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.
14
,
No
.
2
,
J
u
n
e
2
0
2
5
,
p
p
.
373
~
3
8
3
I
SS
N:
2252
-
8
8
1
4
,
DOI
:
1
0
.
1
1
5
9
1
/ijaas
.
v
14
.
i
2
.
pp
373
-
3
8
3
373
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
a
s
.
ia
esco
r
e.
co
m
Co
ntraction
con
t
ro
l
factor
-
ba
sed
g
o
rilla
t
ro
o
p optim
izer
for fe
a
tu
res in
int
rusio
n det
e
ction s
y
stems
Sh
a
lin
i S
ha
rm
a
1
,
Su
priy
a
K
ha
it
a
n
2
,
G
a
y
a
t
ri
H
eg
de
3
,
Div
y
a
Ro
ha
t
g
i
4
,
Nus
ra
t
P
a
rv
ee
n
M
o
ha
m
m
a
d
Ra
f
iqu
e
4
,
Su
ha
s
J
a
na
rda
n L
a
wa
nd
5
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
,
A
c
c
e
n
t
u
r
e
S
o
l
u
t
i
o
n
s,
N
o
i
d
a
,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
o
f
E
n
g
i
n
e
e
r
i
n
g
a
n
d
T
e
c
h
n
o
l
o
g
y
,
B
h
a
r
a
t
i
V
i
d
y
a
p
e
e
t
h
(
D
e
e
m
e
d
t
o
b
e
U
n
i
v
e
r
si
t
y
)
,
N
a
v
i
M
u
m
b
a
i
,
I
n
d
i
a
3
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
a
n
d
B
u
s
i
n
e
ss S
y
st
e
ms,
B
h
a
r
a
t
i
V
i
d
y
a
p
e
e
t
h
(
D
e
e
m
e
d
t
o
b
e
U
n
i
v
e
r
si
t
y
)
,
P
u
n
e
,
I
n
d
i
a
4
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
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
B
h
a
r
a
t
i
V
i
d
y
a
p
e
e
t
h
(
D
e
e
me
d
t
o
b
e
U
n
i
v
e
r
s
i
t
y
)
,
P
u
n
e
,
I
n
d
i
a
5
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
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
X
a
v
i
e
r
I
n
st
i
t
u
t
e
o
f
En
g
i
n
e
e
r
i
n
g
,
M
u
mb
a
i
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Oct
18
,
2
0
2
4
R
ev
is
ed
Feb
5
,
2
0
2
5
Acc
ep
ted
Mar
28
,
2
0
2
5
In
tern
e
t
o
f
th
in
g
s
(I
o
T)
h
a
s
e
v
o
l
v
e
d
in
t
o
a
larg
e
-
sc
a
le
n
e
tw
o
rk
d
u
e
t
o
t
h
e
in
c
re
a
sin
g
n
u
m
b
e
r
o
f
c
o
n
n
e
c
ted
d
e
v
ice
s
a
n
d
m
a
ss
iv
e
a
m
o
u
n
t
o
f
d
a
ta
th
e
y
g
e
n
e
ra
te.
Io
T
n
e
two
rk
s
p
ro
d
u
c
e
m
a
ss
iv
e
a
m
o
u
n
ts
o
f
h
e
tero
g
e
n
e
o
u
s
d
a
ta
fro
m
v
a
rio
u
s
d
e
v
ice
s,
m
a
k
in
g
it
d
iffi
c
u
lt
t
o
i
d
e
n
ti
f
y
re
lev
a
n
t
fe
a
tu
re
s
fo
r
in
tru
si
o
n
d
e
tec
ti
o
n
.
He
n
c
e
,
th
is
re
se
a
rc
h
p
ro
p
o
se
s
th
e
c
o
n
trac
ti
o
n
c
o
n
tr
o
l
fa
c
to
r
-
b
a
se
d
g
o
r
il
la
tro
o
p
o
p
ti
m
i
z
e
r
(CCF
-
G
TO)
fo
r
fe
a
tu
re
se
le
c
ti
o
n
a
n
d
m
u
lt
ip
le
p
a
ra
m
e
tri
c
e
x
p
o
n
e
n
ti
a
l
l
in
e
a
r
u
n
it
s
b
a
se
d
lo
n
g
sh
o
r
t
-
term
m
e
m
o
ry
(M
P
EL
U
-
LS
T
M
)
a
p
p
ro
a
c
h
fo
r
c
las
sifica
ti
o
n
o
f
i
n
tru
si
o
n
d
e
tec
ti
o
n
sy
ste
m
(IDS)
in
I
o
T.
CC
F
-
G
TO.
It
u
se
s
a
d
ju
sta
b
le
p
a
ra
m
e
ters
to
p
rio
ri
ti
z
e
re
lev
a
n
t
in
fo
rm
a
ti
o
n
w
h
il
e
e
li
m
in
a
ti
n
g
u
n
n
e
c
e
ss
a
ry
fe
a
tu
re
s,
m
a
k
in
g
t
h
e
m
o
d
e
l
m
o
re
e
fficie
n
t
a
n
d
re
su
lt
in
g
i
n
b
e
t
ter
c
las
sifica
ti
o
n
a
c
c
u
ra
c
y
.
Th
e
e
x
p
e
rime
n
tal
re
su
lt
s
d
e
m
o
n
stra
te
th
a
t
th
e
M
P
EL
U
-
LS
T
M
a
p
p
r
o
a
c
h
a
c
h
ie
v
e
s
b
e
tt
e
r
a
c
c
u
ra
c
y
o
f
9
9
.
5
6
%
o
n
t
h
e
UN
S
W
-
NB1
5
d
a
tas
e
t
a
s
c
o
m
p
a
re
d
t
o
th
e
e
a
rli
e
r
a
p
p
ro
a
c
h
e
s
l
ik
e
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
two
r
k
with
LS
T
M
(CN
N
-
LS
TM
)
a
n
d
o
p
ti
m
ize
d
d
e
e
p
re
sid
u
a
l
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
two
rk
s
(
DCRN
N).
Th
e
se
fin
d
i
n
g
s
s
u
g
g
e
st
th
a
t
t
h
e
M
P
EL
U
-
L
S
TM
m
e
th
o
d
si
g
n
ifi
c
a
n
t
ly
e
n
h
a
n
c
e
s
th
e
a
c
c
u
ra
c
y
a
n
d
ro
b
u
stn
e
ss
o
f
IDS
i
n
I
o
T
e
n
v
ir
o
n
m
e
n
ts
b
y
a
d
d
re
ss
in
g
issu
e
s
li
k
e
t
h
e
i
d
e
n
t
ifi
c
a
ti
o
n
of
re
lev
a
n
t
fe
a
tu
re
s
a
n
d
fe
a
tu
re
re
d
u
n
d
a
n
c
y
,
c
o
n
tri
b
u
t
in
g
to
m
o
re
e
ffe
c
ti
v
e
a
n
d
se
c
u
re
sy
ste
m
s.
Th
is
re
se
a
rc
h
h
a
s
v
a
lu
a
b
le
imp
li
c
a
ti
o
n
s
fo
r
e
n
h
a
n
c
in
g
th
e
se
c
u
rit
y
b
e
a
rin
g
o
f
I
o
T
in
fra
stru
c
tu
re
.
K
ey
w
o
r
d
s
:
C
o
n
tr
ac
tio
n
co
n
tr
o
l f
ac
to
r
Go
r
illa tr
o
o
p
o
p
tim
izer
I
n
ter
n
et
o
f
th
in
g
s
I
n
tr
u
s
io
n
d
etec
tio
n
s
y
s
tem
L
o
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
Mu
ltip
le
p
ar
am
etr
ic
ex
p
o
n
e
n
tial lin
ea
r
u
n
it
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
:
Sh
alin
i Sh
ar
m
a
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
,
Acc
en
tu
r
e
So
l
u
tio
n
s
DDC4
A
SeaVie
w
T
o
wer
Sect
o
r
1
3
5
No
id
a
,
I
n
d
ia
E
m
ail:
s
h
alin
is
h
ar
m
a1
9
8
0
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
I
n
ter
n
et
o
f
th
in
g
s
(
I
o
T
)
i
n
teg
r
ates
v
ar
io
u
s
em
b
e
d
d
ed
c
o
m
p
u
tin
g
d
e
v
ices
with
in
th
e
p
h
y
s
ical
en
v
ir
o
n
m
en
t,
e
n
ab
lin
g
s
ea
m
less
co
n
n
ec
tiv
ity
b
etwe
en
d
ev
ices,
s
er
v
ices,
an
d
s
y
s
tem
s
.
T
h
is
s
ea
m
less
co
n
n
ec
tiv
ity
h
as b
ec
o
m
e
an
in
teg
r
al
p
ar
t o
f
d
aily
life
,
s
u
p
p
o
r
tin
g
a
b
r
o
ad
r
an
g
e
o
f
ap
p
licati
o
n
s
an
d
im
p
r
o
v
in
g
o
v
er
all
ef
f
icien
c
y
[
1
]
,
[
2
]
.
I
o
T
n
etwo
r
k
s
co
n
tin
u
o
us
ly
in
cl
u
d
e
a
g
r
e
at
er
am
o
u
n
t o
f
h
eter
o
g
e
n
eo
u
s
d
ev
ices w
ith
in
d
iv
id
u
al
co
m
m
u
n
icatio
n
p
r
o
to
co
ls
an
d
d
ata
s
tan
d
ar
d
s
.
W
ith
th
e
f
ast
g
r
o
wth
o
f
th
e
I
n
ter
n
et,
n
etwo
r
k
in
tr
u
s
io
n
ac
tiv
ities
h
ap
p
en
co
n
tin
u
o
u
s
ly
,
a
n
d
c
o
n
v
e
n
tio
n
al
s
ec
u
r
it
y
ap
p
r
o
ac
h
es
ar
e
t
y
p
icall
y
f
o
cu
s
ed
o
n
clo
u
d
s
er
v
ices
[
3
]
.
T
h
ese
ce
n
tr
alize
d
s
ec
u
r
ity
m
ea
s
u
r
es
ar
e
co
m
p
lex
to
m
ee
t
th
e
f
ast
-
d
ev
el
o
p
in
g
r
e
q
u
ir
em
e
n
ts
o
f
n
etwo
r
k
attac
k
s
a
n
d
in
tr
u
s
io
n
s
.
E
n
s
u
r
in
g
th
at
s
ec
u
r
ity
an
d
p
r
i
v
ac
y
r
em
ain
s
ig
n
i
f
ican
t
ch
allen
g
es
f
o
r
I
o
T
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.
14
,
No
.
2
,
J
u
n
e
2
0
2
5
:
3
7
3
-
383
374
d
ev
ices
d
u
e
to
th
eir
lim
it
ed
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
,
m
ak
in
g
c
o
n
v
e
n
tio
n
al
s
ec
u
r
ity
m
ec
h
a
n
is
m
s
in
ap
p
r
o
p
r
iate
f
o
r
th
ese
s
y
s
tem
s
[
4
]
.
T
h
is
m
a
k
es
th
em
v
u
l
n
er
ab
le
t
o
a
b
r
o
ad
r
an
g
e
o
f
c
y
b
er
attac
k
s
s
u
ch
as
d
en
ial
o
f
s
er
v
ice
(
Do
S)
,
d
is
tr
ib
u
ted
Do
S
(
DDo
S)
,
an
d
s
p
o
o
f
in
g
[
5
]
.
T
o
i
d
e
n
t
i
f
y
t
h
e
s
e
m
a
l
i
c
i
o
u
s
a
t
t
a
c
k
s
,
a
n
e
t
w
o
r
k
i
n
t
r
u
s
i
o
n
d
e
t
e
c
ti
o
n
s
y
s
t
e
m
(
N
I
D
S
)
is
i
m
p
o
r
t
a
n
t
t
o
c
o
n
t
r
o
l
t
h
e
o
p
e
r
a
t
i
o
n
s
o
f
t
h
e
I
o
T
n
e
t
w
o
r
k
s
.
N
I
DS
h
as
b
e
e
n
d
e
s
i
g
n
e
d
t
o
d
e
t
e
ct
t
h
e
ad
v
a
n
c
e
d
n
e
t
w
o
r
k
s
,
c
o
m
p
r
i
s
e
d
o
f
I
o
T
n
e
t
w
o
r
k
s
.
T
h
r
o
u
g
h
e
x
a
m
i
n
i
n
g
t
h
e
h
i
n
t
s
o
f
n
e
t
w
o
r
k
i
n
t
r
u
s
i
o
n
s
,
N
I
D
S
h
as
d
e
t
e
c
t
e
d
t
h
e
a
t
t
a
c
k
s
a
n
d
r
a
i
s
e
d
t
h
e
a
l
a
r
m
s
i
n
r
e
a
l
-
ti
m
e
[
6
]
,
[
7
]
.
C
o
n
v
e
n
t
i
o
n
a
l
N
I
D
S
a
p
p
r
o
a
c
h
e
s
h
a
v
e
d
i
f
f
i
c
u
lt
y
i
n
m
a
n
a
g
i
n
g
t
h
e
c
o
m
p
l
e
x
i
t
y
a
n
d
d
i
v
e
r
s
i
t
y
o
f
t
h
e
I
o
T
n
e
t
w
o
r
k
t
r
a
f
f
i
c
,
d
e
s
i
g
n
i
n
g
i
t
c
o
m
p
l
e
x
t
o
d
et
e
r
m
i
n
e
a
b
n
o
r
m
a
l
a
ct
i
v
i
ti
e
s
,
p
a
r
ti
c
u
l
a
r
l
y
i
n
I
o
T
d
e
v
i
c
e
s
.
M
o
r
e
o
v
e
r
,
t
h
e
c
o
n
v
e
n
t
i
o
n
a
l
a
p
p
r
o
a
c
h
es
s
u
f
f
e
r
f
r
o
m
d
r
a
w
b
a
c
k
s
s
u
c
h
a
s
m
a
x
i
m
u
m
f
a
l
s
e
a
l
a
r
m
r
at
e
a
n
d
m
i
n
i
m
u
m
d
e
t
e
c
ti
o
n
r
a
t
e
[
8
]
.
V
a
r
i
o
u
s
r
es
e
a
r
c
h
e
r
s
h
a
v
e
c
o
o
p
e
r
a
t
e
d
o
n
t
h
e
d
e
v
e
l
o
p
m
e
n
t
o
f
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
s
y
s
t
e
m
s
(
I
D
S
)
,
l
e
v
e
r
a
g
i
n
g
t
h
e
p
o
w
e
r
o
f
a
r
t
i
f
i
c
i
a
l
i
n
t
e
l
li
g
e
n
c
e
(
A
I
)
a
p
p
r
o
a
c
h
e
s
[
9
]
,
[
1
0
]
.
M
a
c
h
i
n
e
l
ea
r
n
i
n
g
(
M
L
)
i
s
a
k
i
n
d
o
f
i
n
te
r
d
i
s
ci
p
l
i
n
a
r
y
c
r
o
s
s
-
f
u
n
c
t
i
o
n
a
l
a
r
e
a
t
h
a
t
e
m
u
la
t
es
h
u
m
a
n
i
n
t
e
l
li
g
e
n
c
e
.
H
o
w
e
v
e
r
,
t
h
e
a
b
s
e
n
c
e
o
f
r
e
d
u
n
d
a
n
t
f
e
a
t
u
r
e
s
h
a
m
p
e
r
s
m
a
c
h
i
n
e
-
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
e
s
i
n
e
f
f
e
c
ti
v
e
l
y
d
e
t
e
ct
i
n
g
a
n
d
a
d
d
r
e
s
s
i
n
g
n
o
v
e
l
a
t
t
a
ck
s
i
n
t
h
e
c
u
r
r
e
n
t
I
o
T
n
e
tw
o
r
k
[
1
1
]
–
[
1
3
]
.
T
h
e
NI
DS
was
d
ev
elo
p
ed
u
tili
zin
g
d
ee
p
lear
n
in
g
(
DL
)
ap
p
r
o
ac
h
es,
b
en
ch
m
ar
k
d
atasets
ar
e
p
r
ef
er
r
ed
to
m
ax
im
ize
th
e
d
etec
tio
n
o
f
in
tr
u
s
io
n
s
.
DL
-
b
ased
NI
DS
ar
e
ty
p
ically
tr
ain
ed
b
y
u
tili
zin
g
th
e
r
ec
en
t
d
atasets
d
ev
elo
p
e
d
f
o
r
in
tr
u
s
io
n
d
etec
tio
n
.
T
h
e
e
f
f
ec
tiv
en
ess
o
f
NI
DS
u
s
in
g
DL
ap
p
r
o
ac
h
es
o
f
ten
im
p
r
o
v
es
as
th
e
n
u
m
b
er
o
f
f
ea
tu
r
es
i
n
th
e
d
ata
s
et
en
h
an
ce
s
.
Fu
r
th
e
r
m
o
r
e
,
D
L
m
eth
o
d
s
m
in
im
ize
th
e
s
ize
o
f
th
e
f
ea
tu
r
e
v
ec
to
r
in
to
an
i
d
ea
l
n
u
m
b
er
o
f
ess
e
n
tial
f
ea
tu
r
es
[
1
4
]
,
[
1
5
]
.
I
DS
h
av
e
b
ee
n
b
r
o
a
d
ly
u
tili
ze
d
in
v
ar
io
u
s
s
tu
d
ies
b
ec
au
s
e
o
f
th
eir
co
m
p
lex
ity
i
n
p
r
o
tectin
g
c
o
m
p
u
ter
n
etwo
r
k
s
f
r
o
m
cy
b
er
th
r
ea
ts
.
T
h
is
r
e
s
ea
r
ch
s
u
m
m
ar
izes
DL
ap
p
r
o
ac
h
es
u
s
ed
in
th
e
ex
is
tin
g
wo
r
k
s
f
o
r
th
e
d
esig
n
o
f
I
DS.
Altu
n
a
y
a
n
d
Alb
a
y
r
ak
[
1
6
]
d
ev
elo
p
ed
th
e
th
r
ee
v
a
r
io
u
s
DL
ap
p
r
o
ac
h
es
c
o
n
v
o
lu
ti
o
n
al
n
eu
r
al
n
etwo
r
k
(
C
NN
)
,
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
L
STM
)
,
an
d
th
e
h
y
b
r
id
m
eth
o
d
o
f
C
NN+
L
STM
f
o
r
th
e
I
DS
in
in
d
u
s
tr
ial
I
o
T
n
etwo
r
k
s
.
I
n
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
s
tep
,
th
e
m
is
s
in
g
v
alu
es
wer
e
s
o
lv
ed
an
d
th
e
m
in
-
m
ax
n
o
r
m
aliza
tio
n
s
te
p
was
p
er
f
o
r
m
ed
to
e
n
h
an
c
e
th
e
class
if
icatio
n
p
er
f
o
r
m
an
ce
.
T
h
e
m
i
n
-
m
ax
n
o
r
m
aliza
tio
n
tech
n
iq
u
e
h
ad
t
h
e
b
en
ef
it
o
f
ac
co
m
p
an
y
in
g
all
d
ata
co
n
n
ec
tio
n
s
ef
f
icien
tly
.
Ho
wev
e
r
,
th
e
lack
o
f
a
n
ef
f
ec
tiv
e
f
ea
t
u
r
e
s
elec
tio
n
p
r
o
ce
s
s
to
id
e
n
tify
k
ey
attac
k
f
ea
tu
r
es
le
d
t
o
p
o
o
r
class
if
icatio
n
p
er
f
o
r
m
an
c
e.
Hn
am
te
et
a
l
.
[
1
7
]
in
tr
o
d
u
c
ed
th
e
two
-
s
tag
e
DL
a
p
p
r
o
ac
h
th
r
o
u
g
h
th
e
h
y
b
r
id
m
eth
o
d
o
f
L
STM
an
d
au
t
o
en
c
o
d
er
(
AE
)
f
o
r
th
e
d
etec
tio
n
o
f
I
DS.
T
h
e
d
ata
f
r
o
m
th
e
L
STM
-
AE
ap
p
r
o
ac
h
h
ad
b
ee
n
f
ilter
ed
with
r
esp
ec
t
to
s
o
lv
in
g
th
e
o
v
er
-
f
itti
n
g
an
d
u
n
d
er
-
f
itti
n
g
p
r
o
b
lem
s
.
T
h
e
L
STM
-
AE
ap
p
r
o
ac
h
ef
f
ec
tiv
ely
b
alan
ce
d
th
e
d
im
e
n
s
io
n
ality
r
ed
u
ctio
n
an
d
th
e
f
ea
tu
r
e
r
eten
tio
n
in
th
e
h
ig
h
ly
b
alan
ce
d
d
atasets
.
Fu
r
th
er
m
o
r
e
,
th
e
L
STM
-
AE
ap
p
r
o
ac
h
is
e
f
f
ec
tiv
e
in
id
en
tify
in
g
s
ig
n
if
ica
n
t
an
o
m
alies
in
n
etwo
r
k
tr
af
f
ic,
wh
ich
ca
n
b
e
i
n
d
icativ
e
o
f
f
u
t
u
r
e
cy
b
e
r
-
attac
k
s
.
Ho
wev
er
,
th
e
s
elec
tio
n
o
f
im
p
o
r
tan
t
n
et
wo
r
k
attac
k
f
ea
tu
r
es
f
r
o
m
th
e
r
aw
d
ata
is
im
p
o
r
tan
t
to
attain
b
etter
r
esu
lts
.
Ku
m
ar
et
a
l
.
[
1
8
]
p
r
esen
ted
t
h
e
d
ee
p
r
esid
u
al
co
n
v
o
l
u
tio
n
a
l
n
eu
r
al
n
etwo
r
k
(
DC
R
NN)
f
o
r
s
ec
u
r
ity
en
h
an
ce
m
e
n
t
in
I
DS,
wh
ic
h
was
f
in
e
-
tu
n
e
d
th
r
o
u
g
h
a
n
im
p
r
o
v
e
d
g
az
elle
o
p
tim
izatio
n
a
lg
o
r
ith
m
(
I
GOA)
.
A
n
o
v
el
b
in
a
r
y
GOA
(
NB
GOA)
was u
s
ed
f
o
r
th
e
f
ea
tu
r
e
s
elec
tio
n
p
r
o
ce
d
u
r
e
to
r
em
o
v
e
r
ed
u
n
d
an
t f
ea
tu
r
es f
r
o
m
d
ata
u
tili
ze
d
i
n
th
e
h
in
d
r
an
ce
class
if
icatio
n
p
r
o
ce
d
u
r
es.
Ho
wev
er
,
NB
GOA
wa
s
co
m
p
lex
with
f
ea
tu
r
e
r
eten
tio
n
in
im
b
alan
ce
d
d
atasets
,
n
eg
ativ
ely
im
p
ac
tin
g
th
e
I
DS
ac
cu
r
ac
y
an
d
g
e
n
er
aliza
tio
n
.
W
an
g
et
a
l
.
[
1
9
]
d
ev
elo
p
e
d
th
e
R
esNet,
tr
an
s
f
o
r
m
er
,
an
d
b
i
d
ir
ec
tio
n
al
L
STM
(
B
iLST
M)
ap
p
r
o
ac
h
f
o
r
th
e
I
DS,
wh
ich
to
o
k
o
u
t
b
o
th
s
p
atial
an
d
tem
p
o
r
al
f
e
atu
r
es
o
f
th
e
n
etwo
r
k
tr
af
f
ic
.
T
h
e
s
p
atial
f
ea
tu
r
e
ex
t
r
ac
tio
n
ap
p
r
o
ac
h
was
estab
lis
h
ed
th
r
o
u
g
h
R
esNet
an
d
th
e
tem
p
o
r
al
f
ea
t
u
r
e
ex
tr
ac
ti
o
n
ap
p
r
o
ac
h
was e
s
tab
lis
h
ed
th
r
o
u
g
h
B
iLST
M
to
ex
tr
ac
t
th
e
f
ea
tu
r
es.
E
v
en
t
u
ally
,
s
p
atio
tem
p
o
r
al
f
ea
tu
r
es
w
er
e
in
v
o
l
v
ed
to
attain
th
e
att
ac
k
d
etec
tio
n
a
n
d
class
if
icatio
n
.
Ho
wev
er
,
id
e
n
tify
in
g
a
p
p
r
o
p
r
iate
clea
n
i
n
g
an
d
p
r
ep
r
o
ce
s
s
in
g
m
eth
o
d
s
f
o
r
th
e
p
r
ev
ailin
g
n
etwo
r
k
tr
af
f
ic
d
ata
is
cr
u
cia
l
f
o
r
ef
f
ec
tiv
ely
tr
ain
i
n
g
an
d
test
in
g
th
e
class
if
ier
o
n
ac
tu
al
n
etwo
r
k
tr
af
f
ic
.
Halb
o
u
n
i
et
a
l
.
[
2
0
]
in
tr
o
d
u
ce
d
th
e
s
tack
ed
C
NN
an
d
L
ST
M
ap
p
r
o
ac
h
es
b
ased
o
n
b
atch
n
o
r
m
aliza
tio
n
(
B
N)
an
d
d
r
o
p
o
u
t
la
y
er
s
f
o
r
th
e
I
D
S.
T
h
e
C
NN
co
u
l
d
ex
tr
ac
t
th
e
s
p
atial
f
ea
tu
r
es
an
d
L
STM
ex
t
r
ac
ted
th
e
tem
p
o
r
al
f
ea
tu
r
es
to
d
esig
n
th
e
h
y
b
r
i
d
I
DS
ap
p
r
o
ac
h
.
T
h
e
C
NN
an
d
L
STM
ef
f
ec
tiv
ely
s
o
lv
ed
th
e
o
v
er
f
itti
n
g
t
h
r
o
u
g
h
th
e
m
in
im
izatio
n
o
f
s
o
m
e
t
r
ai
n
ab
le
p
ar
am
ete
r
s
an
d
to
e
n
h
an
ce
th
e
g
en
er
aliza
tio
n
.
Ho
wev
e
r
,
th
e
s
tack
ed
C
NN
an
d
L
STM
ap
p
r
o
ac
h
with
B
N
led
to
m
in
im
iz
in
g
th
e
in
ter
p
r
etab
ilit
y
in
I
DS
ap
p
licatio
n
s
.
Fro
m
th
is
o
v
er
v
iew,
v
ar
io
u
s
lim
itatio
n
s
h
a
v
e
b
ee
n
id
en
tifie
d
:
th
e
lack
o
f
a
f
e
atu
r
e
s
elec
t
io
n
p
r
o
ce
s
s
,
ch
all
en
g
es
with
f
ea
tu
r
e
r
eten
tio
n
in
im
b
alan
ce
d
d
atasets
,
an
d
m
in
im
ized
in
te
r
p
r
eta
b
ilit
y
.
Ho
wev
er
,
I
o
T
n
etwo
r
k
s
p
r
o
d
u
ce
m
ass
iv
e
am
o
u
n
ts
o
f
h
eter
o
g
en
eo
u
s
d
ata
f
r
o
m
v
ar
io
u
s
d
ev
ices,
m
ak
in
g
it
d
if
f
icu
lt
to
id
en
tify
r
elev
an
t
f
ea
tu
r
es
f
o
r
in
tr
u
s
io
n
d
etec
tio
n
.
T
o
o
v
e
r
c
o
m
e
th
is
p
r
o
b
lem
,
t
h
is
r
esear
ch
p
r
o
p
o
s
es
th
e
co
n
tr
ac
tio
n
c
o
n
tr
o
l
f
ac
t
o
r
-
b
ased
g
o
r
illa
tr
o
o
p
o
p
tim
izer
(
C
C
F
-
GT
O)
f
o
r
f
ea
t
u
r
e
s
elec
tio
n
an
d
m
u
ltip
le
p
a
r
am
etr
ic
e
x
p
o
n
en
tial
lin
ea
r
u
n
it
(
MPE
L
U
)
b
ased
L
STM
(
-
)
ap
p
r
o
ac
h
f
o
r
th
e
class
if
icatio
n
o
f
I
DS
in
I
o
T
.
T
o
o
v
e
r
co
m
e
th
ese
p
r
o
b
lem
s
,
th
is
r
esear
ch
p
r
o
p
o
s
es
th
e
C
C
F
-
GT
O
f
o
r
s
elec
tin
g
r
elev
an
t
f
ea
tu
r
es,
al
o
n
g
with
th
e
MPE
L
U
-
L
STM
f
o
r
class
if
y
in
g
I
DS.
MPE
L
U
en
a
b
les
L
STM
to
a
d
ap
tiv
ely
lear
n
m
o
r
e
ef
f
ec
tiv
ely
o
v
er
lo
n
g
s
eq
u
en
ce
s
,
wh
ich
is
im
p
o
r
tan
t
f
o
r
d
etec
tin
g
d
is
tin
ctiv
e
p
atter
n
s
in
tim
e
-
s
er
ies
d
ata
with
in
I
DS.
T
h
e
p
r
im
ar
y
h
ig
h
lig
h
ts
o
f
th
is
r
esear
ch
ar
e
as
tr
ails
:
i)
A
r
o
b
u
s
t
p
r
e
-
p
r
o
ce
s
s
in
g
s
tep
is
p
er
f
o
r
m
ed
by
in
teg
r
atin
g
th
e
d
ata
clea
n
in
g
an
d
s
o
lv
in
g
m
is
s
in
g
v
alu
es.
T
h
en
,
th
e
m
i
n
-
m
ax
n
o
r
m
aliza
ti
o
n
is
p
er
f
o
r
m
ed
t
o
en
s
u
r
e
u
n
if
o
r
m
f
ea
tu
r
e
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
C
o
n
tr
a
ctio
n
co
n
tr
o
l fa
cto
r
-
b
a
s
ed
g
o
r
illa
tr
o
o
p
o
p
timiz
er fo
r
fea
tu
r
es in
in
tr
u
s
io
n
…
(
S
h
a
lin
i S
h
a
r
ma
)
375
co
n
tr
ib
u
tio
n
.
T
h
ese
tech
n
iq
u
e
s
en
h
an
ce
th
e
d
ataset
q
u
ality
,
r
eliab
ilit
y
,
an
d
class
if
ier
ef
f
ec
tiv
en
ess
;
ii)
T
h
e
C
C
F
is
p
r
o
p
o
s
ed
in
to
th
e
GT
O
ap
p
r
o
ac
h
f
o
r
ef
f
ec
tiv
ely
d
et
er
m
in
in
g
th
e
a
p
p
r
o
p
r
iate
f
e
atu
r
es
wh
ile
r
em
o
vi
n
g
r
ed
u
n
d
an
cies.
T
h
is
o
p
tim
izatio
n
ap
p
r
o
ac
h
e
n
h
an
ce
s
th
e
cla
s
s
if
icatio
n
ac
cu
r
ac
y
;
a
n
d
iii)
T
h
e
MPE
L
U
-
L
STM
ap
p
r
o
ac
h
in
teg
r
ates
th
e
s
eq
u
e
n
ce
lear
n
in
g
ca
p
ab
ilit
ies
o
f
L
S
T
M
with
th
e
non
-
lin
ea
r
f
lex
i
b
ilit
y
o
f
th
e
MPE
L
U
ac
tiv
atio
n
f
u
n
cti
o
n
f
o
r
th
e
cl
ass
if
icatio
n
o
f
n
etwo
r
k
attac
k
s
.
T
h
is
in
teg
r
atio
n
en
h
an
ce
s
in
tr
u
s
io
n
d
etec
tio
n
ef
f
ec
tiv
en
ess
f
o
r
b
o
th
s
h
o
r
t a
n
d
lo
n
g
-
ter
m
s
eq
u
e
n
ce
s
,
r
ea
ch
i
n
g
s
u
p
er
io
r
d
etec
tio
n
r
ates
(
D
R
)
.
T
h
is
r
esear
ch
is
o
r
g
an
ized
a
s
f
o
llo
ws:
s
ec
t
io
n
2
o
u
tlin
es
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
.
Sectio
n
3
p
r
esen
ts
th
e
f
ea
t
u
r
e
s
e
lectio
n
u
s
in
g
C
C
F
-
GT
O.
Sectio
n
4
d
em
o
n
s
tr
ates
th
e
r
esu
lts
an
d
d
is
cu
s
s
io
n
.
Sectio
n
5
co
n
clu
d
es a
n
o
v
er
all
r
esear
c
h
.
2.
P
RO
P
O
SE
D
M
E
T
H
O
DO
L
O
G
Y
T
h
is
r
esear
ch
p
r
o
p
o
s
es
th
e
ef
f
ec
tiv
e
C
C
F
-
GT
O
f
o
r
s
elec
t
in
g
th
e
r
elev
an
t
f
ea
tu
r
es
an
d
MPE
L
U
-
L
STM
f
o
r
t
h
e
class
if
icatio
n
o
f
I
DS
in
I
o
T
s
y
s
tem
s
.
T
h
is
r
esear
ch
in
cl
u
d
es
f
o
u
r
s
ig
n
i
f
ican
t
p
h
ases
:
d
ata
co
llectio
n
,
p
r
e
-
p
r
o
ce
s
s
in
g
,
f
ea
tu
r
e
s
elec
tio
n
,
an
d
class
if
icati
o
n
.
L
STM
ca
n
r
ec
o
g
n
ize
p
atte
r
n
s
an
d
co
r
r
elatio
n
s
in
th
e
tim
e
-
s
er
ies
d
ata
o
f
n
etwo
r
k
tr
af
f
ic
o
r
d
ev
ice
b
eh
a
v
i
o
r
.
T
h
r
o
u
g
h
lea
r
n
in
g
th
ese
p
at
ter
n
s
,
th
e
L
STM
h
as
d
is
tin
g
u
is
h
ed
b
etwe
en
n
o
r
m
al
an
d
v
ar
i
o
u
s
ty
p
es
o
f
attac
k
s
.
Fig
u
r
e
1
d
eter
m
i
n
es
th
e
wo
r
k
i
n
g
o
f
t
h
e
p
r
o
p
o
s
ed
m
eth
o
d
.
Fig
u
r
e
1
.
W
o
r
k
i
n
g
o
n
th
e
p
r
o
p
o
s
ed
m
eth
o
d
2
.
1
.
Da
t
a
c
o
llect
io
n
T
h
e
p
r
im
ar
y
f
o
c
u
s
o
f
th
is
r
esear
ch
is
d
ata
co
llectio
n
,
wh
ich
is
co
n
s
id
er
ed
to
esti
m
ate
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
.
T
h
e
im
p
o
r
tan
ce
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
i
n
th
is
r
esear
ch
is
id
e
n
tifie
d
u
s
in
g
th
e
two
b
e
n
ch
m
ar
k
d
atasets
s
u
ch
as
UNSW
-
N
B
1
5
[
2
1
]
a
n
d
C
I
C
-
I
DS
-
2
0
1
7
[
2
2
]
.
T
h
e
d
etailed
d
escr
ip
tio
n
o
f
th
ese
d
atasets
is
d
escr
ib
ed
as f
o
llo
ws.
2
.
1
.
1
.
UNSW
-
NB
1
5
T
h
is
d
ataset
in
v
o
lv
es
r
ec
o
r
d
s
o
f
b
en
i
g
n
tr
af
f
ic
alo
n
g
with
n
in
e
ty
p
es
o
f
attac
k
s
,
in
clu
d
in
g
an
aly
s
is
,
Do
S,
b
ac
k
d
o
o
r
,
a
n
d
f
u
zz
es
.
I
t
was
p
r
o
d
u
ce
d
th
r
o
u
g
h
t
h
e
Au
s
tr
alian
C
en
tr
e
f
o
r
C
y
b
er
Secu
r
ity
(
AC
C
S)
in
2
0
1
5
.
T
h
e
r
ec
o
r
d
s
ar
e
ac
q
u
i
r
e
d
f
r
o
m
v
ar
i
o
u
s
r
ea
l
-
wo
r
l
d
we
b
s
ites
,
in
v
o
lv
in
g
S
y
m
a
n
tec
C
o
r
p
o
r
ati
o
n
,
c
o
m
m
o
n
v
u
ln
er
ab
ilit
ies
an
d
ex
p
o
s
u
r
es
(
C
VE
)
,
an
d
Mic
r
o
s
o
f
t
C
o
r
p
o
r
atio
n
.
T
ab
le
1
r
ep
r
esen
ts
th
e
n
u
m
b
er
o
f
s
am
p
les
o
f
th
e
UNSW
-
NB
1
5
d
ataset.
T
ab
le
1
.
Nu
m
b
er
o
f
s
am
p
les o
f
th
e
UNSW
-
NB
1
5
d
ataset
A
t
t
a
c
k
s
N
u
mb
e
r
o
f
sam
p
l
e
s
B
e
n
i
g
n
56
,
000
G
e
n
e
r
i
c
40
,
000
W
o
r
ms
1
3
0
D
o
S
12
,
264
A
n
a
l
y
s
i
s
2
,
0
0
0
Ex
p
l
o
i
t
s
33
,
393
B
a
c
k
d
o
o
r
1
,
7
4
6
R
e
c
o
n
n
a
i
ss
a
n
c
e
10
,
491
F
u
z
z
e
r
s
18
,
184
S
h
e
l
l
c
o
d
e
1
,
1
3
3
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.
14
,
No
.
2
,
J
u
n
e
2
0
2
5
:
3
7
3
-
383
376
2
.
1
.
2
.
CIC
-
I
DS
-
2017
T
h
e
C
I
C
-
I
DS
-
2
0
1
7
d
ataset
is
en
v
is
io
n
ed
to
s
o
lv
e
th
e
lack
o
f
a
r
ea
l
-
tim
e
n
etwo
r
k
tr
af
f
ic
d
ataset
f
o
r
th
e
esti
m
atio
n
o
f
I
DS.
T
h
is
d
ataset
co
m
p
r
is
es
o
f
m
o
s
t
p
r
esen
t
an
d
a
p
p
r
o
p
r
iate
d
ata
f
o
r
t
esti
n
g
th
e
s
ec
u
r
ity
s
y
s
tem
s
.
T
h
is
d
ataset
in
v
o
lv
es
2
,
8
3
0
,
0
0
0
s
am
p
les,
o
f
wh
ic
h
,
1
9
.
7
0
%
ar
e
attac
k
s
an
d
th
e
r
em
ain
in
g
8
0
.
3
0
%
ar
e
b
en
ig
n
.
T
h
er
e
ar
e
1
4
v
a
r
io
u
s
ass
au
lt
ty
p
es
an
d
1
n
o
r
m
al
class
.
T
h
e
8
4
f
ea
tu
r
es
ar
e
r
ec
o
v
er
ed
f
r
o
m
a
d
ev
elo
p
e
d
n
etwo
r
k
tr
af
f
ic
an
d
th
e
f
in
al
co
lu
m
n
o
f
th
e
d
ataset
in
v
o
lv
es th
e
m
u
lticlas
s
lab
el.
T
ab
le
2
r
ep
r
esen
ts
th
e
n
u
m
b
er
o
f
s
am
p
les o
f
th
e
C
I
C
-
I
DS
-
2
0
1
7
d
ataset.
T
ab
le
2
.
Nu
m
b
er
o
f
s
am
p
les o
f
C
I
C
-
I
DS
-
2017
A
t
t
a
c
k
s
N
u
mb
e
r
o
f
sam
p
l
e
s
B
e
n
i
g
n
2
,
2
6
0
,
3
6
0
B
o
t
1
,
9
4
3
D
D
o
S
1
2
7
,
0
8
2
D
o
S
G
o
l
d
e
n
Ey
e
10
,
289
D
o
S
H
u
l
k
2
2
9
,
1
9
8
F
TP
-
P
a
r
a
t
o
r
7
,
8
9
4
D
o
S
sl
o
w
l
o
r
i
s
5
,
7
7
1
D
o
s Sl
o
w
h
t
t
p
t
e
s
t
5
,
4
8
5
I
n
f
i
l
t
r
a
t
i
o
n
34
H
e
a
r
t
b
l
e
e
d
11
P
o
r
t
S
c
a
n
1
5
7
,
7
0
3
SSH
-
P
a
r
a
t
o
r
5
,
8
6
1
W
e
b
A
t
t
a
c
k
-
X
S
S
6
4
8
W
e
b
A
t
t
a
ck
-
B
r
u
t
e
F
o
r
c
e
1
,
4
9
7
W
e
b
A
t
t
a
c
k
-
S
QL
I
n
j
e
c
t
i
o
n
21
2
.
2
.
P
re
-
pro
ce
s
s
ing
T
h
e
in
p
u
t
f
r
o
m
th
e
UNSW
-
NB
1
5
an
d
C
I
C
-
I
DS
-
2
0
1
7
d
at
asets
is
p
r
o
v
id
ed
f
o
r
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
s
tep
.
Her
e,
d
ata
clea
n
in
g
an
d
m
in
-
m
ax
n
o
r
m
aliza
tio
n
ar
e
p
e
r
f
o
r
m
ed
t
o
en
h
an
ce
th
e
class
if
icatio
n
p
er
f
o
r
m
an
ce
.
T
h
e
d
etailed
in
f
o
r
m
atio
n
o
f
th
ese
tech
n
i
q
u
es i
s
d
escr
ib
ed
in
th
e
f
o
llo
win
g
.
2
.
2
.
1
.
Da
t
a
clea
nin
g
Data
clea
n
in
g
is
th
e
p
r
o
ce
d
u
r
e
o
f
d
eter
m
in
in
g
o
r
elim
in
atin
g
er
r
o
r
s
,
ir
r
eg
u
lar
ities
,
an
d
d
is
cr
ep
an
cies
in
d
ata
b
ef
o
r
e
it
is
u
tili
z
ed
f
o
r
m
o
d
elin
g
.
I
t
is
an
im
p
o
r
tan
t
s
tep
in
d
ata
p
r
e
p
r
o
ce
s
s
in
g
,
p
ar
ticu
lar
ly
f
o
r
DL
ap
p
licatio
n
s
.
T
h
e
co
llected
d
a
tasets
co
n
tain
th
e
n
u
m
b
er
o
f
m
is
s
in
g
v
alu
es
in
a
f
ew
f
ea
tu
r
e
co
lu
m
n
s
.
I
n
th
is
p
h
ase,
all
u
n
f
illed
ce
lls
in
a
f
e
atu
r
e
co
lu
m
n
a
r
e
o
cc
u
p
ied
with
“0
”.
E
v
er
y
ca
teg
o
r
ical
v
al
u
e
is
d
em
o
n
s
tr
ated
as
a
p
ar
ticu
lar
n
u
m
er
ical
v
alu
e
an
d
an
alter
atio
n
p
r
o
ce
d
u
r
e
i
s
em
p
lo
y
ed
.
I
n
th
is
p
h
ase,
th
e
d
ata
th
at
co
n
tain
s
m
is
s
in
g
v
alu
es a
r
e
r
em
o
v
ed
,
a
n
d
th
en
,
th
e
m
in
-
m
ax
n
o
r
m
ali
za
tio
n
is
p
er
f
o
r
m
ed
.
2
.
2
.
2
.
M
in
-
m
a
x
no
r
m
a
l
iza
t
io
n
T
h
e
m
in
-
m
a
x
n
o
r
m
aliza
tio
n
t
ec
h
n
iq
u
e
is
p
er
f
o
r
m
e
d
to
s
u
p
p
o
r
t
th
e
d
e
v
elo
p
m
e
n
t
o
f
n
e
u
r
al
n
etwo
r
k
s
m
o
s
t
d
ep
e
n
d
ab
ly
.
T
h
is
ap
p
r
o
a
ch
h
as
t
h
e
b
en
ef
it
o
f
p
er
f
o
r
m
i
n
g
all
d
ata
c
o
n
n
ec
tio
n
s
ef
f
ec
ti
v
ely
.
Ho
we
v
er
,
th
e
f
ea
tu
r
e
v
alu
es a
r
e
p
r
o
v
id
ed
i
n
a
r
an
g
e
b
etwe
e
n
0
a
n
d
1
in
d
iv
i
d
u
ally
[
2
3
]
.
T
h
is
ap
p
r
o
ac
h
is
e
x
p
r
ess
ed
in
(
1
)
.
=
−
−
(
1
)
W
h
er
e
d
em
o
n
s
tr
ates
th
e
n
o
r
m
alize
d
d
ata;
illu
s
tr
ates
th
e
ac
tu
al
v
alu
e
o
f
th
e
f
ea
tu
r
e;
a
n
d
an
d
th
e
f
ea
tu
r
e’
s
m
ax
im
u
m
an
d
m
ax
im
u
m
v
al
u
es.
T
h
is
ap
p
r
o
ac
h
s
u
p
p
o
r
ts
m
ak
in
g
s
u
r
e
th
a
t
all
f
ea
tu
r
es
p
er
f
o
r
m
u
n
if
o
r
m
ly
to
th
e
lear
n
in
g
p
r
o
ce
s
s
o
f
th
e
m
o
d
el.
T
h
en
,
n
o
r
m
alize
d
d
ata
a
r
e
p
r
o
v
id
e
d
f
o
r
th
e
f
u
r
th
er
p
r
o
ce
s
s
.
3.
F
E
AT
U
RE
S
E
L
E
C
T
I
O
N
U
SI
NG
CCF
-
G
T
O
Af
ter
d
ata
p
r
e
-
p
r
o
ce
s
s
in
g
,
th
e
ap
p
r
o
p
r
iate
f
ea
t
u
r
es
ar
e
s
elec
ted
b
y
u
tili
zin
g
th
e
m
et
a
-
h
eu
r
is
tic
o
p
tim
izatio
n
alg
o
r
ith
m
.
T
h
is
p
r
o
ce
s
s
s
u
p
p
o
r
ts
a
class
if
ier
to
en
h
an
ce
class
if
icatio
n
p
er
f
o
r
m
an
ce
.
GT
O
alg
o
r
ith
m
is
p
r
o
p
o
s
ed
f
o
r
th
e
f
ea
tu
r
e
s
elec
tio
n
p
r
o
ce
s
s
.
T
h
e
C
C
F
-
GT
O
m
eth
o
d
en
h
an
ce
s
th
e
co
n
v
en
tio
n
al
GT
O
b
y
in
teg
r
atin
g
a
n
ad
a
p
tiv
e
m
ec
h
a
n
is
m
th
at
f
in
e
-
t
u
n
es
th
e
b
alan
ce
b
etwe
en
ex
p
lo
r
ati
o
n
an
d
ex
p
lo
itatio
n
,
r
esu
ltin
g
in
im
p
r
o
v
e
d
p
er
f
o
r
m
an
ce
in
id
en
tify
in
g
o
p
tim
a
l
o
r
n
ea
r
-
o
p
tim
al
s
o
lu
tio
n
s
.
GT
O
is
a
n
atu
r
e
-
en
co
u
r
a
g
ed
ap
p
r
o
ac
h
th
at
p
r
eten
d
s
a
s
o
cial
b
eh
av
io
r
o
f
th
e
g
o
r
illas
.
T
h
is
alg
o
r
ith
m
is
in
s
p
ir
ed
b
y
th
e
n
atu
r
al
in
tellig
en
ce
o
f
th
e
g
o
r
illas
.
GT
O
co
m
p
r
is
es
two
s
ig
n
if
ican
t
p
h
ases
s
u
ch
as
ex
p
lo
r
atio
n
an
d
e
x
p
lo
itatio
n
.
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
C
o
n
tr
a
ctio
n
co
n
tr
o
l fa
cto
r
-
b
a
s
ed
g
o
r
illa
tr
o
o
p
o
p
timiz
er fo
r
fea
tu
r
es in
in
tr
u
s
io
n
…
(
S
h
a
lin
i S
h
a
r
ma
)
377
Var
io
u
s
o
p
er
ativ
es
co
m
p
ete
with
o
p
tim
izatio
n
o
p
er
atio
n
s
f
o
r
th
e
g
o
r
illa’
s
b
eh
av
i
o
r
in
th
is
ap
p
r
o
ac
h
.
Du
r
in
g
th
e
ex
p
lo
r
atio
n
p
h
ase,
th
r
ee
o
p
er
atio
n
s
ar
e
tak
en
o
u
t
s
u
ch
as
s
h
if
tin
g
to
an
u
n
ex
p
lo
r
e
d
p
o
s
itio
n
,
s
h
if
tin
g
to
war
d
o
th
e
r
g
o
r
illas
,
an
d
m
o
v
in
g
to
a
f
a
m
iliar
lo
ca
tio
n
.
Fo
llo
win
g
a
s
ilv
er
b
ac
k
as
well
a
s
s
tr
iv
in
g
with
ad
u
lt
f
em
ales a
r
e
ass
u
m
ed
to
en
h
an
ce
s
ea
r
ch
ef
f
ec
tiv
en
ess
[
2
4
]
,
[
2
5
]
.
C
o
m
m
u
n
icatio
n
am
o
n
g
th
e
s
i
lv
er
b
ac
k
s
an
d
o
th
er
g
o
r
illas
is
a
s
ig
n
if
ican
t
p
ar
t
o
f
d
ec
is
io
n
-
m
ak
in
g
.
Hen
ce
,
to
en
h
a
n
ce
th
e
e
x
p
lo
r
a
tio
n
ca
p
ab
il
ity
o
f
GT
O,
th
is
r
e
s
ea
r
ch
p
r
o
p
o
s
es th
e
C
C
F s
tr
ateg
y
to
s
im
u
late
th
is
ass
o
ciatio
n
.
T
o
s
o
lv
e
th
e
lo
ca
l
o
p
tim
u
m
p
r
o
b
lem
,
th
is
r
esear
ch
m
o
v
es
a
s
in
g
le
s
o
lu
tio
n
to
a
p
o
s
itio
n
o
f
an
o
th
er
o
p
tim
al
s
o
lu
tio
n
,
h
e
n
ce
th
at
th
e
lo
ca
l
s
p
ac
e
i
s
ef
f
ec
tiv
ely
ex
p
lo
r
e
d
.
T
h
e
C
C
F
im
p
r
o
v
es
th
e
ex
p
lo
r
atio
n
an
d
ex
p
lo
itatio
n
ca
p
ab
ilit
ies
o
f
th
e
GT
O,
allo
win
g
it
to
e
f
f
ec
tiv
ely
s
elec
t
th
e
m
o
s
t
r
elev
a
n
t
f
ea
tu
r
es
f
o
r
I
DS,
r
ed
u
ci
n
g
r
e
d
u
n
d
a
n
cy
.
T
h
r
o
u
g
h
s
elec
tin
g
o
n
ly
th
e
m
o
s
t
in
f
o
r
m
ativ
e
f
ea
tu
r
es,
C
C
F
-
GT
O
m
in
im
izes
th
e
d
im
e
n
s
io
n
ality
o
f
th
e
d
ataset,
lead
in
g
to
f
aster
tr
ain
in
g
an
d
in
f
er
en
ce
p
r
o
ce
s
s
es.
Fu
r
th
er
m
o
r
e,
r
an
d
o
m
n
ess
is
u
tili
ze
d
to
m
o
v
e
th
e
s
o
lu
tio
n
to
an
ex
p
l
o
r
atio
n
ar
ea
th
at
is
n
o
n
-
o
b
tain
ed
th
r
o
u
g
h
an
ap
p
r
o
ac
h
.
T
h
is
s
tr
ateg
y
ap
p
r
o
ac
h
c
o
n
tai
n
s
th
e
ca
p
ab
ilit
y
to
tak
e
awa
y
o
f
lo
ca
l
o
p
tim
al
s
o
lu
tio
n
,
h
en
ce
it
o
b
tain
s
th
e
m
o
s
t
r
ea
lis
tic
an
d
s
ig
n
if
ican
t
s
o
lu
tio
n
.
Hen
ce
,
s
im
u
lates
an
ar
b
itra
r
y
m
o
v
em
en
t
p
r
o
ce
d
u
r
e
o
f
a
g
o
r
illa
to
en
h
an
ce
a
s
o
lu
tio
n
’
s
q
u
al
ity
.
p
r
eten
d
s
th
e
d
e
g
r
ee
o
f
ex
p
e
r
tis
e
co
n
tr
o
lled
th
r
o
u
g
h
th
e
g
o
r
il
las
in
th
is
p
h
ase.
T
h
er
e
ar
e
b
in
ar
y
ty
p
es
o
f
ex
p
lo
r
atio
n
o
f
u
n
i
d
en
tifie
d
ar
ea
s
.
I
f
|
|
≥
0
.
5
,
th
en
th
e
b
eh
av
io
r
o
f
th
e
g
o
r
illas
to
ex
p
l
o
r
e
u
n
id
en
tifie
d
ar
ea
s
i
s
s
im
u
lated
ef
f
ec
tiv
ely
b
as
ed
o
n
th
eir
p
er
c
ep
tio
n
.
I
n
th
e
in
itializatio
n
s
tag
e,
th
e
GT
O
ap
p
r
o
ac
h
r
a
n
d
o
m
l
y
g
en
er
ates
th
e
p
o
p
u
latio
n
f
r
o
m
th
e
n
o
r
m
alize
d
f
ea
t
u
r
es
an
d
th
e
p
o
s
itio
n
o
f
th
e
s
ilv
er
b
ac
k
g
o
r
illa.
An
o
th
er
f
o
r
m
o
f
ex
p
lo
r
atio
n
in
v
o
lv
es
n
a
v
ig
atin
g
u
n
f
a
m
iliar
p
o
s
itio
n
s
b
ased
o
n
t
h
e
g
o
r
illas
'
co
n
v
er
s
atio
n
al
ex
p
e
r
ien
ce
s
with
o
n
e
an
o
th
er
,
aim
in
g
to
r
ed
u
ce
th
e
b
lin
d
n
atu
r
e
o
f
ex
p
lo
r
at
io
n
.
T
h
e
p
ar
am
ete
r
C
C
F
m
ain
tain
s
h
o
w
th
e
g
o
r
illas
s
elec
t
am
o
n
g
th
ese
s
tr
ateg
ies,
s
ig
n
if
ican
tly
ex
ten
d
in
g
th
e
g
o
r
illa’
s
ex
p
lo
r
atio
n
o
f
u
n
f
am
iliar
ar
e
as
an
d
en
h
an
cin
g
th
e
s
ea
r
ch
s
p
ac
e
o
f
t
h
e
alg
o
r
ith
m
.
T
h
e
g
o
r
illas
s
till
n
ee
d
t
o
co
n
n
ec
t
to
en
s
u
r
e
c
o
n
s
is
ten
t
ex
p
er
ien
ce
s
an
d
m
i
n
im
ize
th
e
lim
itatio
n
s
o
f
th
is
r
esear
ch
.
T
h
e
s
p
ec
if
ic
u
p
d
ate
is
ex
p
r
ess
ed
in
(
2
)
.
=
−
_
−
(
2
)
W
h
er
e
d
em
o
n
s
tr
ates
a
f
itn
ess
(
ac
cu
r
ac
y
)
v
al
u
e
o
f
th
g
o
r
illa
s
;
_
d
em
o
n
s
tr
ates
th
e
f
itn
ess
v
alu
e
o
f
s
ilv
er
b
ac
k
;
an
d
illu
s
t
r
ates a
v
er
ag
e
f
itn
ess
v
alu
e
o
f
wh
o
le
g
o
r
illas
.
W
h
en
>
1
,
th
e
u
p
d
ate
(
3
)
an
d
(
4
)
.
W
h
er
e
d
em
o
n
s
tr
ates
th
e
r
an
d
o
m
n
u
m
b
e
r
am
o
n
g
0
an
d
1
;
(
1
,
)
illu
s
tr
ates
th
e
r
an
d
o
m
v
ec
to
r
b
y
th
e
d
im
e
n
s
io
n
r
a
n
g
in
g
b
etwe
en
0
an
d
1
th
r
o
u
g
h
u
n
c
h
an
g
i
n
g
d
is
tr
ib
u
tio
n
;
illu
s
tr
ates
th
e
d
im
en
s
io
n
ality
is
s
u
e;
d
ep
icts
th
e
ar
b
itra
r
y
g
o
r
illa in
d
iv
i
d
u
al
;
an
d
illu
s
tr
ate
s
th
e
ar
b
itra
r
y
v
ec
to
r
with
th
e
p
r
o
b
lem
d
im
en
s
io
n
p
r
o
d
u
ce
d
in
an
i
n
ter
v
al
[
1
−
|
|
,
|
|
]
with
th
e
s
tatic
d
is
tr
ib
u
tio
n
.
An
esti
m
atio
n
o
f
is
f
o
r
m
u
lated
in
(
5
)
.
=
[
(
−
)
×
(
|
|
−
)
×
(
1
,
)
]
2
+
,
|
|
≥
0
.
5
(
3
)
=
(
−
)
×
,
|
|
<
0
.
5
(
4
)
=
(
[
|
|
,
|
|
,
1
,
]
)
(
5
)
Similar
ly
,
to
e
n
h
an
ce
th
e
p
er
f
o
r
m
an
ce
o
f
ex
p
lo
r
atio
n
wh
en
≤
1
,
th
e
c
u
r
r
e
n
t
g
o
r
illa
in
d
i
v
id
u
als
ar
e
f
u
s
ed
with
r
a
n
d
o
m
l
y
s
elec
ted
g
o
r
illa
in
d
iv
id
u
als.
T
h
is
p
r
o
ce
s
s
in
cr
ea
s
es
ex
p
er
im
en
ta
l
p
ar
am
eter
s
wh
ile
in
teg
r
atin
g
th
e
i
n
f
lu
en
ce
o
f
t
h
e
cu
r
r
en
t
g
o
r
illa in
d
i
v
id
u
als.
A
p
o
s
itio
n
u
p
d
ate
is
f
o
r
m
u
late
d
in
(
6
)
.
=
2
+
(
−
2
)
×
+
[
/
(
×
(
1
,
)
)
]
×
(
1
−
)
(
6
)
T
h
is
p
h
ase
s
ig
n
if
ican
tly
m
o
v
ed
a
s
o
lu
tio
n
o
f
t
h
e
p
r
ese
n
t
in
d
iv
id
u
al
to
an
a
r
b
itra
r
y
in
d
iv
id
u
al
s
o
lu
tio
n
.
A
p
ar
am
ete
r
d
ef
in
es
a
s
m
all
r
an
g
e
o
f
m
o
v
em
en
t,
allo
win
g
f
o
r
s
ig
n
if
ican
t
ex
p
lo
r
atio
n
o
f
t
h
e
lo
ca
l
s
p
ac
e
b
etwe
en
two
s
o
lu
tio
n
s
to
id
en
tif
y
th
e
b
est
s
o
lu
t
i
o
n
.
T
h
is
im
p
r
o
v
es
th
e
ca
p
ab
ilit
y
to
e
x
p
lo
r
e
s
ig
n
if
ican
tly
wh
ile
b
r
o
ad
ly
el
im
in
atin
g
b
lin
d
s
ea
r
ch
es.
T
h
e
n
,
th
e
s
elec
ted
f
ea
tu
r
es
ar
e
f
ed
in
to
th
e
f
u
r
th
e
r
p
r
o
ce
s
s
.
T
h
e
p
ar
am
eter
s
o
f
th
e
p
r
o
p
o
s
ed
C
C
F
-
GT
O
ap
p
r
o
ac
h
in
clu
d
e
a
p
o
p
u
latio
n
s
ize
r
an
g
in
g
f
r
o
m
5
0
to
1
0
0
,
a
n
u
m
b
er
o
f
iter
atio
n
s
b
e
twee
n
1
0
0
an
d
5
0
0
,
a
C
C
F
b
etwe
en
0
.
1
an
d
1
.
0
,
an
d
an
ex
p
lo
r
atio
n
p
ar
am
eter
th
at
is
ad
ju
s
ted
ac
co
r
d
in
g
to
th
e
C
C
F v
alu
e.
3
.
1
.
Cla
s
s
if
ica
t
io
n
T
h
e
s
elec
ted
f
ea
tu
r
es
f
r
o
m
t
h
e
in
p
u
t
d
ata
ar
e
p
r
o
v
id
ed
a
s
in
p
u
t
to
th
e
class
if
icatio
n
p
r
o
ce
s
s
to
class
if
y
th
e
d
ata
in
to
two
ca
t
eg
o
r
ies
n
o
r
m
al
a
n
d
attac
k
.
T
h
e
MPE
L
U
-
L
STM
ac
tiv
atio
n
f
u
n
cti
o
n
i
n
tr
o
d
u
ce
s
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.
14
,
No
.
2
,
J
u
n
e
2
0
2
5
:
3
7
3
-
383
378
m
u
ltip
le
p
ar
am
eter
s
w
h
ich
all
o
ws
th
e
L
STM
n
etwo
r
k
to
ca
p
tu
r
e
co
m
p
lex
n
o
n
-
lin
ea
r
r
elatio
n
s
h
ip
s
with
in
th
e
d
ata
m
o
r
e
ef
f
e
ctiv
ely
.
T
h
is
c
ap
ab
ili
ty
is
im
p
o
r
tan
t
f
o
r
d
et
ec
tin
g
in
tr
icate
p
atter
n
s
in
n
e
two
r
k
tr
af
f
ic
wh
ich
r
ep
r
esen
ts
p
o
s
s
ib
le
in
tr
u
s
io
n
s
.
A
d
etailed
ex
p
lan
atio
n
o
f
t
h
e
MPE
L
U
-
L
STM
is
p
r
o
v
id
ed
as
f
o
llo
ws.
T
h
e
p
r
im
ar
y
p
u
r
p
o
s
e
o
f
L
STM
is
to
s
o
lv
e
lo
n
g
-
ter
m
d
ep
en
d
en
cy
p
r
o
b
lem
s
.
I
n
th
e
co
n
v
en
t
io
n
al
L
STM
ap
p
r
o
ac
h
,
th
er
e
ar
e
4
lay
er
s
:
2
-
in
p
u
t,
1
f
o
r
g
o
t
,
an
d
1
o
u
tp
u
t
g
ate.
T
h
e
in
p
u
t
g
ates
wo
r
k
co
m
b
in
ed
to
ch
o
o
s
e
an
in
p
u
t
th
at
is
ex
ten
d
ed
to
t
h
eir
s
tate.
Acc
o
r
d
in
g
to
th
e
p
r
esen
t
ce
ll
s
tate,
a
f
o
r
g
et
g
ate
id
e
n
tifie
s
wh
ich
p
ast
ce
ll
s
tate
s
s
h
o
u
ld
b
e
d
is
ca
r
d
ed
.
T
h
e
n
,
an
o
u
t
p
u
t
g
ate
d
ec
id
es
wh
ich
d
ata
will
b
e
tr
an
s
m
itted
b
y
th
ese
g
ates.
A
m
em
o
r
y
ce
ll
u
n
it
is
d
esig
n
ed
with
in
p
u
t,
o
u
t
p
u
t
,
an
d
f
o
r
g
o
t
g
ates
u
tili
ze
d
to
ef
f
ec
tiv
ely
ev
o
k
e
an
d
f
o
r
g
et
in
p
u
t
d
ata.
On
ce
a
n
in
p
u
t
is
s
en
t
th
r
o
u
g
h
a
m
e
m
o
r
y
ce
ll
u
n
it,
d
ata
is
s
ig
n
if
ic
an
tly
f
o
r
g
o
tten
as
well
as
s
to
r
ed
.
On
ce
an
in
p
u
t
is
ex
p
r
ess
ed
a
s
=
(
1
,
2
,
…
,
−
1
,
)
an
d
o
u
tp
u
t
i
s
(
1
,
2
,
…
,
−
1
,
)
,
th
e
g
ates a
r
e
ex
p
r
ess
ed
in
(
7
)
.
(
)
=
(
+
)
(
7
)
T
he
i
n
i
t
i
a
l
p
h
a
s
e
i
n
t
h
e
L
S
T
M
p
r
o
c
e
d
u
r
e
i
s
t
o
s
e
n
d
b
y
f
o
r
g
o
t
t
e
n
g
a
t
e
.
D
a
t
a
i
n
a
m
e
m
o
r
y
u
n
i
t
o
f
a
n
e
a
r
l
i
e
r
c
e
l
l
i
s
i
d
e
n
t
i
f
i
e
d
w
h
e
n
s
e
n
t
b
y
t
h
i
s
g
a
t
e
i
f
i
s
s
e
n
t
t
o
f
u
r
t
h
e
r
p
r
o
g
r
e
s
s
i
o
n
o
r
r
e
j
e
c
t
e
d
.
A
f
o
r
g
o
t
g
a
t
e
i
s
f
o
r
m
u
l
a
t
e
d
i
n
(
8
)
.
=
(
.
[
−
1
,
]
+
)
(
8
)
W
h
er
e
d
en
o
tes
th
e
weig
h
t
m
atr
ix
o
f
f
o
r
g
o
t
g
ate
an
d
d
em
o
n
s
tr
ates
a
b
ias.
An
o
th
er
p
h
ase
is
to
u
p
d
ate
th
e
d
ata
b
y
a
c
o
n
s
titu
en
t
in
p
u
t
g
ate
ex
ten
d
ed
to
a
m
em
o
r
y
u
n
it.
I
n
th
is
p
r
o
ce
d
u
r
e,
a
v
alu
e
to
b
e
u
p
d
ated
is
id
en
tifie
d
th
r
o
u
g
h
a
s
ig
m
o
i
d
f
u
n
ctio
n
.
Mo
r
eo
v
e
r
,
a
p
r
o
b
ab
l
e
r
eg
en
er
atio
n
v
ec
to
r
ce
ll
s
tate
is
p
r
o
d
u
ce
d
in
ℎ
lay
er
.
T
h
e
in
p
u
t
an
d
ce
ll st
at
e
is
f
o
r
m
u
lated
i
n
(
9
)
an
d
(
1
0
)
.
=
(
.
[
−
1
,
]
+
)
(
9
)
=
.
−
1
+
.
ℎ
(
.
[
−
1
,
]
+
)
(
1
0
)
W
h
er
e
th
e
v
alu
e
o
f
as
th
e
v
ec
to
r
is
ac
q
u
ir
ed
f
r
o
m
[
0
,
1
]
;
,
[
−
1
,
]
an
d
as
lear
n
ed
p
ar
a
m
eter
s
ar
e
u
tili
ze
d
in
in
p
u
t
g
ates;
d
em
o
n
s
tr
ates
th
e
ce
ll
s
tate
m
atr
ix
weig
h
t;
an
d
d
em
o
n
s
tr
ates
b
ias.
I
n
th
is
p
r
o
ce
d
u
r
e,
th
e
u
p
d
ate
o
f
th
e
ce
ll
s
tate
is
r
estru
c
tu
r
ed
af
te
r
d
eter
m
in
in
g
wh
ich
p
o
r
tio
n
s
o
f
th
e
d
ata
ar
e
r
etain
ed
an
d
wh
ic
h
ar
e
d
is
ca
r
d
ed
.
An
o
u
tp
u
t
g
ate
in
th
is
p
r
o
ce
d
u
r
e
will
id
en
tify
r
ec
en
t
u
p
d
ate
d
ata
in
a
ce
ll,
th
en
it
will
b
e
h
an
d
led
as
L
STM
o
u
tp
u
t.
An
o
u
tp
u
t
g
ate
is
esti
m
ated
in
th
e
f
i
n
al
p
h
as
e
o
f
th
e
L
STM
p
r
o
ce
d
u
r
e
th
r
o
u
g
h
a
s
ig
m
o
id
f
u
n
ctio
n
b
y
r
em
o
te
weig
h
t
m
atr
ix
o
u
tp
u
t
g
ate
r
e
p
r
esen
te
d
th
r
o
u
g
h
an
d
[
−
1
,
]
an
d
d
en
o
tin
g
a
b
ias.
I
n
th
is
p
r
o
ce
d
u
r
e,
an
o
u
tp
u
t
is
ac
q
u
ir
ed
f
r
o
m
m
u
ltip
lied
an
d
ℎ
o
u
tp
u
t
is
co
n
s
eq
u
e
n
tial w
ith
o
u
tp
u
t
is
f
o
r
m
u
lated
in
(
1
1
)
an
d
(
1
2
)
.
=
(
0
.
[
−
1
,
]
+
)
(
1
1
)
=
.
ℎ
(
)
(
1
2
)
MPE
L
U
is
an
ac
tiv
atio
n
f
u
n
ct
io
n
th
at
p
r
o
p
o
s
es
to
s
im
p
lify
an
d
u
n
if
y
a
r
ec
tifie
d
lin
ea
r
u
n
it
(
R
eL
U)
an
d
E
L
U.
T
h
e
s
ig
n
if
ica
n
t
aim
is
to
b
etter
class
if
icatio
n
ef
f
ec
tiv
en
ess
.
MPE
L
U
is
ca
p
ab
le
o
f
f
lex
i
b
ly
c
h
an
g
i
n
g
am
o
n
g
th
e
R
eL
U
an
d
E
L
U,
m
ak
in
g
h
y
p
er
p
ar
a
m
ete
r
lear
n
a
b
le
to
f
u
r
th
er
e
n
h
an
ce
its
r
ea
lis
tic
ca
p
ab
ilit
y
an
d
tu
n
e
th
e
f
u
n
ctio
n
s
h
a
p
e.
MPE
L
U
is
f
o
r
m
u
lated
in
(
1
3
)
.
(
)
=
{
>
0
(
−
1
)
≤
0
(
1
3
)
W
h
er
e
>
0
.
MPE
L
U
allo
ws
th
e
L
STM
to
ad
ap
t
to
v
ar
io
u
s
ty
p
es
o
f
d
ata,
en
a
b
lin
g
it
to
ca
p
tu
r
e
co
m
p
lex
,
non
-
lin
ea
r
p
atter
n
s
m
o
r
e
ef
f
e
ctiv
ely
.
MPE
L
U
in
teg
r
at
es
th
e
b
en
ef
its
o
f
p
a
r
am
etr
ic
ac
tiv
atio
n
f
u
n
ctio
n
s
with
th
e
ad
v
an
ta
g
es
o
f
ex
p
o
n
e
n
ti
al
lin
ea
r
u
n
its
,
p
r
o
v
id
in
g
a
f
lex
ib
le
an
d
ef
f
ec
tiv
e
ac
tiv
at
io
n
f
u
n
ctio
n
.
T
h
is
f
lex
ib
ilit
y
im
p
r
o
v
es
th
e
L
ST
M
ca
p
ab
ilit
y
to
lear
n
in
tr
icate
p
atter
n
s
in
th
e
d
ata,
lea
d
in
g
t
o
th
e
m
o
s
t
ef
f
icien
t
class
if
icatio
n
.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
is
im
p
l
em
en
ted
o
n
Py
t
h
o
n
3
.
1
0
.
1
2
s
o
f
twar
e
to
o
ls
an
d
s
y
s
tem
s
p
ec
if
icatio
n
with
W
in
d
o
ws
1
0
6
4
-
b
it
OS,
I
n
tel
co
r
e
i5
p
r
o
ce
s
s
o
r
,
an
d
1
6
GB
R
AM
.
T
h
e
p
r
o
p
o
s
ed
I
DS
class
if
ica
tio
n
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
C
o
n
tr
a
ctio
n
co
n
tr
o
l fa
cto
r
-
b
a
s
ed
g
o
r
illa
tr
o
o
p
o
p
timiz
er fo
r
fea
tu
r
es in
in
tr
u
s
io
n
…
(
S
h
a
lin
i S
h
a
r
ma
)
379
ap
p
r
o
ac
h
u
tili
ze
s
v
a
r
io
u
s
p
er
f
o
r
m
an
ce
m
etr
ics
to
v
alid
ate
th
e
s
y
s
tem
's
ef
f
ec
tiv
en
ess
.
P
er
f
o
r
m
a
n
ce
m
etr
ics
lik
e
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
al
l,
F1
-
s
co
r
e
,
an
d
d
etec
tio
n
r
ate
ar
e
u
tili
ze
d
f
o
r
esti
m
at
in
g
th
e
p
r
o
p
o
s
ed
m
eth
o
d
.
T
h
e
m
ath
em
atica
l
ex
p
r
ess
io
n
f
o
r
ev
er
y
m
etr
ic
is
d
escr
ib
ed
as
th
e
f
o
llo
win
g
(
1
4
)
to
(
1
7
)
.
W
h
er
e
d
en
o
tes
th
e
tr
u
e
p
o
s
itiv
e
,
s
ig
n
if
ies th
e
tr
u
e
n
eg
ativ
e
,
is
th
e
f
alse p
o
s
itiv
e
,
r
ef
er
s
to
th
e
f
alse n
eg
ati
v
e
.
=
+
+
+
+
(
1
4
)
=
+
(
1
5
)
=
+
(
1
6
)
1
−
=
2
2
+
+
(
1
7
)
4
.
1
.
P
er
f
o
r
m
a
nce
a
na
ly
s
is
T
h
e
im
p
o
r
ta
n
ce
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
is
esti
m
ated
u
s
in
g
v
ar
io
u
s
p
er
f
o
r
m
a
n
ce
m
etr
ics
b
ased
o
n
UNSW
-
N
B
1
5
an
d
C
I
C
-
I
DS
-
2
0
1
7
d
atasets
.
T
ab
le
3
d
em
o
n
s
tr
ates
th
e
an
aly
s
is
o
f
th
e
f
ea
tu
r
e
s
elec
tio
n
ap
p
r
o
ac
h
es.
T
h
e
ex
is
tin
g
o
p
ti
m
izatio
n
-
b
ased
f
ea
tu
r
e
s
elec
tio
n
ap
p
r
o
ac
h
es
lik
e
Go
ld
en
J
ac
k
al
o
p
tim
izatio
n
(
GJO),
Ho
n
ey
B
ad
g
er
o
p
tim
i
za
tio
n
(
HB
O)
,
wh
ale
o
p
tim
iz
atio
n
alg
o
r
ith
m
(
W
OA)
,
a
n
d
GT
O
ar
e
co
m
p
ar
ed
an
d
esti
m
ated
with
th
e
C
C
F
-
GT
O
ap
p
r
o
ac
h
.
C
C
F
-
GT
O
em
p
lo
y
s
r
an
d
o
m
n
ess
to
m
o
v
e
s
o
lu
tio
n
s
i
n
to
u
n
ex
p
l
o
r
ed
r
eg
io
n
s
,
h
elp
i
n
g
t
o
p
r
e
v
en
t
p
r
em
atu
r
e
co
n
v
er
g
e
n
ce
an
d
en
s
u
r
i
n
g
a
m
o
r
e
th
o
r
o
u
g
h
ex
p
l
o
r
atio
n
o
f
th
e
s
ea
r
ch
s
p
ac
e.
T
h
is
f
lex
ib
il
ity
en
a
b
les
th
e
C
C
F
-
GT
O
to
a
d
ap
t
to
v
ar
io
u
s
o
p
tim
izatio
n
p
r
o
b
lem
s
,
m
a
k
in
g
it
ap
p
r
o
p
r
iate
f
o
r
co
m
p
le
x
a
n
d
d
y
n
am
ic
f
ea
tu
r
e
s
elec
tio
n
s
ce
n
ar
io
s
.
I
n
t
h
e
UNSW
-
NB
1
5
d
ataset,
th
e
p
r
o
p
o
s
ed
C
C
F
-
GT
O
ap
p
r
o
ac
h
attain
s
a
b
etter
ac
cu
r
ac
y
o
f
9
9
.
5
6
%
a
n
d
DR
o
f
9
9
.
4
5
%.
I
n
th
e
C
I
C
-
I
DS
-
2
0
1
7
d
ataset,
th
e
p
r
o
p
o
s
ed
C
C
F
-
GT
O
ap
p
r
o
ac
h
r
ea
ch
es a
b
etter
ac
cu
r
ac
y
o
f
9
9
.
9
4
% a
n
d
DR
o
f
9
9
.
8
0
%
r
esp
ec
tiv
ely
.
T
ab
le
3
.
An
aly
s
is
o
f
f
ea
tu
r
e
s
e
lectio
n
ap
p
r
o
ac
h
es
D
a
t
a
s
e
t
M
e
t
h
o
d
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
sc
o
r
e
(
%)
D
R
(
%)
U
N
S
W
-
N
B
1
5
G
JO
9
1
.
2
0
8
5
.
3
4
8
5
.
0
9
8
6
.
9
2
8
7
.
2
9
H
B
O
9
3
.
1
2
8
7
.
1
1
8
7
.
6
2
8
8
.
2
9
9
1
.
2
9
W
O
A
9
5
.
4
8
8
9
.
2
9
9
0
.
8
7
9
0
.
8
2
9
3
.
0
2
G
TO
9
7
.
3
9
9
1
.
7
0
9
3
.
1
2
9
2
.
1
2
9
5
.
6
3
CCF
-
G
TO
9
9
.
5
6
9
3
.
2
9
9
5
.
2
0
9
4
.
2
5
9
9
.
4
5
C
I
C
-
I
D
S
-
2
0
1
7
G
JO
9
2
.
4
5
8
7
.
6
5
8
6
.
7
8
8
9
.
3
0
9
1
.
3
2
H
B
O
9
4
.
2
1
8
9
.
9
8
8
8
.
3
8
9
1
.
3
3
9
4
.
2
1
W
O
A
9
6
.
2
9
9
0
.
1
2
9
2
.
0
9
9
3
.
4
2
9
5
.
6
4
G
TO
9
8
.
3
3
9
2
.
3
8
9
4
.
3
6
9
5
.
6
7
9
7
.
8
7
CCF
-
G
TO
9
9
.
9
4
9
9
.
6
9
9
9
.
7
1
9
9
.
7
0
9
9
.
8
0
T
ab
le
4
d
em
o
n
s
tr
ates
th
e
an
aly
s
is
o
f
th
e
d
if
f
e
r
en
t
class
if
ier
s
.
T
h
e
ex
is
tin
g
class
if
ier
s
lik
e
C
NN,
g
ated
r
ec
u
r
r
en
t
u
n
it
(
GR
U)
,
r
ec
u
r
r
en
t
n
eu
r
al
n
etwo
r
k
(
R
N
N
)
,
an
d
L
STM
ar
e
co
m
p
a
r
ed
an
d
esti
m
ated
with
th
e
MPE
L
U
-
L
STM
ap
p
r
o
ac
h
.
MPE
L
U
in
teg
r
ates
th
e
ad
v
an
tag
es
o
f
R
eL
U
an
d
E
L
U
f
u
n
c
tio
n
s
,
p
r
o
v
i
d
in
g
a
m
o
r
e
f
lex
ib
le
ac
ti
v
atio
n
f
u
n
ctio
n
.
T
h
is
f
lex
ib
ilit
y
en
h
a
n
ce
s
th
e
L
STM
'
s
ca
p
ab
ilit
y
t
o
m
o
d
el
in
t
r
icate
d
ep
en
d
e
n
cies,
wh
ich
is
cr
u
ci
al
f
o
r
d
etec
tin
g
p
atter
n
s
in
I
DS.
I
n
th
e
UNSW
-
NB
1
5
d
ataset,
th
e
p
r
o
p
o
s
ed
MPE
L
U
-
L
STM
ap
p
r
o
ac
h
atta
in
s
a
b
etter
ac
cu
r
ac
y
o
f
9
9
.
5
6
%,
p
r
ec
is
io
n
o
f
9
3
.
2
9
%,
r
ec
all
o
f
9
5
.
2
0
%,
F1
-
s
co
r
e
o
f
9
4
.
2
5
%
,
an
d
DR
o
f
9
9
.
4
5
%.
I
n
th
e
C
I
C
-
I
DS
-
2
0
1
7
d
ataset,
th
e
p
r
o
p
o
s
ed
MPE
L
U
-
L
STM
ap
p
r
o
ac
h
attain
s
a
b
etter
ac
cu
r
ac
y
o
f
9
9
.
9
4
%,
p
r
ec
is
io
n
o
f
9
9
.
6
9
%,
r
ec
all
o
f
9
9
.
7
1
%,
F1
-
s
co
r
e
o
f
9
9
.
7
0
%
,
an
d
DR
o
f
9
9
.
8
0
% r
esp
ec
tiv
el
y
.
T
ab
le
4
.
An
aly
s
is
o
f
d
if
f
er
e
n
t
class
if
ier
s
D
a
t
a
s
e
t
M
e
t
h
o
d
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
sc
o
r
e
(
%)
D
R
(
%)
U
N
S
W
-
N
B
1
5
C
N
N
9
2
.
4
4
8
8
.
0
1
8
8
.
2
3
8
7
.
4
5
9
2
.
3
4
G
R
U
9
5
.
6
3
9
0
.
2
2
9
0
.
7
6
9
0
.
3
9
9
4
.
5
6
R
N
N
9
6
.
4
9
9
1
.
3
2
9
2
.
1
8
9
2
.
4
5
9
6
.
3
2
LSTM
9
8
.
4
5
9
2
.
2
3
9
4
.
2
2
9
3
.
5
6
9
7
.
5
4
M
P
ELU
-
LST
M
9
9
.
5
6
9
3
.
2
9
9
5
.
2
0
9
4
.
2
5
9
9
.
4
5
C
I
C
-
I
D
S
-
2
0
1
7
C
N
N
9
3
.
5
6
9
1
.
3
2
9
1
.
3
0
9
1
.
2
9
9
4
.
2
9
G
R
U
9
5
.
3
5
9
3
.
1
2
9
2
.
3
2
9
3
.
2
4
9
5
.
2
6
R
N
N
9
7
.
3
4
9
5
.
3
8
9
4
.
2
0
9
4
.
3
2
9
6
.
2
2
LSTM
9
8
.
4
5
9
7
.
3
9
9
5
.
3
3
9
6
.
4
7
9
8
.
7
6
M
P
ELU
-
LST
M
9
9
.
9
4
9
9
.
6
9
9
9
.
7
1
9
9
.
7
0
9
9
.
8
0
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.
14
,
No
.
2
,
J
u
n
e
2
0
2
5
:
3
7
3
-
383
380
T
ab
le
5
d
em
o
n
s
tr
ates
th
e
p
er
f
o
r
m
an
ce
a
n
aly
s
is
o
f
L
STM
with
d
if
f
er
en
t
ac
tiv
atio
n
f
u
n
c
tio
n
s
.
T
h
e
L
STM
is
co
m
p
ar
e
d
an
d
esti
m
ated
with
th
e
d
if
f
e
r
en
t
ac
tiv
ati
o
n
f
u
n
ctio
n
s
lik
e
E
L
U,
R
eL
U,
Par
am
etr
ic
R
eL
U
(
PR
eL
U)
,
L
ea
k
y
R
eL
U
(
L
R
eL
U)
,
an
d
MPE
L
U.
T
h
r
o
u
g
h
ad
ju
s
tin
g
th
ese
p
ar
a
m
eter
s
,
MPE
L
U
ca
p
tu
r
es
a
b
r
o
ad
r
an
g
e
o
f
n
o
n
-
lin
ea
r
r
e
latio
n
s
h
ip
s
in
th
e
d
ata,
allo
win
g
it
to
m
o
d
el
th
e
s
u
b
tle
p
atter
n
s
th
at
o
t
h
er
ac
tiv
atio
n
f
u
n
ctio
n
s
o
f
ten
m
is
s
.
T
h
i
s
ad
a
p
tab
ilit
y
en
a
b
les
th
e
MPE
L
U
with
L
STM
to
attain
b
etter
r
esu
lts
in
d
eter
m
in
i
n
g
t
h
e
u
n
d
er
ly
in
g
p
a
tter
n
s
in
th
e
d
ata.
T
ab
le
5
.
An
aly
s
is
o
f
L
STM
wi
th
d
if
f
er
e
n
t a
ctiv
atio
n
f
u
n
ctio
n
s
D
a
t
a
s
e
t
M
e
t
h
o
d
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
sc
o
r
e
(
%)
D
R
(
%)
U
N
S
W
-
N
B
1
5
ELU
-
LSTM
9
3
.
1
2
8
9
.
9
0
9
0
.
1
2
9
0
.
3
2
9
1
.
4
7
R
e
LU
-
LST
M
9
5
.
4
3
9
0
.
3
2
9
1
.
7
8
9
1
.
8
2
9
4
.
2
3
P
R
e
LU
-
LST
M
9
7
.
3
9
9
1
.
3
1
9
3
.
2
1
9
2
.
1
6
9
5
.
2
3
LR
e
LU
-
LS
TM
9
6
.
1
9
9
2
.
9
1
9
4
.
2
3
9
3
.
2
3
9
7
.
6
5
M
P
ELU
-
LST
M
9
9
.
5
6
9
3
.
2
9
9
5
.
2
0
9
4
.
2
5
9
9
.
4
5
C
I
C
-
I
D
S
-
2
0
1
7
ELU
-
LSTM
9
3
.
5
8
9
3
.
7
6
9
4
.
3
7
9
2
.
1
2
9
4
.
3
2
R
e
LU
-
LST
M
9
5
.
2
9
9
4
.
2
1
9
5
.
2
9
9
4
.
2
2
9
5
.
4
3
P
R
e
LU
-
LST
M
9
6
.
4
2
9
7
.
5
3
9
7
.
4
6
9
5
.
2
9
9
6
.
5
4
LR
e
LU
-
LS
TM
9
7
.
1
2
9
8
.
4
2
9
8
.
3
2
9
7
.
6
7
9
8
.
7
5
M
P
ELU
-
LST
M
9
9
.
9
4
9
9
.
6
9
9
9
.
7
1
9
9
.
7
0
9
9
.
8
0
4
.
2
.
Co
m
pa
ra
t
iv
e
a
na
ly
s
is
I
n
th
is
s
ec
tio
n
,
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
p
r
o
p
o
s
ed
MPE
L
U
-
L
STM
ap
p
r
o
ac
h
is
co
m
p
a
r
ed
with
th
e
ex
is
tin
g
m
eth
o
d
s
u
s
in
g
UNSW
-
NB
1
5
an
d
C
I
C
-
I
DS2
0
1
7
d
atasets
.
T
ab
le
6
d
em
o
n
s
tr
ates
th
e
co
m
p
ar
ativ
e
an
aly
s
is
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
.
T
h
e
ex
is
tin
g
m
eth
o
d
s
s
u
ch
as
C
N
N+
L
STM
[
1
6
]
,
Op
tim
ized
DR
C
N
N
[
1
8
]
,
R
es
-
T
r
an
I
DS
[
1
9
]
,
an
d
C
NN
-
L
STM
[
2
0
]
ar
e
co
m
p
ar
ed
a
n
d
esti
m
ated
with
t
h
e
p
r
o
p
o
s
ed
MPE
L
U
-
L
STM
m
eth
o
d
in
ter
m
s
o
f
v
ar
io
u
s
p
er
f
o
r
m
a
n
ce
m
etr
ics.
T
h
e
lear
n
ab
le
p
ar
am
eter
s
o
f
MPE
L
U
p
r
o
v
id
e
f
lex
ib
ilit
y
,
im
p
r
o
v
in
g
th
e
L
STM
'
s
ca
p
ab
ilit
y
to
id
en
tif
y
n
o
n
-
lin
ea
r
r
ela
tio
n
s
h
ip
s
in
I
DS
d
ata
f
o
r
en
h
a
n
ce
d
class
if
icatio
n
ac
cu
r
ac
y
.
T
ab
le
6
.
C
o
m
p
a
r
ativ
e
an
aly
s
is
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
(
NA
=
n
o
t
ap
p
licab
le
)
D
a
t
a
s
e
t
M
e
t
h
o
d
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
sc
o
r
e
(
%)
D
R
(
%)
U
N
S
W
-
N
B
1
5
C
N
N
+
LST
M
[
1
6
]
NA
9
2
.
9
1
9
3
.
1
0
9
3
.
0
0
NA
O
p
t
i
mi
z
e
d
D
R
C
N
N
[
1
8
]
9
9
.
0
6
NA
NA
NA
9
8
.
9
9
C
N
N
-
LSTM
[
2
0
]
9
3
.
7
8
NA
NA
NA
9
4
.
5
3
P
r
o
p
o
se
d
M
P
ELU
-
LST
M
9
9
.
5
6
9
3
.
2
9
9
5
.
2
0
9
4
.
2
5
9
9
.
4
5
C
I
C
-
I
D
S
-
2
0
1
7
Res
-
Tr
a
n
I
D
S
[
1
9
]
9
9
.
1
5
NA
NA
NA
NA
C
N
N
-
LSTM
[
2
0
]
9
9
.
6
4
NA
NA
NA
9
9
.
7
0
P
r
o
p
o
se
d
M
P
ELU
-
LST
M
9
9
.
9
4
9
9
.
6
9
9
9
.
7
1
9
9
.
7
0
9
9
.
8
0
4
.
3
.
Dis
cus
s
io
n
T
h
is
s
ec
tio
n
d
is
cu
s
s
es
th
e
lim
itatio
n
s
o
f
ex
is
tin
g
w
o
r
k
s
an
d
ex
p
lain
s
h
o
w
th
e
p
r
o
p
o
s
ed
MPE
L
U
-
L
STM
ap
p
r
o
ac
h
ad
d
r
ess
es
th
ese
lim
itatio
n
s
,
alo
n
g
with
its
ad
v
an
tag
es.
T
h
e
lim
itatio
n
s
o
f
th
e
ex
is
tin
g
wo
r
k
s
s
u
ch
as
lack
o
f
p
er
f
o
r
m
in
g
th
e
f
ea
tu
r
e
s
elec
tio
n
p
r
o
ce
s
s
,
ch
allen
g
in
g
with
f
ea
tu
r
e
r
eten
tio
n
in
im
b
alan
ce
d
d
atasets
,
an
d
m
in
i
m
ized
in
ter
p
r
etab
ilit
y
.
Hen
ce
,
th
is
r
esear
ch
aim
s
to
p
r
o
p
o
s
e
th
e
C
C
F
-
GT
O
f
o
r
s
elec
tin
g
th
e
in
f
o
r
m
ativ
e
f
ea
t
u
r
es
an
d
MPE
L
U
-
LS
T
M
f
o
r
th
e
class
if
icatio
n
o
f
I
DS
f
o
r
s
o
lv
in
g
th
e
a
b
o
v
e
-
m
en
tio
n
ed
lim
itatio
n
s
f
r
o
m
th
e
liter
atu
r
e
s
u
r
v
ey
.
C
C
F
-
GT
O
m
ain
tain
s
a
b
etter
b
alan
ce
b
etwe
en
ex
p
lo
r
atio
n
an
d
ex
p
lo
itatio
n
.
T
h
is
b
alan
ce
is
im
p
o
r
tan
t
in
f
ea
tu
r
e
s
elec
ti
o
n
p
r
o
ce
s
s
es
to
s
o
lv
e
lo
ca
l
o
p
tim
a
p
r
o
b
lem
s
an
d
m
ak
e
s
u
r
e
a
c
o
m
p
r
eh
en
s
iv
e
s
e
ar
ch
f
o
r
th
e
m
o
s
t
r
elev
an
t
f
ea
tu
r
es.
T
h
e
a
d
ap
tiv
e
m
ec
h
an
is
m
in
tr
o
d
u
ce
d
b
y
th
e
C
C
F
im
p
r
o
v
es
th
e
c
o
n
v
e
r
g
en
ce
r
ate
o
f
th
e
o
p
tim
izatio
n
p
r
o
ce
s
s
,
av
o
id
in
g
lo
ca
l
o
p
tim
a
a
n
d
ac
h
ie
v
in
g
g
lo
b
al
o
p
tim
izatio
n
f
o
r
f
ea
tu
r
e
s
ele
c
tio
n
.
T
h
e
MPE
L
U
ac
tiv
atio
n
f
u
n
ctio
n
p
r
o
d
u
ce
s
s
u
p
p
le
m
en
tar
y
p
a
r
am
eter
s
lear
n
ed
d
u
r
in
g
t
h
e
tr
ain
i
n
g
p
r
o
ce
s
s
,
en
ab
lin
g
m
o
r
e
f
le
x
ib
ilit
y
in
th
e
ac
tiv
atio
n
b
eh
av
i
o
r
.
T
h
ese
p
ar
am
eter
s
allo
w
th
e
ac
tiv
atio
n
f
u
n
ctio
n
to
ad
ap
t
its
s
h
ap
e
b
ased
o
n
d
ata
d
is
tr
ib
u
tio
n
,
d
esig
n
in
g
it
well
ap
p
r
o
p
r
iate
f
o
r
ca
p
tu
r
in
g
co
m
p
lex
,
n
o
n
-
lin
ea
r
r
elatio
n
s
h
ip
s
in
I
DS
d
ata.
T
h
e
u
tili
za
tio
n
o
f
th
e
MPE
L
U
ac
tiv
atio
n
f
u
n
ctio
n
im
p
r
o
v
es
th
e
n
o
n
-
lin
ea
r
m
o
d
e
lin
g
ca
p
ab
ilit
y
o
f
L
STM
,
e
n
ab
lin
g
it
to
b
etter
ca
p
tu
r
e
p
atter
n
s
in
I
DS
d
ata.
T
h
is
f
lex
ib
i
lity
en
h
an
ce
s
th
e
L
STM
’
s
ca
p
ab
ilit
y
to
m
o
d
el
co
m
p
lex
r
elatio
n
s
h
ip
s
in
I
DS
d
ata,
r
esu
ltin
g
in
b
etter
class
if
icatio
n
p
er
f
o
r
m
a
n
ce
.
T
h
e
p
r
o
p
o
s
ed
MPE
L
U
-
L
STM
ap
p
r
o
ac
h
attain
s
a
b
etter
D
R
o
f
9
9
.
4
5
%
an
d
9
9
.
8
0
%
o
n
UNSW
-
NB
1
5
an
d
C
I
C
-
I
DS
-
2
0
1
7
d
atasets
.
Ho
wev
er
,
th
e
e
x
is
tin
g
m
eth
o
d
o
f
o
p
tim
ized
DR
C
N
N
[
1
8
]
an
d
C
NN
-
L
STM
[
2
0
]
att
ain
ed
th
e
less
DR
o
f
9
8
.
9
9
%
a
n
d
9
4
.
5
3
%
in
th
e
UNSW
-
NB
1
5
d
ataset,
wh
e
r
ea
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
C
o
n
tr
a
ctio
n
co
n
tr
o
l fa
cto
r
-
b
a
s
ed
g
o
r
illa
tr
o
o
p
o
p
timiz
er fo
r
fea
tu
r
es in
in
tr
u
s
io
n
…
(
S
h
a
lin
i S
h
a
r
ma
)
381
in
th
e
C
I
C
-
I
DS
-
2
0
1
7
d
ataset,
C
NN
-
L
STM
[
2
0
]
attain
ed
th
e
less
DR
o
f
9
9
.
7
0
%
r
esp
ec
tiv
ely
.
T
h
ese
r
esu
lts
d
em
o
n
s
tr
at
e
th
at
th
e
p
r
o
p
o
s
ed
C
C
F
-
GT
O
ap
p
r
o
ac
h
attain
s
b
etter
r
esu
lts
as
co
m
p
ar
ed
to
th
e
ex
is
tin
g
m
eth
o
d
s
b
y
s
elec
tin
g
th
e
m
o
s
t
r
elev
an
t
f
ea
tu
r
es.
T
h
e
C
C
F
-
GT
O
ap
p
r
o
ac
h
e
n
s
u
r
es
th
at
th
e
I
DS
m
o
d
el
is
n
o
t
o
v
er
wh
elm
ed
by
ir
r
elev
a
n
t
d
ata,
lead
in
g
to
en
ha
n
ce
d
d
etec
tio
n
r
a
tes,
ac
cu
r
ac
y
,
a
n
d
o
v
e
r
all
m
o
d
el
p
er
f
o
r
m
an
ce
.
5.
CO
NCLU
SI
O
N
I
DS
is
cr
u
cial
in
th
e
r
ea
lm
o
f
d
ata
p
r
o
tectio
n
f
o
r
I
o
T
,
p
lay
i
n
g
a
v
ital
r
o
le
in
s
ec
u
r
in
g
u
s
er
d
ata
an
d
p
r
o
tectin
g
i
n
tellectu
al
d
ev
ice
s
.
Nev
er
th
eless
,
tr
ad
itio
n
al
I
DS
b
ased
o
n
s
tatis
tics
an
d
ex
p
er
t
s
y
s
tem
s
ar
e
co
m
p
lex
to
m
ee
t
s
ec
u
r
ity
r
e
q
u
ir
em
en
ts
f
o
r
r
ap
id
n
etwo
r
k
d
ev
elo
p
m
e
n
t
an
d
th
e
co
n
tin
u
o
u
s
g
r
o
wth
o
f
lar
g
e
d
ata.
Hen
ce
,
th
is
r
esear
ch
p
r
o
p
o
s
es
th
e
C
C
F
-
GT
O
f
o
r
s
elec
tin
g
th
e
in
f
o
r
m
ativ
e
f
ea
tu
r
es
an
d
MPE
L
U
-
L
STM
f
o
r
th
e
class
if
icatio
n
o
f
I
D
S.
T
h
e
ad
d
itio
n
al
p
ar
am
eter
s
o
f
MPE
L
U
p
r
o
v
id
e
s
ig
n
if
ican
t
f
lex
ib
ilit
y
in
ac
tiv
atio
n
f
u
n
ctio
n
s
,
en
a
b
lin
g
th
e
L
STM
to
ad
a
p
t
b
etter
to
v
ar
io
u
s
d
ata
p
atter
n
s
in
I
DS,
a
n
d
lead
in
g
to
b
etter
class
if
icatio
n
p
er
f
o
r
m
an
ce
.
T
h
e
ex
p
er
im
en
tal
r
esu
lts
illu
s
t
r
at
e
th
at
th
e
p
r
o
p
o
s
ed
MPE
L
U
-
L
STM
m
eth
o
d
attain
s
b
etter
ac
cu
r
ac
ies
o
f
9
9
.
5
6
%
an
d
9
9
.
9
4
%
o
n
UNSW
-
NB
1
5
an
d
C
I
C
-
I
DS
-
2
0
1
7
d
ata
s
ets
as
co
m
p
ar
ed
to
th
e
ex
is
tin
g
m
eth
o
d
s
s
u
ch
as
o
p
tim
ized
DR
C
NN
an
d
C
N
N
-
L
STM
.
T
h
e
f
u
tu
r
e
wo
r
k
in
v
o
lv
es
a
h
y
b
r
id
D
L
ap
p
r
o
ac
h
to
en
h
an
ce
th
e
o
v
e
r
a
ll c
la
s
s
if
icatio
n
p
er
f
o
r
m
an
ce
f
o
r
I
DS in
I
o
T
.
ACK
NO
WL
E
DG
E
M
E
NT
T
h
e
au
th
o
r
s
wo
u
ld
lik
e
to
th
an
k
o
u
r
s
u
p
er
v
is
o
r
,
Pro
f
.
Sy
e
d
Z
ee
s
h
an
Hu
s
s
ain
,
f
o
r
h
is
c
o
n
tin
u
o
u
s
m
o
tiv
atio
n
an
d
s
u
p
p
o
r
t th
r
o
u
g
h
o
u
t th
is
r
esear
ch
.
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
e
d.
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
Sh
alin
i Sh
ar
m
a
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Su
p
r
iy
a
Kh
aitan
✓
✓
✓
✓
✓
✓
✓
Gay
atr
i H
eg
d
e
✓
✓
✓
✓
✓
✓
✓
✓
Div
y
a
R
o
h
atg
i
✓
✓
✓
✓
✓
✓
✓
Nu
s
r
at
Par
v
ee
n
Mo
h
am
m
ad
R
af
iq
u
e
✓
✓
✓
✓
✓
Su
h
as Jan
ar
d
an
L
awa
n
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
e
n
l
y
av
ailab
le
in
[
UNSW
-
NB
1
5
an
d
C
I
C
-
I
DS2
0
2
7
]
at
h
ttp
s
://re
s
ea
r
ch
.
u
n
s
w.
ed
u
.
au
/p
r
o
jects/u
n
s
w
-
n
b
1
5
-
d
ataset
an
d
h
ttp
s
://www.
k
ag
g
le.
co
m
/d
atasets
/ch
eth
u
h
n
/n
etwo
r
k
-
i
n
tr
u
s
io
n
-
d
ataset.
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.
14
,
No
.
2
,
J
u
n
e
2
0
2
5
:
3
7
3
-
383
382
RE
F
E
R
E
NC
E
S
[
1
]
S
.
R
o
y
,
J.
Li
,
B
.
J.
C
h
o
i
,
a
n
d
Y
.
B
a
i
,
“
A
l
i
g
h
t
w
e
i
g
h
t
s
u
p
e
r
v
i
se
d
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
me
c
h
a
n
i
sm
f
o
r
I
o
T
n
e
t
w
o
r
k
s,
”
F
u
t
u
re
G
e
n
e
r
a
t
i
o
n
C
o
m
p
u
t
e
r
S
y
s
t
e
m
s
,
v
o
l
.
1
2
7
,
p
p
.
2
7
6
–
2
8
5
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
f
u
t
u
r
e
.
2
0
2
1
.
0
9
.
0
2
7
.
[
2
]
A
.
M
o
m
a
n
d
,
S
.
U
.
Ja
n
,
a
n
d
N
.
R
a
mz
a
n
,
“
A
B
C
N
N
-
I
D
S
:
a
t
t
e
n
t
i
o
n
-
b
a
se
d
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
f
o
r
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
i
n
I
o
T
n
e
t
w
o
r
k
s,”
W
i
re
l
e
ss
Pe
rs
o
n
a
l
C
o
m
m
u
n
i
c
a
t
i
o
n
s
,
v
o
l
.
1
3
6
,
n
o
.
4
,
p
p
.
1
9
8
1
–
2
0
0
3
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
2
7
7
-
0
2
4
-
1
1
2
6
0
-
7.
[
3
]
Y
.
Y
a
n
g
,
J
.
C
h
e
n
g
,
Z
.
Li
u
,
H
.
Li
,
a
n
d
G
.
X
u
,
“
A
m
u
l
t
i
-
c
l
a
s
si
f
i
c
a
t
i
o
n
d
e
t
e
c
t
i
o
n
m
o
d
e
l
f
o
r
i
mb
a
l
a
n
c
e
d
d
a
t
a
i
n
N
I
D
S
b
a
se
d
o
n
r
e
c
o
n
st
r
u
c
t
i
o
n
a
n
d
f
e
a
t
u
r
e
m
a
t
c
h
i
n
g
,
”
J
o
u
r
n
a
l
o
f
C
l
o
u
d
C
o
m
p
u
t
i
n
g
,
v
o
l
.
1
3
,
n
o
.
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
8
6
/
s
1
3
6
7
7
-
023
-
0
0
5
8
4
-
7.
[
4
]
Y
.
K
a
y
o
d
e
S
a
h
e
e
d
,
A
.
I
d
r
i
s
A
b
i
o
d
u
n
,
S
.
M
i
sr
a
,
M
.
K
r
i
st
i
a
n
s
e
n
H
o
l
o
n
e
,
a
n
d
R
.
C
o
l
o
mo
-
P
a
l
a
c
i
o
s,
“
A
mac
h
i
n
e
l
e
a
r
n
i
n
g
-
b
a
s
e
d
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
f
o
r
d
e
t
e
c
t
i
n
g
i
n
t
e
r
n
e
t
o
f
t
h
i
n
g
s
n
e
t
w
o
r
k
a
t
t
a
c
k
s
,
”
Al
e
x
a
n
d
ri
a
En
g
i
n
e
e
ri
n
g
J
o
u
r
n
a
l
,
v
o
l
.
6
1
,
n
o
.
1
2
,
p
p
.
9
3
9
5
–
9
4
0
9
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
e
j
.
2
0
2
2
.
0
2
.
0
6
3
.
[
5
]
M
.
A
.
Ta
l
u
k
d
e
r
et
a
l
.
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
-
b
a
s
e
d
n
e
t
w
o
r
k
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
f
o
r
b
i
g
a
n
d
i
m
b
a
l
a
n
c
e
d
d
a
t
a
u
s
i
n
g
o
v
e
r
s
a
mp
l
i
n
g
,
st
a
c
k
i
n
g
f
e
a
t
u
r
e
e
m
b
e
d
d
i
n
g
a
n
d
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
,
”
J
o
u
r
n
a
l
o
f
Bi
g
D
a
t
a
,
v
o
l
.
1
1
,
n
o
.
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
8
6
/
s
4
0
5
3
7
-
0
2
4
-
0
0
8
8
6
-
w.
[
6
]
M
.
S
a
r
h
a
n
,
S
.
La
y
e
g
h
y
,
N
.
M
o
u
st
a
f
a
,
M
.
G
a
l
l
a
g
h
e
r
,
a
n
d
M
.
P
o
r
t
ma
n
n
,
“
F
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
f
o
r
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
-
b
a
se
d
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
i
n
I
o
T
n
e
t
w
o
r
k
s,
”
D
i
g
i
t
a
l
C
o
m
m
u
n
i
c
a
t
i
o
n
s
a
n
d
N
e
t
w
o
r
k
s
,
v
o
l
.
1
0
,
n
o
.
1
,
p
p
.
2
0
5
–
2
1
6
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
d
c
a
n
.
2
0
2
2
.
0
8
.
0
1
2
.
[
7
]
S
.
La
y
e
g
h
y
,
M
.
B
a
k
t
a
sh
mo
t
l
a
g
h
,
a
n
d
M
.
P
o
r
t
m
a
n
n
,
“
D
I
-
N
I
D
S
:
d
o
mai
n
i
n
v
a
r
i
a
n
t
n
e
t
w
o
r
k
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
s
y
st
e
m,”
K
n
o
w
l
e
d
g
e
-
B
a
se
d
S
y
s
t
e
m
s
,
v
o
l
.
2
7
3
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
k
n
o
sy
s
.
2
0
2
3
.
1
1
0
6
2
6
.
[
8
]
N
.
D
a
t
-
T
h
i
n
h
,
H
.
X
u
a
n
-
N
i
n
h
,
a
n
d
L.
K
i
m
-
H
u
n
g
,
“
M
i
d
S
i
o
t
:
a
m
u
l
t
i
st
a
g
e
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
s
y
s
t
e
m
f
o
r
i
n
t
e
r
n
e
t
o
f
t
h
i
n
g
s
,
”
Wi
re
l
e
ss
C
o
m
m
u
n
i
c
a
t
i
o
n
s
a
n
d
M
o
b
i
l
e
C
o
m
p
u
t
i
n
g
,
v
o
l
.
2
0
2
2
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
5
5
/
2
0
2
2
/
9
1
7
3
2
9
1
.
[
9
]
J.
L
i
,
H
.
Z
h
a
n
g
,
Y
.
Li
u
,
a
n
d
Z.
Li
u
,
“
S
e
mi
-
s
u
p
e
r
v
i
se
d
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
f
r
a
mew
o
r
k
f
o
r
n
e
t
w
o
r
k
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
,
”
J
o
u
r
n
a
l
o
f
S
u
p
e
r
c
o
m
p
u
t
i
n
g
,
v
o
l
.
7
8
,
n
o
.
1
1
,
p
p
.
1
3
1
2
2
–
1
3
1
4
4
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
2
2
7
-
0
2
2
-
0
4
3
9
0
-
x.
[
1
0
]
M
.
V
i
s
h
w
a
k
a
r
ma
a
n
d
N
.
K
e
ssw
a
n
i
,
“
A
n
e
w
t
w
o
-
p
h
a
s
e
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
s
y
s
t
e
m
w
i
t
h
n
a
ï
v
e
B
a
y
e
s
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
f
o
r
d
a
t
a
c
l
a
ss
i
f
i
c
a
t
i
o
n
a
n
d
e
l
l
i
p
t
i
c
e
n
v
e
l
o
p
me
t
h
o
d
f
o
r
a
n
o
ma
l
y
d
e
t
e
c
t
i
o
n
,
”
D
e
c
i
si
o
n
An
a
l
y
t
i
c
s
J
o
u
r
n
a
l
,
v
o
l
.
7
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
d
a
j
o
u
r
.
2
0
2
3
.
1
0
0
2
3
3
.
[
1
1
]
M
.
S
a
j
i
d
e
t
a
l
.
,
“
En
h
a
n
c
i
n
g
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
:
a
h
y
b
r
i
d
mac
h
i
n
e
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
,
”
J
o
u
rn
a
l
o
f
C
l
o
u
d
C
o
m
p
u
t
i
n
g
,
v
o
l
.
1
3
,
n
o
.
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
8
6
/
s1
3
6
7
7
-
024
-
0
0
6
8
5
-
x.
[
1
2
]
V
.
G
o
w
d
h
a
ma
n
a
n
d
R
.
D
h
a
n
a
p
a
l
,
“
A
n
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
sy
s
t
e
m
f
o
r
w
i
r
e
l
e
ss
se
n
so
r
n
e
t
w
o
r
k
s
u
si
n
g
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
S
o
f
t
C
o
m
p
u
t
i
n
g
,
v
o
l
.
2
6
,
n
o
.
2
3
,
p
p
.
1
3
0
5
9
–
1
3
0
6
7
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
5
0
0
-
0
2
1
-
0
6
4
7
3
-
y.
[
1
3
]
Y
.
W
a
n
g
,
G
.
S
u
n
,
X
.
C
a
o
,
a
n
d
J.
Y
a
n
g
,
“
A
n
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
s
y
st
e
m
f
o
r
t
h
e
i
n
t
e
r
n
e
t
o
f
t
h
i
n
g
s
b
a
s
e
d
o
n
t
h
e
e
n
s
e
mb
l
e
o
f
u
n
s
u
p
e
r
v
i
s
e
d
t
e
c
h
n
i
q
u
e
s
,
”
Wi
r
e
l
e
ss
C
o
m
m
u
n
i
c
a
t
i
o
n
s
a
n
d
Mo
b
i
l
e
C
o
m
p
u
t
i
n
g
,
v
o
l
.
2
0
2
2
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
5
5
/
2
0
2
2
/
8
6
1
4
9
0
3
.
[
1
4
]
V
.
R
a
v
i
,
R
.
C
h
a
g
a
n
t
i
,
a
n
d
M
.
A
l
a
z
a
b
,
“
D
e
e
p
l
e
a
r
n
i
n
g
f
e
a
t
u
r
e
f
u
si
o
n
a
p
p
r
o
a
c
h
f
o
r
a
n
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
s
y
s
t
e
m
i
n
S
D
N
-
b
a
se
d
I
o
T
n
e
t
w
o
r
k
s,”
I
EEE
I
n
t
e
r
n
e
t
o
f
T
h
i
n
g
s
M
a
g
a
z
i
n
e
,
v
o
l
.
5
,
n
o
.
2
,
p
p
.
2
4
–
2
9
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
I
O
TM
.
0
0
3
.
2
2
0
0
0
0
1
.
[
1
5
]
S
.
A
.
B
a
k
h
s
h
,
M
.
A
.
K
h
a
n
,
F
.
A
h
m
e
d
,
M
.
S
.
A
l
s
h
e
h
r
i
,
H
.
A
l
i
,
a
n
d
J.
A
h
ma
d
,
“
E
n
h
a
n
c
i
n
g
I
o
T
n
e
t
w
o
r
k
s
e
c
u
r
i
t
y
t
h
r
o
u
g
h
d
e
e
p
l
e
a
r
n
i
n
g
-
p
o
w
e
r
e
d
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
sy
st
e
m
,
”
I
n
t
e
r
n
e
t
o
f
T
h
i
n
g
s (
N
e
t
h
e
rl
a
n
d
s)
,
v
o
l
.
2
4
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
o
t
.
2
0
2
3
.
1
0
0
9
3
6
.
[
1
6
]
H
.
C
.
A
l
t
u
n
a
y
a
n
d
Z.
A
l
b
a
y
r
a
k
,
“
A
h
y
b
r
i
d
C
N
N
+
LST
M
b
a
se
d
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
sy
s
t
e
m
f
o
r
i
n
d
u
s
t
r
i
a
l
I
o
T
n
e
t
w
o
r
k
s,”
En
g
i
n
e
e
ri
n
g
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
a
n
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
,
v
o
l
.
3
8
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
e
s
t
c
h
.
2
0
2
2
.
1
0
1
3
2
2
.
[
1
7
]
V
.
H
n
a
mt
e
,
H
.
N
h
u
n
g
-
N
g
u
y
e
n
,
J.
H
u
ssa
i
n
,
a
n
d
Y
.
H
w
a
-
K
i
m,
“
A
n
o
v
e
l
t
w
o
-
st
a
g
e
d
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
f
o
r
n
e
t
w
o
r
k
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
:
LST
M
-
A
E,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
1
1
,
p
p
.
3
7
1
3
1
–
3
7
1
4
8
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
3
.
3
2
6
6
9
7
9
.
[
1
8
]
G
.
S
.
C
.
K
u
m
a
r
,
R
.
K
.
K
u
mar
,
K
.
P
.
V
.
K
u
mar,
N
.
R
.
S
a
i
,
a
n
d
M
.
B
r
a
h
m
a
i
a
h
,
“
D
e
e
p
r
e
si
d
u
a
l
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
:
a
n
e
f
f
i
c
i
e
n
t
t
e
c
h
n
i
q
u
e
f
o
r
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
s
y
st
e
m,”
E
x
p
e
rt
S
y
s
t
e
m
s
w
i
t
h
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
2
3
8
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
sw
a
.
2
0
2
3
.
1
2
1
9
1
2
.
[
1
9
]
S
.
W
a
n
g
,
W
.
X
u
,
a
n
d
Y
.
Li
u
,
“
R
e
s
-
Tr
a
n
B
i
LST
M
:
a
n
i
n
t
e
l
l
i
g
e
n
t
a
p
p
r
o
a
c
h
f
o
r
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
i
n
t
h
e
i
n
t
e
r
n
e
t
o
f
t
h
i
n
g
s
,
”
C
o
m
p
u
t
e
r
N
e
t
w
o
r
k
s
,
v
o
l
.
2
3
5
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mn
e
t
.
2
0
2
3
.
1
0
9
9
8
2
.
[
2
0
]
A
.
H
a
l
b
o
u
n
i
,
T.
S
.
G
u
n
a
w
a
n
,
M
.
H
.
H
a
b
a
e
b
i
,
M
.
H
a
l
b
o
u
n
i
,
M
.
K
a
r
t
i
w
i
,
a
n
d
R
.
A
h
m
a
d
,
“
C
N
N
-
LST
M
:
h
y
b
r
i
d
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
f
o
r
n
e
t
w
o
r
k
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
s
y
st
e
m,”
I
EEE
Ac
c
e
ss
,
v
o
l
.
1
0
,
p
p
.
9
9
8
3
7
–
9
9
8
4
9
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
2
.
3
2
0
6
4
2
5
.
[
2
1
]
N
.
M
o
u
s
t
a
f
a
,
“
U
N
S
W
-
N
B
1
5
d
a
t
a
set
,
”
U
N
S
W
S
y
d
n
e
y
,
2
0
2
1
.
(
a
c
c
e
sse
d
A
u
g
.
2
0
,
2
0
2
4
)
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
r
e
se
a
r
c
h
.
u
n
sw
.
e
d
u
.
a
u
/
p
r
o
j
e
c
t
s
/
u
n
sw
-
n
b
1
5
-
d
a
t
a
s
e
t
.
[
2
2
]
H
.
N
.
C
h
e
t
h
a
n
,
“
C
I
C
-
I
D
S
2
0
2
7
d
a
t
a
se
t
,
”
K
a
g
g
l
e
,
2
0
1
7
.
(
a
c
c
e
sse
d
A
u
g
.
2
0
,
2
0
2
4
)
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
w
w
w
.
k
a
g
g
l
e
.
c
o
m
/
d
a
t
a
se
t
s/
c
h
e
t
h
u
h
n
/
n
e
t
w
o
r
k
-
i
n
t
r
u
si
o
n
-
d
a
t
a
s
e
t
.
[
2
3
]
S
.
S
o
n
g
ma
,
T
.
S
a
t
h
u
p
h
a
n
,
a
n
d
T
.
P
a
mu
t
h
a
,
“
O
p
t
i
m
i
z
i
n
g
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
s
y
st
e
ms
i
n
t
h
r
e
e
p
h
a
s
e
s
o
n
t
h
e
C
S
E
-
C
I
C
-
I
D
S
-
2
0
1
8
d
a
t
a
se
t
,
”
C
o
m
p
u
t
e
rs
,
v
o
l
.
1
2
,
n
o
.
1
2
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
c
o
m
p
u
t
e
r
s
1
2
1
2
0
2
4
5
.
[
2
4
]
T.
W
u
e
t
a
l
.
,
“
A
mo
d
i
f
i
e
d
g
o
r
i
l
l
a
t
r
o
o
p
s
o
p
t
i
m
i
z
e
r
f
o
r
g
l
o
b
a
l
o
p
t
i
m
i
z
a
t
i
o
n
p
r
o
b
l
e
m,
”
A
p
p
l
i
e
d
S
c
i
e
n
c
e
s
,
v
o
l
.
1
2
,
n
o
.
1
9
,
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
a
p
p
1
2
1
9
1
0
1
4
4
.
[
2
5
]
E.
S
.
G
h
i
t
h
a
n
d
F
.
A
.
A
.
T
o
l
b
a
,
“
Tu
n
i
n
g
P
I
D
c
o
n
t
r
o
l
l
e
r
s
b
a
s
e
d
o
n
h
y
b
r
i
d
a
r
i
t
h
me
t
i
c
o
p
t
i
m
i
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m
a
n
d
a
r
t
i
f
i
c
i
a
l
g
o
r
i
l
l
a
t
r
o
o
p
o
p
t
i
mi
z
a
t
i
o
n
f
o
r
m
i
c
r
o
-
r
o
b
o
t
i
c
s
s
y
st
e
ms
,”
I
E
EE
Ac
c
e
ss
,
v
o
l
.
1
1
,
p
p
.
2
7
1
3
8
–
2
7
1
5
4
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
3
.
3
2
5
8
1
8
7
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
S
h
a
li
n
i
S
h
a
r
m
a
c
o
m
p
lete
d
P
h
.
D.
i
n
C
o
m
p
u
ter
S
c
ien
c
e
fr
o
m
Ja
m
ia
M
il
li
a
Isla
m
ia
Un
iv
e
rsity
,
Ne
w
De
lh
i,
I
n
d
ia.
S
h
e
a
c
q
u
ired
a
M
a
ste
r
o
f
Tec
h
n
o
l
o
g
y
in
In
f
o
rm
a
ti
o
n
Tec
h
n
o
l
o
g
y
fro
m
G
u
ru
G
o
b
i
n
d
S
in
g
h
Un
iv
e
rsity
,
De
lh
i
i
n
2
0
1
1
.
S
h
e
wo
r
k
e
d
a
s
As
sista
n
t
P
ro
fe
ss
o
r
at
S
h
a
rd
a
Un
i
v
e
rsity
d
u
rin
g
2
0
1
1
-
2
0
1
5
.
S
h
e
h
a
s
se
rv
e
d
IT
o
rg
a
n
iza
ti
o
n
s
in
f
o
g
a
i
n
In
d
ia,
IRIS
S
o
f
twa
re
a
n
d
Ac
c
e
n
tu
re
fro
m
2
0
1
5
o
n
wa
rd
s.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
sh
a
li
n
ish
a
rm
a
1
9
8
0
@
g
m
a
il
.
c
o
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.