el'
s
ef
f
ec
tiv
en
ess
ag
ain
s
t
u
n
s
ee
n
d
ata.
Ach
iev
in
g
1
0
0
%
ac
cu
r
ac
y
w
ith
th
e
g
en
er
ate
d
m
alicio
u
s
s
am
p
les
d
em
o
n
s
tr
ates
th
at
th
e
GAN
ef
f
ec
tiv
ely
ca
p
tu
r
ed
th
e
u
n
d
er
l
y
in
g
d
is
tr
ib
u
tio
n
o
f
m
alicio
u
s
d
ata.
Ho
w
ev
er
,
th
e
9
6
%
ac
cu
r
ac
y
with
b
en
ig
n
d
ata
s
u
g
g
ests
s
o
m
e
r
o
o
m
f
o
r
im
p
r
o
v
e
m
en
t,
p
o
s
s
ib
ly
d
u
e
to
th
e
in
h
er
en
t
d
iv
er
s
ity
a
n
d
c
o
m
p
lex
ity
o
f
b
en
ig
n
d
ata.
T
h
is
ev
alu
atio
n
h
ig
h
lig
h
ts
th
e
m
o
d
el's
r
o
b
u
s
tn
ess
an
d
ab
ilit
y
to
g
en
er
alize
to
u
n
s
ee
n
d
ata,
ev
e
n
wh
en
th
at
d
ata
is
s
y
n
th
etica
lly
g
en
er
ated
.
Fig
u
r
e
4
.
T
h
e
lo
s
s
o
f
th
e
m
alicio
u
s
g
en
er
ated
d
ata
Fig
u
r
e
5
.
T
h
e
lo
s
s
o
f
th
e
b
en
ig
n
g
en
er
ate
d
d
ata
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
E
n
h
a
n
ci
n
g
ma
lw
a
r
e
d
etec
tio
n
w
ith
g
en
etic
a
lg
o
r
ith
ms a
n
d
g
en
era
tive
…
(
A
b
id
Dh
iy
a
E
d
d
i
n
e
)
3071
Fig
u
r
e
6
.
So
m
e
o
f
th
e
g
en
e
r
ated
m
alicio
u
s
d
ata
Fig
u
r
e
7
.
So
m
e
o
f
th
e
g
en
e
r
ated
b
en
ig
n
d
ata
4.
CO
M
P
ARI
SO
N,
L
I
M
I
T
A
T
I
O
NS A
ND
F
UT
U
RE
WO
R
K
S
As
a
s
im
p
le
co
m
p
ar
is
o
n
with
p
r
ev
io
u
s
wo
r
k
s
,
o
u
r
r
esear
c
h
is
ch
ar
ac
ter
ized
b
y
its
u
s
e
o
f
g
e
n
etic
alg
o
r
ith
m
t
o
s
elec
t
th
e
m
o
s
t
im
p
o
r
tan
t
f
ea
tu
r
es
in
ad
d
itio
n
to
r
ely
in
g
o
n
th
e
s
m
allest
p
o
s
s
ib
le
n
u
m
b
e
r
o
f
f
ea
tu
r
es
(
1
8
)
,
wh
ic
h
allo
ws
u
s
to
s
h
o
w
m
o
r
e
ac
cu
r
ate
r
esu
lts
d
u
r
in
g
th
e
tr
ain
in
g
p
r
o
ce
s
s
.
Ou
r
wo
r
k
is
also
ch
ar
ac
ter
ized
b
y
a
d
d
in
g
an
o
t
h
er
ev
alu
atio
n
s
tag
e
b
ased
o
n
t
h
e
u
s
e
o
f
g
en
e
r
ativ
e
ad
v
er
s
ar
i
al
n
etwo
r
k
s
in
o
r
d
e
r
ev
alu
ate
th
e
ef
f
ec
tiv
e
n
ess
o
f
th
e
m
o
d
el,
u
n
lik
e
o
t
h
er
wo
r
k
s
th
at
wer
e
s
atis
f
ied
with
ac
cu
r
ac
y
p
r
ec
is
io
n
,
r
ec
al
l
an
d
F1
-
S
co
r
e.
T
h
is
ca
n
co
n
f
ir
m
th
e
m
o
d
el'
s
ef
f
icien
cy
in
wo
r
k
in
g
with
n
ew
an
d
u
n
s
ee
n
d
ata.
As
f
o
r
ac
cu
r
ac
y
,
t
h
e
r
ea
s
o
n
f
o
r
its
d
ec
r
ea
s
e
co
m
p
ar
e
d
to
w
o
r
k
s
[
1
]
,
[
1
5
]
a
n
d
[
1
9
]
is
d
u
e
t
o
o
u
r
u
s
e
o
f
two
d
r
o
p
o
u
t
lay
er
s
wh
ile
tr
ain
i
n
g
o
u
r
d
ee
p
n
e
u
r
al
n
etwo
r
k
,
in
o
r
d
er
to
av
o
id
f
allin
g
in
to
a
s
tate
o
f
o
v
er
f
itti
n
g
.
T
ab
le
7
r
esu
m
e
th
e
co
m
p
ar
is
o
n
o
f
o
u
r
wo
r
k
with
th
e
s
im
ilar
wo
r
k
s
.
On
th
e
o
th
er
s
id
e,
o
u
r
d
ataset
d
o
es n
o
t r
ep
r
esen
t th
e
f
u
ll sp
ec
tr
u
m
o
f
r
ea
l
-
w
o
r
ld
m
alwa
r
e
f
o
r
th
is
,
th
e
m
o
d
el'
s
ef
f
ec
tiv
en
ess
m
ig
h
t
b
e
lim
ited
.
W
e
ca
n
s
ay
also
th
at
GANs
m
ig
h
t
n
o
t
alwa
y
s
ca
p
tu
r
e
th
e
f
u
l
l
d
iv
er
s
ity
o
f
d
ata,
as
in
d
icate
d
b
y
t
h
e
9
6
%
ac
cu
r
ac
y
with
b
en
ig
n
d
ata.
An
o
th
er
s
en
s
itiv
e
p
o
in
t
is
,
th
at
d
ee
p
n
eu
r
al
n
etwo
r
k
s
an
d
GANs
ar
e
o
f
ten
co
n
s
id
er
e
d
b
lack
b
o
x
m
o
d
els,
m
ak
i
n
g
it
d
if
f
icu
lt
to
in
ter
p
r
et
th
ei
r
d
ec
is
io
n
-
m
ak
in
g
p
r
o
ce
s
s
.
T
h
i
s
lack
o
f
tr
an
s
p
a
r
en
cy
ca
n
b
e
a
lim
itatio
n
in
u
n
d
er
s
tan
d
in
g
wh
y
ce
r
tain
d
ata
is
class
if
ied
as
m
alwa
r
e,
T
h
e
im
p
lem
en
tatio
n
o
f
e
x
p
lain
ab
l
e
AI
tech
n
iq
u
es
will
b
e
u
s
ef
u
l
f
o
r
t
h
ese
ca
s
es.
Scalin
g
th
e
m
o
d
el
to
h
a
n
d
le
lar
g
er
d
atasets
o
r
r
ea
l
-
tim
e
d
etec
tio
n
s
ce
n
ar
io
s
m
ig
h
t
p
o
s
e
a
g
o
o
d
c
h
allen
g
e.
Fu
tu
r
e
r
esear
ch
en
d
ea
v
o
r
s
co
u
ld
ad
v
an
ce
th
is
wo
r
k
b
y
f
o
c
u
s
in
g
o
n
th
e
d
ev
el
o
p
m
en
t
an
d
im
p
lem
en
tatio
n
o
f
p
r
ac
tical
to
o
ls
,
s
o
f
twar
e
s
o
lu
t
io
n
s
,
o
r
r
ea
l
-
tim
e
s
y
s
tem
s
th
at
lev
er
ag
e
th
e
p
r
o
p
o
s
ed
m
o
d
e
l
f
o
r
ef
f
ec
tiv
e
an
d
ef
f
icien
t
m
alwa
r
e
d
etec
tio
n
in
o
p
er
atio
n
al
en
v
ir
o
n
m
en
ts
with
th
e
i
n
teg
r
atio
n
o
f
e
x
p
lain
ab
l
e
AI
tech
n
iq
u
es
f
o
r
m
o
r
e
tr
an
s
p
a
r
en
t a
n
d
in
ter
p
r
etab
le
d
ec
is
io
n
-
m
ak
i
n
g
p
r
o
ce
s
s
es
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
25
:
3
0
6
4
-
3
0
7
4
3072
T
ab
le
7
.
C
o
m
p
a
r
is
o
n
o
f
o
u
r
wo
r
k
with
s
im
ilar
wo
r
k
s
Ref
D
a
t
a
s
e
t
F
e
a
t
u
r
e
sel
e
c
t
i
o
n
t
e
c
h
n
i
q
u
e
N
u
mb
e
r
o
f
f
e
a
t
u
r
e
s
Tr
a
i
n
i
n
g
t
e
c
h
n
i
q
u
e
A
c
c
u
r
a
c
y
Ev
a
l
u
a
t
i
o
n
me
t
r
i
c
s
[
1
]
C
I
C
-
M
a
l
M
e
m
-
2
0
2
2
-
52
Lo
g
i
s
t
i
c
r
e
g
r
e
ss
i
o
n
(
L
R
)
9
9
.
9
7
%
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
A
U
C
S
c
o
r
e
[
1
0
]
M
a
l
w
a
r
e
A
p
i
C
a
l
l
-
-
RF
9
6
.
9
%
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
[
1
4
]
C
I
C
-
M
a
l
M
e
m
-
2
0
2
2
-
57
RF
9
9
%
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
[
1
5
]
M
a
l
M
e
m
-
2
0
2
2
-
55
K
N
N
9
9
.
9
7
%
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
[
1
6
]
-
-
-
C
N
N
9
9
.
5
3
%
(
A
U
C
S
c
o
r
e
)
A
U
C
S
c
o
r
e
P
r
e
c
i
s
i
o
n
[
1
7
]
C
I
C
-
M
a
l
M
e
m
2
0
2
2
C
I
C
-
M
a
l
D
r
o
i
d
2
0
2
0
-
-
S
t
a
c
k
i
n
g
t
e
c
h
n
i
q
u
e
i
n
c
l
u
d
i
n
g
:
−
S
u
p
p
o
r
t
v
e
c
t
o
r
mac
h
i
n
e
(
S
V
M
)
−
C
a
t
b
o
o
s
t
−
H
i
st
o
g
r
a
m
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
−
R
a
n
d
o
m
f
o
r
e
s
t
(
R
F
)
9
8
%
9
9
.
9
9
%
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
[
1
8
]
C
I
C
-
M
a
l
M
e
m
-
2
0
2
2
-
55
RF
9
8
.
4
5
%
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
[
1
9
]
C
I
C
-
M
a
l
mem
-
2
0
2
2
-
56
C
N
N
+
B
i
-
LS
TM
9
9
.
9
6
%
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
O
u
r
r
e
se
a
r
c
h
C
I
C
-
M
a
l
M
e
m
2
0
2
2
G
e
n
e
t
i
c
a
l
g
o
r
i
t
h
m
18
DNN
9
9
.
5
6
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
A
U
C
s
c
o
r
e
S
y
n
t
h
e
t
i
c
d
a
t
a
u
s
i
n
g
G
A
N
s
5.
CO
NCLU
SI
O
N
I
n
th
is
r
esear
ch
,
we
u
tili
ze
d
a
g
en
etic
alg
o
r
ith
m
f
o
r
f
ea
tu
r
e
s
elec
tio
n
,
tr
ain
ed
a
d
ee
p
n
eu
r
al
n
etwo
r
k
o
n
th
e
s
elec
ted
f
ea
tu
r
es,
a
n
d
em
p
lo
y
ed
g
e
n
er
ativ
e
ad
v
e
r
s
ar
ial
n
etwo
r
k
s
t
o
e
v
alu
ate
th
e
e
f
f
ec
tiv
en
ess
o
f
o
u
r
m
o
d
el
ag
ain
s
t
th
e
n
ew
a
n
d
u
n
s
ee
n
d
ata
an
d
m
ak
e
it
m
o
r
e
g
en
er
alize
d
.
T
h
e
c
o
m
b
in
atio
n
o
f
f
ea
tu
r
e
s
elec
tio
n
,
DNN
tr
ain
in
g
,
an
d
d
ata
g
e
n
er
atio
n
tech
n
iq
u
es
d
em
o
n
s
tr
ates
a
h
o
lis
tic
an
d
s
o
p
h
is
ticated
ap
p
r
o
ac
h
to
m
alwa
r
e
d
etec
tio
n
.
T
h
e
h
ig
h
ac
cu
r
ac
y
an
d
p
e
r
f
o
r
m
an
ce
m
etr
ics
o
b
ta
in
ed
ac
r
o
s
s
d
if
f
er
en
t
s
tag
es
o
f
th
e
s
tu
d
y
i
n
d
icate
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
.
T
h
ese
r
esu
lts
h
av
e
s
ig
n
if
ican
t
im
p
licatio
n
s
f
o
r
c
y
b
er
s
ec
u
r
ity
,
as
th
ey
p
r
o
v
id
e
in
s
ig
h
ts
in
to
th
e
p
o
ten
tial
o
f
m
ac
h
in
e
lear
n
i
n
g
m
o
d
els
f
o
r
ac
cu
r
atel
y
d
etec
tin
g
m
alwa
r
e
an
d
d
is
tin
g
u
is
h
in
g
it
f
r
o
m
b
e
n
ig
n
d
ata.
Fu
r
th
er
m
o
r
e,
t
h
e
ap
p
lica
tio
n
o
f
GANs
to
g
e
n
er
ate
s
y
n
t
h
etic
d
ata
p
lay
ed
a
cr
u
cial
r
o
le
in
o
u
r
ev
al
u
atio
n
p
r
o
ce
s
s
.
T
h
e
GAN
-
g
e
n
er
ate
d
d
ata
m
im
ick
ed
r
ea
l
-
wo
r
ld
m
alwa
r
e
s
ce
n
ar
io
s
,
p
r
o
v
id
i
n
g
a
r
o
b
u
s
t
an
d
d
iv
er
s
e
test
in
g
en
v
ir
o
n
m
en
t.
T
h
is
allo
wed
u
s
to
r
ig
o
r
o
u
s
ly
ass
es
s
th
e
p
er
f
o
r
m
an
ce
o
f
o
u
r
m
o
d
el
ag
ain
s
t
p
r
ev
i
o
u
s
ly
u
n
s
ee
n
an
d
p
o
ten
tially
s
o
p
h
is
ticated
m
alwa
r
e
v
ar
ian
ts
.
I
n
co
n
clu
s
io
n
,
th
is
r
esear
ch
r
ep
r
esen
ts
a
co
m
p
r
eh
en
s
iv
e
an
d
im
p
ac
tf
u
l
co
n
tr
ib
u
tio
n
to
th
e
f
ield
o
f
m
alwa
r
e
d
etec
tio
n
,
s
h
o
wca
s
in
g
th
e
p
o
ten
tial
o
f
m
ac
h
in
e
lear
n
i
n
g
tech
n
i
q
u
e
s
f
o
r
ad
d
r
ess
in
g
cy
b
er
s
ec
u
r
i
ty
ch
allen
g
es.
T
h
e
im
p
r
ess
iv
e
r
esu
lts
o
b
tain
ed
ac
r
o
s
s
v
ar
io
u
s
s
tag
es
o
f
th
e
s
tu
d
y
u
n
d
er
s
co
r
e
t
h
e
ef
f
ec
ti
v
en
ess
an
d
p
r
o
m
is
e
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
.
ACK
NO
WL
E
DG
M
E
N
T
S
T
h
is
s
ec
tio
n
s
h
o
u
ld
ac
k
n
o
wle
d
g
e
in
d
iv
id
u
als
wh
o
p
r
o
v
id
ed
p
er
s
o
n
al
ass
is
tan
ce
to
th
e
w
o
r
k
b
u
t
d
o
n
o
t
m
ee
t
th
e
c
r
iter
ia
f
o
r
au
th
o
r
s
h
ip
,
d
etailin
g
th
eir
co
n
tr
ib
u
tio
n
s
.
I
t
is
im
p
er
ativ
e
to
o
b
t
ain
co
n
s
en
t
f
r
o
m
all
in
d
iv
id
u
als lis
ted
in
th
e
ac
k
n
o
wled
g
m
en
ts
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
E
n
h
a
n
ci
n
g
ma
lw
a
r
e
d
etec
tio
n
w
ith
g
en
etic
a
lg
o
r
ith
ms a
n
d
g
en
era
tive
…
(
A
b
id
Dh
iy
a
E
d
d
i
n
e
)
3073
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
A
b
id
Dh
iy
a
E
d
d
in
e
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Gh
az
li Ab
d
elk
ad
er
✓
✓
✓
✓
✓
✓
✓
✓
B
o
u
ac
h
e
Mo
u
r
a
d
✓
✓
✓
✓
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
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
in
d
in
g
s
o
f
th
is
s
tu
d
y
ar
e
o
p
en
ly
av
ailab
le
in
Kag
g
le
at
h
ttp
s
:
//w
w
w
.
ka
g
g
le.
co
m/d
a
ta
s
ets/
lu
cc
a
g
o
d
o
y/o
b
fu
s
ca
ted
-
ma
lw
a
r
e
-
mem
o
r
y
-
2022
-
cic
RE
F
E
R
E
NC
E
S
[
1
]
M
.
D
e
n
e
r
,
G
.
O
k
,
a
n
d
A
.
O
r
ma
n
,
“
M
a
l
w
a
r
e
d
e
t
e
c
t
i
o
n
u
s
i
n
g
m
e
mo
r
y
a
n
a
l
y
si
s
d
a
t
a
i
n
b
i
g
d
a
t
a
e
n
v
i
r
o
n
m
e
n
t
,
”
Ap
p
l
i
e
d
S
c
i
e
n
c
e
s
,
v
o
l
.
1
2
,
n
o
.
1
7
,
p
.
8
6
0
4
,
A
u
g
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
a
p
p
1
2
1
7
8
6
0
4
.
[
2
]
A
.
D
a
mo
d
a
r
a
n
,
F
.
D
i
Tr
o
i
a
,
C
.
A
.
V
i
sag
g
i
o
,
T.
H
.
A
u
st
i
n
,
a
n
d
M
.
S
t
a
m
p
,
“
A
c
o
mp
a
r
i
s
o
n
o
f
st
a
t
i
c
,
d
y
n
a
m
i
c
,
a
n
d
h
y
b
r
i
d
a
n
a
l
y
s
i
s
f
o
r
ma
l
w
a
r
e
d
e
t
e
c
t
i
o
n
,
”
J
o
u
r
n
a
l
o
f
C
o
m
p
u
t
e
r
Vi
r
o
l
o
g
y
a
n
d
H
a
c
k
i
n
g
T
e
c
h
n
i
q
u
e
s
,
v
o
l
.
1
3
,
n
o
.
1
,
p
p
.
1
–
1
2
,
D
e
c
.
2
0
1
5
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
4
1
6
-
0
1
5
-
0
2
6
1
-
z.
[
3
]
“
M
a
l
w
a
r
e
st
a
t
i
st
i
c
s
& t
r
e
n
d
s r
e
p
o
r
t
,
”
AV
-
T
ES
T
.
h
t
t
p
s
:
/
/
w
w
w
.
a
v
-
t
e
s
t
.
o
r
g
/
e
n
/
st
a
t
i
st
i
c
s
/
ma
l
w
a
r
e
/
(
a
c
c
e
sse
d
A
p
r
.
0
4
,
2
0
2
4
)
.
[
4
]
M
.
H
.
L.
Lo
u
k
a
n
d
B
.
A
.
Ta
m
a
,
“
Tr
e
e
-
b
a
s
e
d
c
l
a
ssi
f
i
e
r
e
n
s
e
m
b
l
e
s
f
o
r
P
E
mal
w
a
r
e
a
n
a
l
y
si
s
:
A
p
e
r
f
o
r
ma
n
c
e
r
e
v
i
si
t
,
”
A
l
g
o
r
i
t
h
m
s
,
v
o
l
.
1
5
,
n
o
.
9
,
p
.
3
3
2
,
S
e
p
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
a
1
5
0
9
0
3
3
2
.
[
5
]
K
.
S
.
R
o
y
,
T
.
A
h
me
d
,
P
.
B
.
U
d
a
s,
M
.
E.
K
a
r
i
m,
a
n
d
S
.
M
a
j
u
m
d
a
r
,
“
M
a
l
H
y
S
t
a
c
k
:
A
h
y
b
r
i
d
st
a
c
k
e
d
e
n
s
e
m
b
l
e
l
e
a
r
n
i
n
g
f
r
a
m
e
w
o
r
k
w
i
t
h
f
e
a
t
u
r
e
e
n
g
i
n
e
e
r
i
n
g
s
c
h
e
mes
f
o
r
o
b
f
u
sca
t
e
d
ma
l
w
a
r
e
a
n
a
l
y
si
s
,
”
I
n
t
e
l
l
i
g
e
n
t
S
y
st
e
m
s
w
i
t
h
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
2
0
,
p
.
2
0
0
2
8
3
,
N
o
v
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
sw
a
.
2
0
2
3
.
2
0
0
2
8
3
.
[
6
]
R
.
S
i
h
w
a
i
l
,
K
.
O
m
a
r
,
a
n
d
K
.
A
k
r
a
m
Za
i
n
o
l
A
r
i
f
f
i
n
,
“
A
n
e
f
f
e
c
t
i
v
e
m
e
m
o
r
y
a
n
a
l
y
s
i
s
f
o
r
ma
l
w
a
r
e
d
e
t
e
c
t
i
o
n
a
n
d
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
C
o
m
p
u
t
e
rs
,
M
a
t
e
r
i
a
l
s &
a
m
p
;
C
o
n
t
i
n
u
a
,
v
o
l
.
6
7
,
n
o
.
2
,
p
p
.
2
3
0
1
–
2
3
2
0
,
2
0
2
1
,
d
o
i
:
1
0
.
3
2
6
0
4
/
c
m
c
.
2
0
2
1
.
0
1
4
5
1
0
.
[
7
]
D
.
S
mi
t
h
,
S
.
K
h
o
r
s
a
n
d
r
o
o
,
a
n
d
K
.
R
o
y
,
“
L
e
v
e
r
a
g
i
n
g
f
e
a
t
u
r
e
sel
e
c
t
i
o
n
t
o
i
m
p
r
o
v
e
t
h
e
a
c
c
u
r
a
c
y
f
o
r
ma
l
w
a
r
e
d
e
t
e
c
t
i
o
n
,
”
Ju
n
.
2
0
2
3
,
d
o
i
:
1
0
.
2
1
2
0
3
/
r
s.
3
.
r
s
-
3
0
4
5
3
9
1
/
v
1
.
[
8
]
Y
.
G
u
o
,
“
A
r
e
v
i
e
w
o
f
M
a
c
h
i
n
e
L
e
a
r
n
i
n
g
-
b
a
se
d
z
e
r
o
-
d
a
y
a
t
t
a
c
k
d
e
t
e
c
t
i
o
n
:
C
h
a
l
l
e
n
g
e
s
a
n
d
f
u
t
u
r
e
d
i
r
e
c
t
i
o
n
s
,
”
C
o
m
p
u
t
e
r
C
o
m
m
u
n
i
c
a
t
i
o
n
s
,
v
o
l
.
1
9
8
,
p
p
.
1
7
5
–
1
8
5
,
Ja
n
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mc
o
m
.
2
0
2
2
.
1
1
.
0
0
1
.
[
9
]
B
a
b
a
j
i
d
e
T
o
l
u
l
o
p
e
F
a
mi
l
o
n
i
,
“
C
y
b
e
r
sec
u
r
i
t
y
c
h
a
l
l
e
n
g
e
s
i
n
t
h
e
a
g
e
o
f
A
I
:
Th
e
o
r
e
t
i
c
a
l
a
p
p
r
o
a
c
h
e
s
a
n
d
p
r
a
c
t
i
c
a
l
so
l
u
t
i
o
n
s,”
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
& IT
R
e
se
a
rc
h
J
o
u
rn
a
l
,
v
o
l
.
5
,
n
o
.
3
,
p
p
.
7
0
3
–
7
2
4
,
M
a
r
.
2
0
2
4
,
d
o
i
:
1
0
.
5
1
5
9
4
/
c
si
t
r
j
.
v
5
i
3
.
9
3
0
.
[
1
0
]
T.
R
a
j
e
sh
,
K
.
D
i
v
y
a
,
S
.
F
i
r
d
o
u
se,
S
.
A
r
e
e
b
,
a
n
d
N
.
T
h
a
k
u
r
,
“
M
a
l
w
a
r
e
d
e
t
e
c
t
i
o
n
a
n
d
p
r
e
d
i
c
t
i
o
n
s
y
st
e
m
u
si
n
g
a
d
v
a
n
c
e
d
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms,”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
f
o
r
R
e
se
a
rc
h
i
n
A
p
p
l
i
e
d
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
ri
n
g
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
1
1
,
n
o
.
6
,
p
p
.
3
7
1
8
–
3
7
2
9
,
Ju
n
.
2
0
2
3
,
d
o
i
:
1
0
.
2
2
2
1
4
/
i
j
r
a
s
e
t
.
2
0
2
3
.
5
4
1
8
1
.
[
1
1
]
D
.
H
i
t
a
j
,
G
.
P
a
g
n
o
t
t
a
,
F
.
D
e
G
a
s
p
a
r
i
,
L.
D
e
C
a
r
l
i
,
a
n
d
L
.
V
.
M
a
n
c
i
n
i
,
“
M
i
n
e
r
v
a
:
A
f
i
l
e
-
b
a
s
e
d
r
a
n
s
o
mw
a
r
e
d
e
t
e
c
t
o
r
,
”
a
rXi
v
:
2
3
0
1
.
1
1
0
5
0
,
Ja
n
.
2
0
2
3
.
[
1
2
]
S
.
R
i
t
w
i
k
a
a
n
d
K
.
B
.
R
a
j
u
,
“
M
a
l
i
c
i
o
u
s
s
o
f
t
w
a
r
e
d
e
t
e
c
t
i
o
n
a
n
d
a
n
a
l
y
z
a
t
i
o
n
u
si
n
g
t
h
e
v
a
r
i
o
u
s
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms,
”
i
n
2
0
2
2
1
3
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
C
o
m
p
u
t
i
n
g
C
o
m
m
u
n
i
c
a
t
i
o
n
a
n
d
N
e
t
w
o
r
k
i
n
g
T
e
c
h
n
o
l
o
g
i
e
s
(
I
C
C
C
N
T
)
,
O
c
t
.
2
0
2
2
,
p
p
.
1
–
7
.
d
o
i
:
1
0
.
1
1
0
9
/
i
c
c
c
n
t
5
4
8
2
7
.
2
0
2
2
.
9
9
8
4
4
0
2
.
[
1
3
]
“
M
a
l
w
a
r
e
mem
o
r
y
a
n
a
l
y
s
i
s
(
C
I
C
-
M
a
l
M
e
m
-
2
0
2
2
)
,
”
C
a
n
a
d
i
a
n
I
n
st
i
t
u
t
e
f
o
r
C
y
b
e
rs
e
c
u
ri
t
y
.
h
t
t
p
s
:
/
/
w
w
w
.
u
n
b
.
c
a
/
c
i
c
/
d
a
t
a
se
t
s/
mal
m
e
m
-
2
0
2
2
.
h
t
ml
(
a
c
c
e
sse
d
A
p
r
.
0
7
,
2
0
2
4
)
.
[
1
4
]
A
.
M
e
z
i
n
a
a
n
d
R
.
B
u
r
g
e
t
,
“
O
b
f
u
sc
a
t
e
d
ma
l
w
a
r
e
d
e
t
e
c
t
i
o
n
u
si
n
g
d
i
l
a
t
e
d
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
t
w
o
r
k
,
”
i
n
2
0
2
2
1
4
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
g
ress
o
n
U
l
t
r
a
M
o
d
e
r
n
T
e
l
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
a
n
d
C
o
n
t
r
o
l
S
y
st
e
m
s
a
n
d
W
o
rks
h
o
p
s
(
I
C
U
MT)
,
O
c
t
.
2
0
2
2
,
p
p
.
1
1
0
–
1
1
5
.
d
o
i
:
1
0
.
1
1
0
9
/
i
c
u
m
t
5
7
7
6
4
.
2
0
2
2
.
9
9
4
3
4
4
3
.
[
1
5
]
M
.
M
.
A
b
u
a
l
h
a
j
,
A
.
A
.
A
b
u
-
S
h
a
r
e
h
a
,
Q
.
Y
.
S
h
a
m
b
o
u
r
,
A
.
A
l
s
a
a
i
d
a
h
,
S
.
N
.
A
l
-
K
h
a
t
i
b
,
a
n
d
M
.
A
n
b
a
r
,
“
C
u
s
t
o
mi
z
e
d
K
-
n
e
a
r
e
s
t
n
e
i
g
h
b
o
r
s’
a
l
g
o
r
i
t
h
m
f
o
r
mal
w
a
r
e
d
e
t
e
c
t
i
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
D
a
t
a
a
n
d
N
e
t
w
o
rk
S
c
i
e
n
c
e
,
v
o
l
.
8
,
n
o
.
1
,
p
p
.
4
3
1
–
4
3
8
,
2
0
2
4
,
d
o
i
:
1
0
.
5
2
6
7
/
j
.
i
j
d
n
s.
2
0
2
3
.
9
.
0
1
2
.
[
1
6
]
S
.
S
.
V
a
n
j
i
r
e
a
n
d
M
.
La
k
s
h
mi
,
“
A
n
o
v
e
l
me
t
h
o
d
o
f
d
e
t
e
c
t
i
n
g
ma
l
w
a
r
e
o
n
A
n
d
r
o
i
d
mo
b
i
l
e
d
e
v
i
c
e
s
w
i
t
h
e
x
p
l
a
i
n
a
b
l
e
a
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
,
”
B
u
l
l
e
t
i
n
o
f
El
e
c
t
ri
c
a
l
En
g
i
n
e
e
ri
n
g
a
n
d
I
n
f
o
rm
a
t
i
c
s
,
v
o
l
.
1
3
,
n
o
.
3
,
p
p
.
2
0
1
9
–
2
0
2
6
,
J
u
n
.
2
0
2
4
,
d
o
i
:
Evaluation Warning : The document was created with Spire.PDF for Python.