I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
1
5
,
No
.
4
,
A
u
g
u
s
t 2
0
2
6
,
p
p
.
3
7
0
3
~
3
7
1
1
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
1
4
pp
3
7
0
3
-
3
7
1
1
3703
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Deep learni
ng
int
rusio
n det
e
ction
f
o
r so
ft
wa
re
-
d
efin
ed
networking
usi
ng
sy
nthet
ic
mino
rit
y
ov
ersa
mpling
P
ra
j
wa
la
s
im
ha
Sin
du
g
a
t
t
a
Na
g
a
ra
j
a
1
,
Na
v
y
a
Ra
j
a
s
heka
ra
2
,
P
us
hp
a
B
a
ng
a
lo
re
Ra
m
esh
3
,
Druv
a
K
um
a
r
Sid
da
ra
j
u
4
,
Sa
ntho
s
h
K
um
a
r
Ra
ma
cha
nd
ra
g
o
wda
4
,
T
rupti
Sh
ripa
d
T
a
g
a
re
4
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
n
d
E
n
g
i
n
e
e
r
i
n
g
(
C
y
b
e
r
S
e
c
u
r
i
t
y
)
,
S
c
h
o
o
l
o
f
En
g
i
n
e
e
r
i
n
g
,
D
a
y
a
n
a
n
d
a
S
a
g
a
r
U
n
i
v
e
r
si
t
y
,
B
e
n
g
a
l
u
r
u
,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s a
n
d
C
o
m
mu
n
i
c
a
t
i
o
n
En
g
i
n
e
e
r
i
n
g
,
S
c
h
o
o
l
o
f
E
n
g
i
n
e
e
r
i
n
g
,
D
a
y
a
n
a
n
d
a
S
a
g
a
r
U
n
i
v
e
r
si
t
y
,
B
e
n
g
a
l
u
r
u
,
I
n
d
i
a
3
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s a
n
d
T
e
l
e
c
o
mm
u
n
i
c
a
t
i
o
n
E
n
g
i
n
e
e
r
i
n
g
,
D
a
y
a
n
a
n
d
a
S
a
g
a
r
C
o
l
l
e
g
e
o
f
E
n
g
i
n
e
e
r
i
n
g
,
B
e
n
g
a
l
u
r
u
,
I
n
d
i
a
4
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s a
n
d
C
o
m
mu
n
i
c
a
t
i
o
n
En
g
i
n
e
e
r
i
n
g
,
D
a
y
a
n
a
n
d
a
S
a
g
a
r
C
o
l
l
e
g
e
o
f
En
g
i
n
e
e
r
i
n
g
,
B
e
n
g
a
l
u
r
u
,
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
Dec
9
,
2
0
2
4
R
ev
is
ed
Ma
y
2
5
,
2
0
2
6
Acc
ep
ted
J
u
l 9
,
2
0
2
6
Th
is
a
rti
c
le
p
ro
p
o
se
s
a
n
a
d
v
a
n
c
e
d
m
e
th
o
d
fo
r
n
e
tw
o
rk
in
tr
u
sio
n
d
e
tec
ti
o
n
u
sin
g
a
c
o
m
b
in
a
ti
o
n
o
f
re
c
u
rre
n
t
n
e
u
ra
l
n
e
two
r
k
s
(RNN
s),
sp
e
c
ifi
c
a
ll
y
lo
n
g
sh
o
rt
-
term
m
e
m
o
ry
(LS
T
M
),
g
a
t
e
d
re
c
u
rre
n
t
u
n
it
s
(G
RU),
a
n
d
b
i
d
irec
ti
o
n
a
l
lo
n
g
sh
o
rt
-
term
m
e
m
o
ry
(BiL
S
TM
)
m
o
d
e
ls,
e
n
h
a
n
c
e
d
b
y
sy
n
th
e
ti
c
m
in
o
rit
y
o
v
e
rsa
m
p
li
n
g
tec
h
n
i
q
u
e
(
S
M
OT
E
)
to
a
d
d
re
ss
c
las
s
imb
a
lan
c
e
i
n
d
a
tas
e
ts
li
k
e
n
e
two
r
k
se
c
u
rit
y
la
b
o
ra
to
r
y
–
k
n
o
wle
d
g
e
d
isc
o
v
e
ry
in
d
a
tab
a
se
s
(
NSL
-
KDD
)
.
Th
e
m
e
th
o
d
a
ims
to
a
c
c
u
ra
tely
c
las
sify
n
e
two
rk
traffic
b
y
lea
rn
in
g
tem
p
o
ra
l
p
a
tt
e
r
n
s
o
f
b
o
th
n
o
r
m
a
l
a
n
d
m
a
li
c
io
u
s
a
c
ti
v
it
ies
.
S
M
OTE
is
e
m
p
lo
y
e
d
t
o
b
a
lan
c
e
th
e
d
a
tas
e
t,
e
n
su
rin
g
th
a
t
u
n
d
e
rre
p
re
se
n
ted
a
t
tac
k
ty
p
e
s
re
c
e
iv
e
a
d
e
q
u
a
te
m
o
d
e
l
a
tt
e
n
ti
o
n
,
th
e
re
b
y
imp
r
o
v
in
g
m
o
d
e
l
r
o
b
u
stn
e
ss
.
Th
e
p
ro
p
o
se
d
m
o
d
e
ls
(LS
TM
,
G
RU,
a
n
d
BiL
S
TM
)
a
re
train
e
d
a
n
d
e
v
a
lu
a
ted
o
n
th
e
NSL
-
KD
D
d
a
tas
e
t,
with
h
y
p
e
rp
a
ra
m
e
ter
tu
n
i
n
g
p
e
rfo
rm
e
d
th
r
o
u
g
h
Ra
n
d
o
m
ize
d
S
e
a
rc
h
CV
fo
r
o
p
ti
m
a
l
p
e
rfo
rm
a
n
c
e
.
Th
e
re
su
lt
s
sh
o
w
a
sig
n
ifi
c
a
n
t
imp
ro
v
e
m
e
n
t
in
a
c
c
u
ra
c
y
,
p
re
c
isio
n
,
re
c
a
ll
,
a
n
d
F
1
-
s
c
o
re
,
with
BiL
S
TM
d
e
m
o
n
stra
ti
n
g
t
h
e
h
i
g
h
e
st
p
e
rfo
rm
a
n
c
e
,
a
c
h
iev
i
n
g
n
e
a
r
-
p
e
rfe
c
t
c
las
sifica
ti
o
n
re
su
lt
s
(9
9
.
5
%
a
c
c
u
ra
c
y
).
Th
is
m
e
th
o
d
n
o
t
o
n
l
y
m
it
ig
a
tes
th
e
issu
e
o
f
c
las
s
imb
a
lan
c
e
b
u
t
a
lso
lev
e
ra
g
e
s
th
e
p
o
we
r
o
f
RNN
s
fo
r
se
q
u
e
n
c
e
m
o
d
e
li
n
g
,
o
ffe
ri
n
g
a
p
r
o
m
isin
g
s
o
lu
ti
o
n
f
o
r
e
ffe
c
ti
v
e
in
tr
u
sio
n
d
e
tec
ti
o
n
in
m
o
d
e
rn
n
e
two
rk
s.
K
ey
w
o
r
d
s
:
Gate
d
r
ec
u
r
r
en
t
u
n
its
I
n
tr
u
s
i
o
n
d
ete
cti
o
n
R
ec
u
r
r
en
t n
eu
r
al
n
etwo
r
k
s
R
o
b
u
s
tn
ess
Sy
n
th
etic
m
in
o
r
ity
o
v
er
s
am
p
li
n
g
tech
n
iq
u
e
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
:
Pra
jwalasim
h
a
Sin
d
u
g
atta
Nag
ar
aja
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
an
d
E
n
g
in
ee
r
in
g
(
C
y
b
er
S
ec
u
r
ity
)
,
Sch
o
o
l o
f
E
n
g
i
n
ee
r
in
g
Day
an
an
d
a
Sag
a
r
Un
iv
er
s
ity
B
en
g
alu
r
u
,
Kar
n
atak
a,
I
n
d
ia
E
m
ail: p
r
ajwa
lasi
m
h
a.
s
n
1
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
As
th
e
in
ter
n
et
g
ets
lar
g
er
a
n
d
lar
g
er
a
n
d
cy
b
er
attac
k
s
m
o
r
e
s
o
p
h
is
ticated
,
n
etwo
r
k
s
ec
u
r
ity
i
s
b
ec
o
m
in
g
a
s
ig
n
if
ican
t
is
s
u
e
f
o
r
b
o
th
th
e
p
u
b
lic
an
d
p
r
i
v
ate
s
ec
to
r
s
.
T
o
b
e
ab
le
to
d
etec
t
an
d
d
e
f
en
d
ag
ain
s
t
th
ese
ev
er
-
ch
a
n
g
in
g
th
r
ea
ts
,
c
o
m
p
u
ter
n
etwo
r
k
s
m
u
s
t
b
e
m
o
n
ito
r
ed
an
d
p
r
o
tecte
d
u
s
in
g
in
tr
u
s
io
n
d
etec
tio
n
s
y
s
tem
s
(
I
DS)
.
B
u
t
o
ld
d
etec
tio
n
tech
n
iq
u
es
lik
e
s
ig
n
atu
r
e
-
b
ased
an
d
a
n
o
m
aly
d
etec
t
io
n
ar
e
in
ef
f
ec
ti
v
e
ag
ain
s
t
n
o
v
el,
p
o
ly
m
o
r
p
h
ic
,
a
n
d
u
n
k
n
o
w
n
attac
k
s
.
T
o
o
v
er
c
o
m
e
th
ese
d
r
awb
ac
k
s
,
th
e
ap
p
licatio
n
o
f
m
ac
h
i
n
e
lear
n
in
g
(
ML
)
an
d
d
ee
p
lear
n
in
g
(
DL
)
m
eth
o
d
s
h
as
s
p
r
u
n
g
u
p
as
a
p
o
ten
tial
s
o
lu
tio
n
.
Su
ch
ad
v
an
ce
d
tech
n
iq
u
es
ar
e
ab
le
to
e
x
tr
ac
t
in
tr
icate
p
atter
n
s
f
r
o
m
d
ata,
an
d
a
d
ap
tiv
ely
r
esp
o
n
d
to
n
e
w
th
r
ea
ts
,
m
ak
i
n
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
4
,
Au
g
u
s
t 2
0
2
6
:
3
7
0
3
-
3
7
1
1
3704
th
em
p
ar
ticu
lar
ly
s
u
ited
to
th
e
co
n
tex
t
o
f
s
o
f
twar
e
-
d
ef
i
n
ed
n
etwo
r
k
s
(
SDN)
,
in
w
h
ich
th
e
n
etwo
r
k
in
f
r
astru
ctu
r
e
is
p
r
o
g
r
a
m
m
ab
l
e
an
d
ca
n
b
e
d
y
n
am
ically
m
a
n
ag
ed
[
1
]
,
[
2
]
.
R
ec
u
r
r
en
t
n
eu
r
al
n
etwo
r
k
s
(
R
NNs
)
an
d
,
m
o
r
e
p
a
r
ticu
lar
ly
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
L
STM
)
n
etwo
r
k
s
,
h
av
e
b
ee
n
p
r
o
m
is
in
g
an
d
u
s
ed
ef
f
ec
tiv
ely
in
in
tr
u
s
io
n
d
etec
tio
n
.
L
STM
s
ar
e
p
ar
ticu
lar
ly
s
u
itab
le
f
o
r
s
eq
u
en
tial
d
ata,
an
d
in
tr
u
s
io
n
d
etec
tio
n
r
elies
o
n
th
e
d
etec
tio
n
o
f
p
atter
n
s
o
v
er
tim
e.
Su
ch
n
etwo
r
k
s
ca
n
lear
n
lo
n
g
-
te
r
m
d
ep
e
n
d
en
cies
an
d
ar
e
th
u
s
well
s
u
ited
to
d
e
tect
s
u
b
tle
attac
k
s
th
at
ar
e
lik
ely
to
o
cc
u
r
o
v
er
a
lo
n
g
er
tim
ef
r
am
e
[
3
]
.
R
ec
en
t
r
esu
lts
s
h
o
wed
L
STM
s
ar
e
ab
le
to
ef
f
ec
tiv
el
y
id
en
tif
y
b
o
th
k
n
o
w
n
an
d
u
n
k
n
o
wn
n
etwo
r
k
in
tr
u
s
io
n
s
[
4
]
.
Fu
r
th
e
r
,
o
th
er
t
y
p
es
o
f
R
NNs,
s
u
ch
as
g
ated
r
ec
u
r
r
en
t
u
n
it
(
GR
Us
)
h
av
e
also
b
ee
n
u
s
ed
in
I
DS
with
s
im
ilar
ad
v
an
tag
es
in
ter
m
s
o
f
co
m
p
u
tatio
n
al
ef
f
icie
n
cy
an
d
s
im
p
licity
o
v
er
L
STM
s
[
5
]
.
On
e
o
f
th
e
m
ajo
r
d
if
f
icu
lties
in
th
e
ap
p
licatio
n
o
f
R
NN
b
ased
m
o
d
els
in
I
DS
is
th
e
p
r
o
b
lem
o
f
class
im
b
alan
ce
.
T
h
is
is
esp
ec
ially
im
p
o
r
tan
t
wh
en
th
er
e
is
a
lar
g
e
im
b
alan
ce
in
th
e
am
o
u
n
t
o
f
n
o
r
m
al
tr
af
f
ic
an
d
attac
k
tr
af
f
ic
in
t
h
e
d
atab
as
e,
an
d
t
h
e
attac
k
s
ar
e
g
en
er
ally
u
n
d
e
r
-
r
ep
r
esen
ted
.
T
h
ese
d
atasets
ar
e
n
o
t
b
alan
ce
d
,
wh
ich
ca
n
r
esu
lt
in
s
k
ewe
d
m
o
d
els
th
at
f
av
o
u
r
th
e
m
ajo
r
ity
class
(
n
o
r
m
al
tr
af
f
i
c)
,
an
d
n
e
g
lect
th
e
m
in
o
r
ity
class
es
(
attac
k
s
)
.
T
o
ad
d
r
ess
th
is
p
r
o
b
lem
,
th
er
e
ar
e
s
ev
er
al
tech
n
iq
u
es
th
at
ca
n
b
e
u
tili
ze
d
,
s
u
ch
as
s
y
n
th
etic
m
in
o
r
ity
o
v
er
s
am
p
l
in
g
tech
n
iq
u
e
(
SMOT
E
)
.
SM
OT
E
cr
ea
tes
s
y
n
th
etic
s
am
p
l
es
f
o
r
t
h
e
m
in
o
r
ity
class
,
wh
ich
h
elp
s
to
b
alan
ce
th
e
s
am
p
le
s
et
an
d
en
h
an
ce
s
th
e
m
o
d
el'
s
p
er
f
o
r
m
an
ce
in
d
etec
tin
g
r
ar
e
attac
k
s
[
6
]
.
SMOT
E
o
v
er
co
m
es
th
e
im
b
alan
ce
d
p
r
o
b
lem
an
d
im
p
r
o
v
es
th
e
r
ep
r
esen
tatio
n
o
f
m
in
o
r
ity
class
es
s
o
th
at
th
e
I
DS
is
s
tr
o
n
g
to
d
if
f
er
en
t k
in
d
s
o
f
attac
k
s
.
B
id
ir
ec
tio
n
al
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
B
iLST
M
)
n
etwo
r
k
s
ca
n
b
e
u
s
ed
to
f
u
r
th
e
r
im
p
r
o
v
e
p
er
f
o
r
m
an
ce
b
esid
es
ad
d
r
ess
in
g
class
im
b
alan
ce
.
I
n
a
d
d
iti
o
n
to
o
p
er
atin
g
i
n
a
f
o
r
war
d
d
ir
ec
tio
n
,
B
iLST
Ms
ca
n
also
o
p
er
ate
in
a
b
ac
k
war
d
d
ir
ec
tio
n
,
allo
win
g
th
e
n
etw
o
r
k
to
lear
n
a
r
ich
er
s
et
o
f
tem
p
o
r
al
r
elatio
n
s
h
ip
s
.
T
h
is
ca
n
b
e
v
er
y
b
en
ef
icial
in
I
DS
b
ec
au
s
e
it
is
im
p
o
r
tan
t
to
k
n
o
w
th
e
ch
ar
ac
ter
is
tics
o
f
b
o
th
cu
r
r
en
t
an
d
f
u
tu
r
e
t
r
af
f
ic
t
o
id
en
tif
y
lo
n
g
-
r
u
n
n
in
g
attac
k
b
eh
a
v
io
r
s
th
at
ca
n
b
e
im
p
lem
en
ted
o
v
er
a
tim
e
r
an
g
e
[
7
]
.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
in
th
e
co
n
te
x
t o
f
I
DS u
s
in
g
B
iLST
M
an
d
SMOT
E
aim
s
to
:
−
Acc
u
r
ate
an
d
ef
f
ec
tiv
e
in
tr
u
s
i
o
n
d
etec
tio
n
:
u
tili
ze
B
iLST
M
-
SMOT
E
f
o
r
ac
cu
r
ate
d
etec
tio
n
o
f
k
n
o
w
n
an
d
u
n
k
n
o
w
n
n
etwo
r
k
in
tr
u
s
io
n
s
with
h
ig
h
p
r
ec
is
io
n
an
d
r
ec
all.
−
C
las
s
im
b
alan
ce
:
SMOT
E
is
u
s
ed
to
cr
ea
te
s
y
n
th
etic
ex
a
m
p
les
o
f
attac
k
in
s
tan
ce
s
to
ad
d
r
ess
class
im
b
alan
ce
an
d
en
h
a
n
ce
th
e
m
o
d
el'
s
d
etec
tio
n
o
f
r
ar
e
attac
k
s
.
−
R
ea
l
-
tim
e
an
d
s
ca
lab
le
d
etec
ti
o
n
:
th
e
m
o
d
el
s
h
o
u
ld
b
e
ab
le
to
r
u
n
in
r
ea
l
-
tim
e
en
v
ir
o
n
m
e
n
ts
,
h
an
d
lin
g
lar
g
e
v
o
lu
m
es
o
f
n
etwo
r
k
tr
a
f
f
ic
wh
ile
m
ain
tain
in
g
h
ig
h
d
etec
tio
n
r
ates.
Dete
ctio
n
m
o
d
el
s
h
o
u
ld
b
e
s
ca
lab
le
an
d
ab
le
to
p
r
o
ce
s
s
lar
g
e
v
o
lu
m
es
o
f
n
etwo
r
k
t
r
af
f
ic
in
r
ea
l
-
tim
e
wh
ile
m
ain
tain
in
g
h
ig
h
d
etec
tio
n
r
ates.
I
n
th
is
ar
ticle,
a
n
o
v
el
m
eth
o
d
is
p
r
o
p
o
s
ed
to
co
m
b
in
e
th
e
SMOT
E
-
en
h
an
ce
d
R
NN
m
o
d
el
s
,
s
u
ch
as
L
STM
,
GR
U
an
d
B
iLST
M
,
f
o
r
n
etwo
r
k
in
tr
u
s
io
n
d
etec
tio
n
in
SDN
en
v
ir
o
n
m
en
t.
T
h
is
ap
p
r
o
ac
h
u
s
es
SMOT
E
f
o
r
d
ata
b
ala
n
cin
g
a
n
d
is
ca
p
ab
le
o
f
lev
er
a
g
in
g
t
h
e
p
o
wer
o
f
R
NNs
to
m
o
d
el
te
m
p
o
r
al
d
e
p
en
d
e
n
cy
in
n
etwo
r
k
tr
af
f
ic,
o
f
f
er
s
a
co
m
p
r
eh
e
n
s
iv
e
s
o
lu
tio
n
to
d
etec
t
a
wid
e
v
ar
iety
o
f
n
etw
o
r
k
in
t
r
u
s
io
n
s
.
T
h
e
m
o
d
el'
s
p
er
f
o
r
m
a
n
ce
is
test
ed
with
th
e
n
etwo
r
k
s
ec
u
r
ity
lab
o
r
ato
r
y
–
k
n
o
wled
g
e
d
is
co
v
er
y
i
n
d
atab
ases
(
NSL
-
KDD
)
d
ataset,
ac
h
iev
in
g
th
e
b
est
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
an
d
F1
-
s
co
r
es
in
c
o
m
p
ar
is
o
n
to
o
th
er
d
atasets
.
T
h
e
B
iLST
M
m
o
d
el
is
f
o
u
n
d
t
o
g
iv
e
h
ig
h
class
if
icatio
n
ac
cu
r
ac
y
i
n
b
o
t
h
th
e
n
o
r
m
al
an
d
attac
k
class
es
as
co
m
p
ar
ed
to
th
e
co
n
v
en
tio
n
al
R
NN
v
ar
ian
ts
.
B
ased
o
n
th
e
r
esu
lts
,
it
ca
n
b
e
in
f
er
r
ed
th
at
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
wo
u
l
d
p
r
o
v
id
e
a
s
tr
o
n
g
s
y
s
tem
f
o
r
r
e
al
-
tim
e
an
d
a
d
ap
tiv
e
I
DS
f
o
r
m
o
d
e
r
n
n
etwo
r
k
s
.
Mo
r
eo
v
er
,
o
u
r
g
o
al
is
to
em
p
h
asize
th
e
n
ee
d
f
o
r
u
s
in
g
DL
m
eth
o
d
s
an
d
o
v
er
s
am
p
lin
g
ap
p
r
o
ac
h
es
to
g
eth
e
r
to
o
v
er
co
m
e
d
ata
im
b
alan
ce
a
n
d
o
th
er
ch
allen
g
es.
Ho
wev
er
,
with
an
in
c
r
ea
s
in
g
th
r
ea
t
o
f
cy
b
er
-
attac
k
,
it
is
ess
en
tial
th
at
I
DS
ar
e
m
o
r
e
ad
ap
tab
le,
ca
n
p
r
o
ce
s
s
co
m
p
le
x
,
lar
g
e
-
s
ca
le
d
ata
an
d
th
at
th
ey
ar
e
h
ig
h
ly
ac
cu
r
ate
an
d
p
r
ec
is
e
an
d
r
ec
all
in
all
ty
p
es o
f
attac
k
s
.
2.
L
I
T
E
R
AT
U
RE
SU
RVE
Y
W
ith
th
e
r
ap
id
d
e
v
elo
p
m
e
n
t
o
f
n
etwo
r
k
s
ca
le
a
n
d
c
o
m
p
lex
ity
,
n
etwo
r
k
i
n
tr
u
s
io
n
d
et
ec
tio
n
h
as
b
ec
o
m
e
a
h
o
t
to
p
ic
f
o
r
r
ese
ar
ch
.
T
r
ad
itio
n
al
s
ig
n
atu
r
e
-
b
ased
tech
n
iq
u
es
an
d
a
b
asic
an
o
m
al
y
d
etec
tio
n
ap
p
r
o
ac
h
h
av
e
p
r
o
v
en
to
b
e
i
n
ef
f
ec
tiv
e
in
th
e
d
etec
tio
n
o
f
n
o
v
el
attac
k
ty
p
es,
r
esu
ltin
g
i
n
an
in
cr
ea
s
e
o
f
th
e
u
s
e
o
f
ML
an
d
DL
tech
n
iq
u
es
.
T
h
e
ad
v
an
tag
e
o
f
th
ese
m
eth
o
d
s
is
th
at
th
e
p
atter
n
s
ar
e
lea
r
n
ed
au
to
m
atica
lly
f
r
o
m
th
e
d
ata
an
d
th
e
m
et
h
o
d
s
ad
ap
t
to
n
ew
th
r
ea
ts
wi
th
o
u
t
u
s
in
g
p
r
e
-
d
ef
in
ed
attac
k
s
ig
n
atu
r
es.
T
h
e
f
o
llo
win
g
is
a
b
r
ie
f
o
v
e
r
v
iew
o
f
s
o
m
e
im
p
o
r
tan
t liter
atu
r
e
in
th
is
ar
ea
.
T
r
ad
itio
n
al
I
DS
m
eth
o
d
s
ar
e
b
ased
o
n
p
r
ev
io
u
s
ly
k
n
o
wn
a
ttack
p
atter
n
s
,
wh
ich
ar
e
o
n
ly
ef
f
ec
tiv
e
ag
ain
s
t
k
n
o
w
n
attac
k
s
.
B
u
t
th
is
k
in
d
o
f
s
y
s
tem
h
as
d
if
f
icu
lty
in
d
etec
tin
g
n
ew
ze
r
o
-
d
ay
attac
k
s
an
d
p
o
ly
m
o
r
p
h
ic
attac
k
s
th
at
ch
an
g
e
o
v
er
tim
e.
An
o
m
aly
-
b
a
s
ed
m
eth
o
d
s
,
h
o
wev
er
,
id
en
t
if
y
an
o
m
alies
th
at
d
ev
iate
f
r
o
m
n
o
r
m
al
n
etwo
r
k
p
atter
n
s
b
u
t
f
r
eq
u
e
n
tly
p
r
o
d
u
c
e
m
an
y
f
alse
p
o
s
itiv
es
b
ec
au
s
e
to
d
ay
'
s
n
etwo
r
k
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dee
p
lea
r
n
in
g
in
tr
u
s
io
n
d
etec
tio
n
fo
r
s
o
ftw
a
r
e
-
d
efin
ed
…
(
P
r
a
jw
a
la
s
imh
a
S
in
d
u
g
a
tta
N
a
g
a
r
a
ja
)
3705
o
p
er
ate
in
v
ar
y
in
g
m
o
d
es.
C
o
n
s
eq
u
en
tly
,
s
u
c
h
tr
ad
itio
n
al
s
tr
ateg
ies
ar
e
b
ec
o
m
in
g
in
ef
f
ec
tiv
e
ag
ain
s
t
th
e
ad
v
an
ce
d
attac
k
s
[
8
]
.
I
n
v
iew
o
f
th
e
d
r
awb
ac
k
s
o
f
c
o
n
v
e
n
t
io
n
al
I
DS,
r
esear
ch
er
s
h
av
e
t
u
r
n
ed
to
th
e
u
s
e
o
f
ML
tech
n
iq
u
es.
I
n
i
n
tr
u
s
io
n
d
etec
tio
n
,
c
lass
if
icatio
n
alg
o
r
ith
m
s
s
u
ch
as
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
s
(
SVM)
,
d
ec
is
io
n
t
r
ee
s
(
DT
)
,
an
d
r
an
d
o
m
f
o
r
est
s
(
R
F)
ar
e
wid
ely
u
s
ed
b
ec
au
s
e
th
ey
ar
e
ef
f
ec
tiv
e
at
class
if
icatio
n
.
T
h
ese
m
eth
o
d
s
,
h
o
wev
e
r
,
d
ep
en
d
o
n
m
a
n
u
ally
d
esig
n
e
d
f
ea
t
u
r
es,
an
d
r
esu
lts
ar
e
n
o
t a
lway
s
s
ati
s
f
ac
to
r
y
wh
en
wo
r
k
in
g
with
lar
g
e
an
d
co
m
p
l
ex
d
ata
s
ets with
h
ig
h
n
o
n
lin
e
ar
ity
b
etwe
en
f
ea
tu
r
es
[
9
]
,
[
1
0
]
.
I
n
c
o
n
t
r
a
s
t
,
D
L
m
o
d
e
ls
,
p
a
r
t
i
cu
l
a
r
l
y
R
NN
s
,
a
r
e
s
h
o
w
i
n
g
a
lo
t
o
f
p
r
o
m
i
s
e
b
e
c
a
u
s
e
o
f
t
h
e
i
r
c
a
p
a
c
i
t
y
t
o
p
r
o
c
e
s
s
s
e
q
u
e
n
t
i
a
l
d
a
t
a
a
n
d
i
d
en
t
i
f
y
l
o
n
g
-
t
e
r
m
d
e
p
e
n
d
e
n
c
i
e
s
.
A
m
o
n
g
t
h
e
s
e
,
L
S
T
M
n
e
tw
o
r
k
s
a
r
e
w
i
d
e
l
y
u
s
e
d
i
n
i
n
t
r
u
s
i
o
n
d
e
t
e
ct
i
o
n
p
r
o
b
l
e
m
s
wh
e
r
e
t
h
e
a
t
t
a
c
k
s
a
r
e
te
m
p
o
r
a
l
p
a
t
t
e
r
n
s
,
s
u
c
h
a
s
t
r
a
f
f
i
c
b
u
r
s
t
s
an
d
a
t
t
a
c
k
s
i
g
n
at
u
r
e
s
th
at
ch
an
g
e
o
v
er
tim
e
[
1
1
]
,
[
1
2
]
.
T
h
e
GR
Us,
wh
ich
ar
e
a
v
a
r
ian
t
o
f
R
NNs,
also
h
av
e
s
im
ilar
ad
v
an
tag
es,
b
u
t
with
a
less
er
n
u
m
b
er
o
f
p
a
r
am
eter
s
,
th
u
s
b
ein
g
also
m
o
r
e
ec
o
n
o
m
ical
in
ter
m
s
o
f
co
m
p
u
tat
io
n
[
1
3
]
,
[
1
4
]
.
On
e
o
f
th
e
g
r
ea
test
d
if
f
icu
lties
en
co
u
n
ter
e
d
in
lear
n
in
g
m
o
d
e
ls
f
o
r
ac
cu
r
ate
in
tr
u
s
io
n
d
etec
tio
n
is
th
e
class
im
b
alan
ce
is
s
u
e,
wh
ich
o
cc
u
r
s
wh
en
th
e
r
e
is
a
h
ig
h
r
atio
o
f
attac
k
v
er
s
u
s
n
o
r
m
al
tr
af
f
ic.
T
h
is
d
is
p
r
o
p
o
r
tio
n
ca
n
d
r
asti
ca
lly
af
f
ec
t
th
e
m
o
d
el'
s
p
er
f
o
r
m
an
ce
in
m
in
o
r
ity
class
es
(
at
tack
s
)
an
d
h
av
e
a
s
ig
n
if
ican
t
im
p
ac
t
on
th
e
m
ajo
r
ity
class
(
n
o
r
m
al
tr
af
f
ic)
.
I
n
r
esp
o
n
s
e
to
th
is
,
SMOT
E
h
as
b
ee
n
h
ig
h
l
y
u
tili
ze
d
.
T
h
e
SMOT
E
alg
o
r
ith
m
is
d
es
ig
n
ed
to
ad
d
s
y
n
th
etic
i
n
s
tan
ce
s
to
th
e
m
in
o
r
ity
class
es,
th
e
r
eb
y
b
alan
cin
g
th
e
d
ata
an
d
e
n
h
an
cin
g
th
e
m
o
d
el'
s
p
er
f
o
r
m
a
n
ce
in
id
e
n
tify
in
g
r
ar
e
attac
k
ty
p
es
[
1
5
]
.
T
h
e
p
er
f
o
r
m
an
ce
o
f
in
tr
u
s
io
n
d
etec
tio
n
h
as
b
ee
n
f
u
r
th
er
en
h
an
ce
d
b
y
r
ec
en
t
d
e
v
elo
p
m
en
ts
in
B
iLST
M
n
etwo
r
k
s
.
B
iL
STM
s
r
ea
d
s
eq
u
en
ce
s
f
o
r
war
d
an
d
b
ac
k
war
d
,
lear
n
in
g
ab
o
u
t
th
e
co
n
tex
t
n
o
t
o
n
l
y
f
r
o
m
th
e
p
r
ev
io
u
s
s
eq
u
en
ce
o
f
tr
af
f
ic,
b
u
t
also
th
e
f
u
t
u
r
e
o
n
e.
T
h
is
is
esp
ec
ially
u
s
ef
u
l
in
id
en
tify
in
g
ad
v
an
ce
d
attac
k
s
th
at
co
u
ld
i
n
v
o
lv
e
m
u
lti
-
s
tep
ac
tio
n
s
an
d
m
ay
b
e
p
er
f
o
r
m
e
d
at
d
if
f
er
e
n
t
tim
es
,
an
d
r
eq
u
ir
e
co
n
tex
t
f
r
o
m
p
ast
an
d
f
u
tu
r
e
n
etwo
r
k
co
n
d
itio
n
s
.
B
iLST
Ms
h
av
e
b
ee
n
f
o
u
n
d
to
b
e
s
u
p
e
r
io
r
to
co
n
v
en
tio
n
al
L
STM
n
etwo
r
k
s
in
s
ev
er
al
I
DS
s
ce
n
ar
io
s
b
ec
au
s
e
th
ey
ar
e
b
etter
ab
le
to
ca
p
tu
r
e
t
h
e
tem
p
o
r
al
p
atter
n
s
in
n
etwo
r
k
tr
af
f
ic
[
1
6
]
.
So
m
e
s
tu
d
ies
h
av
e
ex
am
i
n
e
d
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
ML
an
d
DL
-
b
ased
m
o
d
els
o
f
I
DS
u
s
in
g
b
en
ch
m
ar
k
d
atasets
s
u
ch
as
K
DD
C
u
p
9
9
,
NSL
-
KDD,
an
d
C
I
C
-
I
DS2
0
1
7
.
A
p
o
p
u
lar
d
ata
s
et
to
test
a
n
etwo
r
k
I
DS
is
th
e
NSL
-
KDD
d
ataset,
wh
ich
co
n
tain
s
a
co
m
p
lete
s
et
o
f
n
o
r
m
al
an
d
attac
k
tr
af
f
i
c
s
am
p
les.
On
th
ese
d
atasets
,
DL
m
o
d
els
s
u
ch
as
L
STM
an
d
B
iLST
M
h
av
e
b
e
en
f
o
u
n
d
to
b
e
m
o
r
e
ac
c
u
r
ate
,
p
r
ec
is
e,
an
d
r
ec
all
an
d
F1
-
s
co
r
e
th
a
n
tr
ad
itio
n
al
ap
p
r
o
ac
h
es.
E
v
en
th
ese
m
o
d
els
h
av
e
p
r
o
b
lem
s
o
f
o
v
er
f
i
ttin
g
an
d
d
em
a
n
d
a
s
u
b
s
tan
tial a
m
o
u
n
t o
f
lab
elled
d
ata
[
1
7
]
.
I
n
r
ec
e
n
t
y
ea
r
s
,
h
y
b
r
id
m
eth
o
d
s
th
at
co
m
b
in
e
a
s
et
o
f
ML
alg
o
r
ith
m
s
with
en
s
em
b
le
m
e
th
o
d
s
h
a
v
e
b
ee
n
p
o
p
u
lar
to
en
h
an
ce
th
e
r
o
b
u
s
tn
ess
an
d
ac
cu
r
ac
y
o
f
I
D
S.
A
co
m
b
in
atio
n
o
f
C
NN
wit
h
L
STM
o
r
GR
U
h
as
also
b
ee
n
r
ec
o
g
n
ized
as
a
h
i
g
h
ly
ef
f
ec
tiv
e
ap
p
r
o
ac
h
;
ap
ar
t
f
r
o
m
C
NN,
wh
ich
is
g
o
o
d
at
ex
tr
ac
tin
g
s
p
atial
f
ea
tu
r
es,
R
NN
is
ca
p
ab
le
o
f
c
ap
tu
r
in
g
tem
p
o
r
al
f
ea
t
u
r
es.
T
h
eir
f
u
s
ed
m
eth
o
d
was
r
ep
o
r
t
ed
to
b
e
ca
p
ab
le
o
f
h
av
in
g
h
i
g
h
er
d
etec
tio
n
p
er
f
o
r
m
an
ce
n
o
t
o
n
ly
f
o
r
th
e
k
n
o
wn
attac
k
s
b
u
t
also
f
o
r
th
e
u
n
k
n
o
wn
o
n
es
with
a
lo
wer
f
alse
p
o
s
itiv
e
r
ate
[
1
8
]
.
R
ec
en
t
s
tu
d
ies
h
av
e
also
ex
p
lo
r
ed
atten
tio
n
-
en
h
a
n
ce
d
B
iLST
M
n
etwo
r
k
s
,
ad
ap
tiv
e
o
v
er
s
am
p
lin
g
s
tr
ateg
ies,
an
d
h
y
b
r
id
DL
ar
ch
it
ec
tu
r
es
f
o
r
in
tr
u
s
io
n
d
etec
tio
n
,
d
em
o
n
s
tr
atin
g
im
p
r
o
v
e
d
r
o
b
u
s
tn
ess
ag
ain
s
t e
v
o
lv
in
g
cy
b
er
th
r
ea
ts
[
1
9
]
–
[
2
3
]
.
T
h
is
r
ev
iew
o
n
th
e
liter
atu
r
e
p
r
esen
ts
a
co
m
p
r
e
h
en
s
iv
e
s
u
r
v
ey
o
f
t
h
e
cu
r
r
en
t
r
esear
ch
p
r
o
g
r
ess
in
n
etwo
r
k
in
tr
u
s
io
n
d
etec
tio
n
b
ased
o
n
ML
an
d
DL
m
eth
o
d
s
.
I
t
h
ig
h
lig
h
ts
th
e
d
is
ad
v
an
t
ag
es
o
f
r
ely
in
g
o
n
tr
ad
itio
n
al
I
DS
s
y
s
tem
s
,
th
e
p
o
s
s
ib
ilit
y
o
f
u
tili
zin
g
m
o
r
e
ad
v
a
n
ce
d
m
eth
o
d
s
lik
e
R
NNs,
L
STM
s
,
an
d
B
iLST
Ms,
an
d
th
e
im
p
o
r
tan
ce
o
f
ad
d
r
ess
in
g
p
r
o
b
lem
s
s
u
ch
as
class
im
b
alan
ce
th
r
o
u
g
h
m
eth
o
d
s
lik
e
SMOT
E
.
Fu
r
th
er
m
o
r
e,
it
illu
s
tr
ates
th
e
ev
o
lu
tio
n
o
f
th
e
u
s
e
o
f
m
o
r
e
th
a
n
o
n
e
ML
m
o
d
el
to
en
h
an
ce
th
e
d
etec
tio
n
p
er
f
o
r
m
a
n
ce
.
3.
P
RO
P
O
SE
D
M
E
T
H
O
D
3
.
1
.
Da
t
a
prepro
ce
s
s
ing
s
t
ep
s
Data
p
r
ep
r
o
ce
s
s
in
g
is
a
cr
itica
l
p
h
ase
i
n
ML
th
at
t
r
an
s
f
o
r
m
s
r
aw
d
ata
in
to
a
clea
n
an
d
s
u
ita
b
le
f
o
r
m
at
f
o
r
m
o
d
el
tr
ain
in
g
.
I
n
th
is
s
tu
d
y
,
th
e
d
ataset
u
n
d
er
g
o
es
two
p
r
im
ar
y
p
r
e
p
r
o
ce
s
s
in
g
tech
n
iq
u
es:
f
ea
tu
r
e
n
o
r
m
aliza
tio
n
an
d
tar
g
et
la
b
el
en
co
d
i
n
g
.
T
h
ese
s
tep
s
en
s
u
r
e
th
at
th
e
in
p
u
t
f
ea
tu
r
es
a
r
e
o
n
a
co
m
p
a
r
ab
le
s
ca
le
an
d
th
at
th
e
ca
teg
o
r
ical
tar
g
et
v
ar
iab
les
ar
e
p
r
o
p
er
ly
f
o
r
m
att
ed
to
m
ee
t
th
e
co
m
p
u
tatio
n
al
r
eq
u
ir
em
e
n
ts
o
f
t
h
e
alg
o
r
ith
m
.
i)
No
r
m
aliza
tio
n
(
m
in
-
m
ax
s
ca
lin
g
)
:
th
e
Min
Ma
x
Scaler
is
u
s
ed
to
s
ca
le
f
ea
tu
r
es
in
to
th
e
r
an
g
e
[
0
,
1
]
.
T
h
is
en
s
u
r
es
th
at
all
f
ea
tu
r
es
co
n
tr
ib
u
te
eq
u
ally
to
th
e
lear
n
in
g
p
r
o
ce
s
s
an
d
p
r
ev
en
ts
d
o
m
in
a
n
ce
b
y
f
ea
tu
r
es
with
lar
g
er
m
ag
n
itu
d
es
(
1
)
.
′
=
−
−
(
1
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
4
,
Au
g
u
s
t 2
0
2
6
:
3
7
0
3
-
3
7
1
1
3706
W
h
er
e
is
o
r
ig
in
al
f
ea
tu
r
e
v
alu
e
,
is
m
in
im
u
m
v
al
u
e
o
f
th
e
f
e
atu
r
e
,
is
m
ax
im
u
m
v
alu
e
o
f
th
e
f
ea
tu
r
e
,
an
d
′
is
n
o
r
m
alize
d
f
ea
tu
r
e
v
alu
e
.
ii)
L
ab
el
en
co
d
in
g
:
la
b
els
ar
e
e
n
co
d
ed
in
to
n
u
m
e
r
ical
v
alu
es
u
s
in
g
th
e
L
ab
elE
n
c
o
d
er
f
r
o
m
th
e
Sk
lear
n
lib
r
ar
y
as
in
(
2
)
.
E
ac
h
u
n
i
q
u
e
class
lab
el
in
is
m
ap
p
ed
to
an
in
teg
er
,
m
ak
in
g
it
s
u
itab
le
f
o
r
m
o
d
els
th
at
r
eq
u
ir
e
n
u
m
er
ical
in
p
u
t
f
o
r
t
ar
g
et
lab
els.
W
h
er
e
is
o
r
ig
in
al
class
lab
els
(
ca
teg
o
r
ical)
,
is
en
co
d
ed
class
lab
els (
n
u
m
er
ic
al)
.
=
(
)
(
2
)
3
.
2
.
Sy
nthet
ic
m
ino
rit
y
o
v
er
s
a
m
pli
ng
t
ec
hn
iqu
e
Un
lik
e
a
s
im
p
le
d
u
p
licatio
n
o
f
m
in
o
r
ity
class
ex
am
p
les,
S
MO
T
E
s
o
lv
es
th
e
is
s
u
e
o
f
class
im
b
alan
ce
b
y
s
y
n
t
h
esizin
g
m
in
o
r
ity
clas
s
ex
am
p
les.
I
t
w
o
r
k
s
b
y
p
ick
i
n
g
a
s
am
p
le
f
r
o
m
m
in
o
r
ity
cl
ass
an
d
f
in
d
i
n
g
t
h
e
k
-
n
ea
r
est
n
eig
h
b
o
r
s
o
f
it.
A
r
an
d
o
m
n
eig
h
b
o
r
is
s
elec
ted
an
d
a
s
y
n
th
etic
s
am
p
le
is
p
r
o
d
u
ce
d
alo
n
g
a
lin
e
s
eg
m
en
t
b
etwe
en
th
e
o
r
ig
in
al
s
am
p
le
an
d
its
n
eig
h
b
o
r
.
A
r
a
n
d
o
m
n
u
m
b
er
λ
,
b
etwe
en
0
a
n
d
1
d
eter
m
in
es
th
e
p
o
s
itio
n
o
f
th
e
s
y
n
th
etic
s
am
p
le.
T
h
is
is
d
o
n
e
u
n
til
th
e
d
esire
d
n
u
m
b
er
o
f
s
am
p
les
is
g
en
e
r
ated
f
o
r
s
y
n
th
etic
d
ata,
th
e
d
ataset
will
b
e
b
alan
ce
d
,
an
d
t
h
e
m
o
d
el
will
b
e
b
etter
at
g
en
er
alizin
g
with
o
u
t
o
v
er
f
itti
n
g
.
R
aw
n
etwo
r
k
tr
af
f
ic
d
ata
is
th
en
n
o
r
m
alize
d
b
etwe
en
0
an
d
1
with
m
in
–
m
ax
n
o
r
m
aliza
tio
n
,
s
u
ch
th
at
th
e
co
n
tr
ib
u
tio
n
o
f
th
e
f
ea
tu
r
es
is
u
n
if
o
r
m
wh
ile
tr
ai
n
in
g
t
h
e
m
o
d
el.
L
a
b
el
en
co
d
in
g
co
n
v
e
r
ts
ca
teg
o
r
ical
class
lab
els
in
to
n
u
m
b
e
r
s
.
T
o
o
v
er
c
o
m
e
th
e
p
r
o
b
lem
o
f
class
im
b
alan
ce
,
SMOT
E
is
u
s
ed
o
n
ly
o
n
th
e
tr
ain
i
n
g
s
et
th
u
s
p
r
ev
en
tin
g
d
ata
leak
ag
e
is
s
u
e.
T
h
e
n
u
m
b
er
o
f
n
ea
r
e
s
t
n
eig
h
b
o
u
r
s
k
was
ch
o
s
en
as
5
,
b
ased
o
n
th
e
p
r
elim
in
ar
y
v
alid
atio
n
e
x
p
er
i
m
en
ts
as
well
as
s
tan
d
ar
d
I
DS
liter
atu
r
e.
Ov
er
s
am
p
lin
g
was
d
o
n
e
f
o
r
m
in
o
r
class
es to
o
b
tain
a
n
ea
r
b
alan
c
ed
class
.
T
o
ac
h
iev
e
r
e
p
r
o
d
u
ci
b
ilit
y
a
f
ix
ed
r
an
d
o
m
s
ee
d
was u
s
ed
.
T
o
im
p
r
o
v
e
th
e
r
ep
r
o
d
u
ci
b
ilit
y
an
d
im
p
lem
en
tatio
n
clar
i
ty
o
f
th
e
p
r
o
p
o
s
ed
in
tr
u
s
io
n
d
etec
tio
n
f
r
am
ewo
r
k
,
th
e
c
o
m
p
lete
co
m
p
u
tatio
n
al
wo
r
k
f
lo
w
is
p
r
es
en
ted
u
s
in
g
al
g
o
r
ith
m
ic
p
s
eu
d
o
co
d
e
.
Alg
o
r
ith
m
1
d
escr
ib
es
th
e
SMOT
E
-
b
ased
m
in
o
r
ity
class
o
v
e
r
s
am
p
lin
g
p
r
o
ce
d
u
r
e
,
wh
ile
Alg
o
r
ith
m
2
p
r
esen
ts
th
e
u
n
if
ied
tr
ain
in
g
p
r
o
ce
d
u
r
e
f
o
r
th
e
L
STM
,
GR
U,
an
d
B
iLST
M
ar
ch
itectu
r
es.
T
h
ese
alg
o
r
ith
m
s
p
r
o
v
id
e
a
co
n
cise
d
escr
ip
tio
n
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
a
n
d
f
ac
ilit
ate
f
u
tu
r
e
im
p
lem
en
tatio
n
an
d
c
o
m
p
ar
is
o
n
b
y
o
th
er
r
esear
ch
e
r
s
.
Alg
o
r
ith
m
1
.
Ps
eu
d
o
co
d
e
o
f
S
MO
T
E
-
b
ased
m
in
o
r
it
y
class
o
v
er
s
am
p
lin
g
INPUT: Dataset D, minority class samples M, oversampling ratio r, number of neighbors k
OUTPUT: Augmented dataset with synthetic samples
SET S = empty set
num_synthetic_samples = r * |M|
FOR each sample xi in M:
neighbors = find_k_nearest_neighbors(xi, M, k)
FOR n samples in neighbors:
SELECT xj randomly from neighbors
λ = random(0, 1) # Generate a random value between 0 and 1
x_new = xi + λ * (xj
-
xi)
ADD x_new to S
END FOR
END FOR
RETURN D + S
Alg
o
r
ith
m
2
.
Ps
eu
d
o
co
d
e
o
f
u
n
if
ied
DL
tr
ain
in
g
p
r
o
ce
d
u
r
e
f
o
r
L
STM
,
GR
U
,
an
d
B
iLST
M
INPUT: Sequence X = {x1, x2, ..., xT}, model_type ("LSTM", "GRU", "BiLSTM")
OUTPUT: Final hidden state(s) h_final
INITIALIZE:
h0 = initial_hidden_state # Initialize hidden state
IF model_type == "LSTM" OR model_type == "BiLSTM":
C0 = initial_cell_state # Initialize cell state
IF model_type == "LSTM":
FOR t = 1 to T: # Iterate over sequence
ft = sigmoid(Wf
* [ht
-
1, xt] + bf) # Forget gate
it = sigmoid(Wi * [ht
-
1, xt] + bi) # Input gate
Ct_candidate = tanh(WC * [ht
-
1, xt] + bC) # Candidate cell state
Ct = ft * Ct
-
1 + it * Ct_candidate # Update cell state
ot = sigmoid(Wo * [ht
-
1, xt] + bo) # Output gate
ht = ot * tanh(Ct) # Update hidden state
END FOR
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dee
p
lea
r
n
in
g
in
tr
u
s
io
n
d
etec
tio
n
fo
r
s
o
ftw
a
r
e
-
d
efin
ed
…
(
P
r
a
jw
a
la
s
imh
a
S
in
d
u
g
a
tta
N
a
g
a
r
a
ja
)
3707
h_final = ht
ELSE IF model_type == "GRU":
FOR t = 1 to T: # Iterate over sequence
zt = sigmoid(Wz * [ht
-
1, xt] + bz) # Update gate
rt = sigmoid(Wr * [ht
-
1, xt] + br) # Reset gate
ht_candidate = tanh(Wh * [rt * ht
-
1, xt] + bh) # Candidate hidden state
ht = (1
-
zt) * ht
-
1 + zt * ht_candidate # Update hidden state
END FOR
h_final = ht
ELSE IF model_type == "BiLSTM":
# Forward LSTM
INITIALIZE h0_f, C0_f # Initialize forward hidden and cell states
FOR t = 1 to T:
Compute ht_f using LSTM equations for forward direction
END FOR
# Backward LSTM
INITIALIZE h0_b, C0_b # Initialize backward hidden and cell states
FOR t = T to 1:
Compute ht_b using LSTM equations for backward direction
END FOR
# Combine Forward and Backward Outputs
FOR t = 1 to T:
h_combined_t = concatenate(ht_f, ht_b) # Combine forward and backward states
END FOR
h_final = h_combined_t
RETURN h_final
3
.
3
.
Rec
urre
nt
neura
l net
wo
rk
s
R
NNs
ar
e
d
esig
n
ed
t
o
p
r
o
c
ess
s
eq
u
en
tial
d
ata
b
y
u
s
in
g
a
h
id
d
e
n
s
tate
to
r
etain
i
n
f
o
r
m
atio
n
f
r
o
m
p
r
ev
i
o
u
s
s
tep
s
.
T
h
ese
m
o
d
els
o
f
ten
f
ac
e
ch
allen
g
e
s
s
u
ch
as
v
an
is
h
in
g
g
r
ad
ien
ts
an
d
d
if
f
icu
lties
with
lo
n
g
-
ter
m
d
ep
e
n
d
en
cie
s
.
T
o
r
eso
lv
e
th
ese
is
s
u
es,
v
ar
ian
ts
lik
e
L
STM
,
G
R
U
,
an
d
B
iLST
M
ar
e
co
m
m
o
n
l
y
u
tili
ze
d
.
T
h
e
p
r
o
p
o
s
ed
in
tr
u
s
io
n
d
ete
ctio
n
f
r
am
ewo
r
k
f
ir
s
t
p
r
e
p
r
o
ce
s
s
es
th
e
NSL
-
KDD
d
atase
t
th
r
o
u
g
h
n
o
r
m
aliza
tio
n
an
d
lab
el
en
co
d
in
g
.
SMOT
E
is
th
en
ap
p
lied
ex
clu
s
iv
ely
to
th
e
tr
ain
in
g
d
ataset
to
b
alan
ce
m
in
o
r
ity
attac
k
class
es with
o
u
t in
tr
o
d
u
cin
g
d
ata
leak
ag
e.
T
h
e
b
alan
ce
d
d
ata
ar
e
s
u
b
s
eq
u
e
n
tly
u
s
ed
f
o
r
tr
ain
in
g
th
e
s
elec
ted
R
NN
ar
ch
itectu
r
e
(
L
STM
,
GR
U,
o
r
B
iLST
M)
.
Hy
p
er
p
ar
a
m
eter
o
p
tim
izatio
n
is
p
er
f
o
r
m
ed
u
s
in
g
R
an
d
o
m
ized
Sear
ch
C
V,
wh
ile
ea
r
ly
s
to
p
p
in
g
p
r
e
v
en
ts
o
v
e
r
f
itti
n
g
.
Fin
ally
,
th
e
tr
ain
e
d
m
o
d
el
is
ev
alu
ate
d
u
s
in
g
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e.
3
.
4
.
E
a
rly
s
t
o
pp
ing
E
ar
ly
s
to
p
p
i
n
g
p
r
ev
en
ts
o
v
er
f
i
ttin
g
b
y
h
altin
g
tr
ain
in
g
wh
en
th
e
v
alid
atio
n
lo
s
s
s
to
p
s
im
p
r
o
v
in
g
f
o
r
a
s
p
ec
if
ied
n
u
m
b
er
o
f
ep
o
c
h
s
(
)
:
(
ℎ
)
ℎ
]
T
h
is
en
s
u
r
es
th
e
m
o
d
el
is
tr
ain
ed
u
p
t
o
th
e
o
p
tim
al
p
o
i
n
t
o
f
g
e
n
er
aliza
tio
n
wh
ile
s
a
v
in
g
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
.
T
h
e
NSL
-
KDD
d
ataset
was
d
iv
id
ed
in
to
7
0
%
tr
ain
in
g
,
1
5
%
v
alid
atio
n
,
an
d
1
5
%
test
in
g
s
ets.
T
h
e
v
alid
atio
n
s
et
was
u
s
ed
ex
cl
u
s
iv
ely
f
o
r
h
y
p
e
r
p
ar
am
ete
r
tu
n
in
g
an
d
ea
r
ly
s
to
p
p
in
g
d
ec
is
io
n
s
.
Hy
p
er
p
a
r
am
eter
tu
n
in
g
was
p
er
f
o
r
m
e
d
u
s
in
g
R
an
d
o
m
ized
Sear
ch
C
V
with
5
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
o
n
th
e
tr
ain
in
g
d
ata.
T
h
e
ex
p
lo
r
ed
h
y
p
er
p
ar
am
eter
r
a
n
g
es
ar
e
s
u
m
m
ar
ized
in
T
ab
le
1
.
E
ar
ly
s
to
p
p
in
g
with
a
p
atien
ce
o
f
1
0
ep
o
ch
s
was
em
p
lo
y
ed
to
p
r
ev
e
n
t
o
v
e
r
f
itti
n
g
.
T
r
ai
n
in
g
was
c
o
n
d
u
cted
o
f
f
lin
e,
wh
ile
in
f
e
r
en
ce
wa
s
ev
alu
ated
o
n
th
e
h
eld
-
o
u
t te
s
t set.
T
ab
le
1
.
Hy
p
er
p
ar
a
m
eter
s
ea
r
ch
s
p
ac
e
P
a
r
a
me
t
e
r
V
a
l
u
e
s
H
i
d
d
e
n
u
n
i
t
s
{6
4
,
1
2
8
,
2
5
6
}
Le
a
r
n
i
n
g
r
a
t
e
{0
.
0
0
1
,
0
.
0
0
0
5
,
0
.
0
0
0
1
}
B
a
t
c
h
si
z
e
{3
2
,
6
4
,
1
2
8
}
D
r
o
p
o
u
t
r
a
t
e
{0
.
2
,
0
.
3
,
0
.
5
}
O
p
t
i
mi
z
e
r
A
d
a
m
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
4
,
Au
g
u
s
t 2
0
2
6
:
3
7
0
3
-
3
7
1
1
3708
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
tech
n
iq
u
e
was
test
ed
o
n
th
e
NSL
-
KDD
d
ataset
f
o
r
in
tr
u
s
io
n
d
etec
tio
n
ex
p
er
im
en
ts
,
an
d
th
e
m
etr
ics
lik
e
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
,
an
d
F1
-
s
co
r
e
wer
e
ca
lcu
lated
to
ev
alu
ate
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
tech
n
iq
u
e.
Fo
r
r
ef
er
en
ce
,
T
ab
le
2
d
escr
ib
es
v
ar
io
u
s
m
o
d
els
an
d
m
eth
o
d
s
in
th
e
co
n
tex
t
o
f
I
DS
b
ased
o
n
th
e
f
o
u
r
p
er
f
o
r
m
a
n
ce
p
ar
am
e
ter
s
:
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e.
T
h
ese
m
etr
ics
ar
e
es
s
en
tial
to
ev
alu
ate
th
e
p
o
ten
tial
o
f
m
o
d
els
in
d
etec
tin
g
tr
u
e
p
o
s
i
tiv
e
ca
s
es
wh
ile
m
in
im
izin
g
th
e
g
en
er
atio
n
o
f
f
alse p
o
s
itiv
es
[
2
4
]
,
[
2
5
]
.
T
ab
le
2
.
Per
f
o
r
m
an
ce
co
m
p
a
r
is
o
n
o
f
in
tr
u
s
io
n
d
etec
tio
n
m
o
d
els b
ased
o
n
k
ey
m
etr
ics
M
o
d
e
l
/
met
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
(
%)
P
r
o
p
o
se
d
B
i
LS
TM
+
S
M
O
TE
9
9
.
1
2
9
8
.
9
5
9
9
.
0
5
9
8
.
9
6
Tr
a
d
i
t
i
o
n
a
l
LS
TM
-
b
a
s
e
d
I
D
S
9
7
.
8
0
9
7
.
6
0
9
7
.
7
0
9
7
.
6
5
G
R
U
-
b
a
s
e
d
mo
d
e
l
9
6
.
5
0
9
6
.
4
0
9
6
.
3
0
9
6
.
3
5
RF
(
b
a
se
l
i
n
e
)
9
2
.
0
0
9
1
.
8
0
9
1
.
9
0
9
1
.
8
5
N
a
ï
v
e
B
a
y
e
s
I
D
S
8
5
.
0
0
8
3
.
7
0
8
4
.
5
0
8
4
.
1
0
T
h
e
p
r
o
p
o
s
ed
B
iLST
M
+
SMOT
E
m
o
d
el
i
n
th
is
wo
r
k
s
h
o
ws
o
u
ts
tan
d
in
g
ac
c
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
an
d
F1
-
s
co
r
e
m
ea
s
u
r
e,
wh
ic
h
ar
e
9
9
.
1
2
%,
9
8
.
9
5
%,
9
9
.
0
5
%,
an
d
9
8
.
9
6
%
,
r
esp
ec
tiv
ely
.
T
h
e
r
esu
lts
u
n
d
er
lin
e
th
e
r
o
b
u
s
tn
ess
an
d
e
f
f
ec
tiv
e
n
ess
o
f
th
e
m
o
d
el
to
ac
cu
r
ately
d
etec
t
in
tr
u
s
io
n
s
ev
en
in
th
e
p
r
esen
ce
o
f
h
ig
h
-
class
im
b
alan
ce
s
.
Su
c
h
co
m
b
i
n
atio
n
as
B
iLST
M
with
SMOT
E
ca
n
g
iv
e
DL
p
o
wer
f
u
l
s
eq
u
e
n
ce
m
o
d
elin
g
f
ea
t
u
r
es,
an
d
it
also
ca
n
b
e
a
g
o
o
d
s
o
lu
tio
n
f
o
r
th
e
"a
ttack
"
s
am
p
les
b
ein
g
less
in
in
tr
u
s
io
n
d
etec
tio
n
d
atasets
.
T
h
e
v
er
y
p
r
ec
is
e
r
es
u
lt
s
h
o
ws
th
e
m
o
d
el
is
r
ea
lly
g
o
o
d
at
d
ec
r
ea
s
in
g
f
alse
p
o
s
itiv
es,
an
d
t
h
e
h
i
g
h
r
ec
all
m
ea
n
s
it
is
ca
p
ab
le
o
f
d
etec
tin
g
tr
u
e
in
tr
u
s
io
n
ca
s
es
q
u
ite
well.
T
h
e
h
ar
m
o
n
ized
F1
-
s
co
r
e
u
n
d
er
li
n
es
th
e
all
-
ar
o
u
n
d
ac
cu
r
ac
y
an
d
r
o
b
u
s
tn
ess
o
f
th
e
m
o
d
el.
T
h
e
tr
ad
itio
n
al
L
STM
-
b
ased
I
DS
is
a
s
tan
d
ar
d
DL
-
b
ased
I
DS
wh
ich
h
as
an
ac
cu
r
ac
y
o
f
9
7
.
8
0
%,
a
p
r
ec
is
io
n
o
f
9
7
.
6
0
%,
a
r
ec
all
o
f
9
7
.
7
0
%,
an
d
a
n
F1
-
s
co
r
e
o
f
9
7
.
6
5
%.
T
h
is
m
et
h
o
d
wo
r
k
s
well
o
n
all
m
etr
ics,
b
u
t
n
o
t
as
g
o
o
d
as
th
e
p
r
o
p
o
s
ed
B
iLST
M
+
SMOT
E
m
o
d
el.
Su
ch
a
d
i
f
f
er
en
ce
is
d
u
e
t
o
th
e
f
ac
t
t
h
at
th
e
tr
ad
itio
n
al
L
STM
m
o
d
el
d
o
es
n
o
t
em
p
lo
y
ad
v
an
ce
d
d
a
ta
-
b
alan
cin
g
tech
n
iq
u
es,
wh
i
ch
r
esu
lts
in
less
ef
f
ec
tiv
en
ess
in
d
ata
im
b
alan
ce
d
d
atasets
.
Mo
r
e
o
v
er
,
th
e
b
id
ir
ec
tio
n
al
m
ec
h
an
is
m
o
f
th
e
B
iLST
M
s
tr
u
ctu
r
e
allo
ws th
e
p
r
o
p
o
s
ed
m
o
d
el
to
in
clu
d
e
p
ast an
d
f
u
tu
r
e
tim
e
m
em
o
r
y
,
wh
ich
f
u
r
th
er
im
p
r
o
v
e
s
th
e
g
en
er
aliza
tio
n
ca
p
ab
ilit
ies
o
f
th
e
m
o
d
el
wh
en
co
m
p
ar
ed
to
th
e
u
n
i
d
ir
ec
tio
n
al
L
STM
.
T
h
e
GR
U
-
b
ased
m
o
d
el
ac
h
iev
es
an
ac
cu
r
ac
y
o
f
9
6
.
5
0
%,
a
p
r
ec
is
io
n
o
f
9
6
.
4
0
%,
a
r
ec
all
o
f
9
6
.
3
0
%,
an
d
an
F1
-
s
co
r
e
o
f
9
6
.
3
5
%.
GR
U
h
as
b
ee
n
k
n
o
wn
f
o
r
its
co
m
p
u
tatio
n
al
ef
f
icien
cy
an
d
its
s
im
p
le
ar
ch
itectu
r
e
wh
en
co
m
p
ar
e
d
to
L
STM
,
b
u
t
it
d
o
esn
'
t
ca
p
tu
r
e
lo
n
g
ter
m
d
ep
en
d
en
ce
in
s
eq
u
en
tial
d
ata
as
well.
T
h
is
is
ev
id
en
t
f
r
o
m
th
e
lo
wer
m
etr
ics
o
f
th
e
GR
U
m
o
d
el
wh
e
n
co
m
p
ar
ed
with
L
STM
-
b
ased
m
o
d
els.
Ho
we
v
er
,
it
is
s
till
a
co
m
p
etitiv
e
ap
p
r
o
ac
h
,
esp
ec
ially
in
ca
s
es with
lim
ited
co
m
p
u
tatio
n
al
ca
p
ac
ity
.
T
h
e
b
aselin
e
m
o
d
el
in
RF
h
as
an
ac
cu
r
ac
y
o
f
9
2
.
0
0
%,
a
p
r
ec
is
io
n
o
f
9
1
.
8
0
%,
a
r
ec
all
o
f
9
1
.
9
0
%,
an
d
an
F1
-
s
co
r
e
o
f
9
1
.
8
5
%.
T
h
is
m
o
d
el
is
a
ty
p
ical
ML
m
o
d
el
th
at
h
as
b
ee
n
a
p
p
lied
in
I
DS
b
ec
au
s
e
o
f
its
en
s
em
b
le
lear
n
in
g
n
atu
r
e
an
d
r
esis
tan
ce
to
o
v
er
f
itti
n
g
.
B
u
t
it
h
as
b
ee
n
f
o
u
n
d
to
b
e
less
ef
f
ec
tiv
e
th
an
DL
m
o
d
els,
wh
ich
in
d
icate
s
th
at
it
m
ay
b
e
less
s
u
itab
le
f
o
r
co
m
p
le
x
an
d
h
ig
h
-
d
im
en
s
io
n
al
n
etwo
r
k
d
at
a
p
r
o
ce
s
s
in
g
.
RF
h
as
d
if
f
icu
lties
d
is
ce
r
n
in
g
tem
p
o
r
al
a
n
d
s
eq
u
en
tial
p
atter
n
s
in
d
ata
th
at
ar
e
im
p
o
r
tan
t
f
o
r
th
e
ac
cu
r
ate
d
etec
tio
n
o
f
d
y
n
a
m
ic
an
d
ev
o
lv
in
g
in
tr
u
s
io
n
p
att
er
n
s
.
T
h
e
n
aïv
e
B
ay
es
I
DS
m
o
d
el
h
as
th
e
lo
west
ac
cu
r
ac
y
o
f
8
5
.
0
0
%,
th
e
lo
we
s
t
p
r
ec
is
io
n
o
f
8
3
.
7
0
%,
th
e
lo
west
r
ec
all
o
f
8
4
.
5
0
%
,
an
d
th
e
lo
west
F1
-
s
co
r
e
o
f
8
4
.
1
0
%
in
t
h
e
co
m
p
ar
is
o
n
.
T
h
e
n
aïv
e
B
ay
es
class
if
ier
'
s
s
i
m
p
licity
in
m
ak
in
g
its
ass
u
m
p
tio
n
s
,
s
u
ch
as
f
ea
tu
r
e
in
d
ep
en
d
en
ce
,
m
ak
es
it
less
ef
f
ec
tiv
e
in
ca
p
tu
r
in
g
th
e
co
m
p
lex
r
elatio
n
s
h
ip
s
b
etwe
en
f
ea
tu
r
es
in
n
etwo
r
k
tr
af
f
ic
d
ata.
T
h
is
lead
s
to
a
h
ig
h
f
alse
p
o
s
itiv
e
an
d
f
alse
n
eg
ativ
e
r
ate
an
d
is
th
er
ef
o
r
e
l
ess
ap
p
r
o
p
r
iate
f
o
r
to
d
ay
'
s
in
tr
u
s
io
n
d
etec
tio
n
a
p
p
licatio
n
s
wh
ich
r
e
q
u
ir
e
h
i
g
h
p
r
ec
is
io
n
a
n
d
r
ec
all.
T
h
e
tab
le
h
ig
h
lig
h
ts
th
e
im
p
o
r
tan
t
b
en
ef
it
o
f
DL
m
eth
o
d
s
co
m
p
ar
e
d
to
tr
ad
itio
n
al
ML
m
o
d
els
w
h
en
u
s
in
g
d
ata
-
b
alan
cin
g
m
eth
o
d
s
s
u
ch
as
SMOT
E
to
im
p
r
o
v
e
p
er
f
o
r
m
a
n
ce
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
B
iLST
M
+
SMOT
E
is
th
e
b
est
in
co
m
p
ar
is
o
n
,
with
its
ab
il
ity
to
ef
f
ec
tiv
ely
ca
p
tu
r
e
th
e
tem
p
o
r
al
d
ep
e
n
d
en
cy
,
o
v
er
co
m
e
class
im
b
alan
ce
,
an
d
ac
h
iev
e
s
ig
n
if
ican
t
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e
.
T
h
e
im
p
r
o
v
em
e
n
t
o
v
er
t
h
e
co
n
v
e
n
tio
n
al
L
STM
an
d
GR
U
m
o
d
els
u
n
d
er
s
co
r
es
th
e
s
ig
n
if
ican
ce
o
f
a
r
ch
itectu
r
al
en
h
an
ce
m
en
ts
an
d
p
r
e
p
r
o
ce
s
s
in
g
m
eth
o
d
s
in
in
tr
u
s
io
n
d
etec
tio
n
.
i)
Scalab
ilit
y
an
d
r
ea
l
-
tim
e
f
ea
s
ib
ilit
y
:
wh
ile
tr
ain
in
g
is
m
o
r
e
ex
p
e
n
s
iv
e
f
o
r
B
iLST
M
m
o
d
els,
in
f
er
en
ce
laten
cy
is
s
till
ac
ce
p
tab
le,
f
itti
n
g
th
e
p
ar
ad
ig
m
to
b
e
d
e
p
lo
y
ed
at
th
e
SDN
co
n
tr
o
ller
le
v
el
in
n
ea
r
-
r
ea
l
tim
e.
T
r
ain
in
g
is
d
o
n
e
o
f
f
lin
e
,
an
d
r
ea
l
-
tim
e
in
f
er
en
ce
ca
n
b
e
ad
d
e
d
to
ce
n
tr
alize
d
SDN
m
o
n
ito
r
i
n
g
s
y
s
tem
s
.
C
o
m
p
r
ess
io
n
o
f
th
e
m
o
d
el
o
r
in
f
er
e
n
ce
u
s
in
g
e
d
g
e
s
co
u
ld
b
e
a
n
ex
ten
s
io
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dee
p
lea
r
n
in
g
in
tr
u
s
io
n
d
etec
tio
n
fo
r
s
o
ftw
a
r
e
-
d
efin
ed
…
(
P
r
a
jw
a
la
s
imh
a
S
in
d
u
g
a
tta
N
a
g
a
r
a
ja
)
3709
ii)
Data
s
et
lim
itat
io
n
s
an
d
co
n
ce
p
t
d
r
if
t:
NSL
-
KDD
allo
ws
f
o
r
a
f
air
co
m
p
ar
is
o
n
with
p
r
ev
io
u
s
wo
r
k
,
b
u
t
it
d
o
esn
'
t
ac
cu
r
ately
r
ep
r
esen
t
t
h
e
b
eh
av
io
r
o
f
to
d
a
y
'
s
n
etwo
r
k
s
.
I
n
r
ea
l
-
wo
r
ld
tr
af
f
ic,
co
n
ce
p
t
d
r
if
t
o
cc
u
r
s
f
r
o
m
ch
a
n
g
es in
attac
k
s
tr
ateg
i
es a
n
d
u
s
er
'
s
b
eh
av
io
r
.
T
h
is
m
u
s
t b
e
s
u
s
tain
ed
th
r
o
u
g
h
p
er
i
o
d
ic
r
etr
ain
in
g
,
on
-
lin
e
ed
u
ca
tio
n
o
r
b
y
d
r
i
f
t d
etec
tio
n
.
5.
CO
NCLU
SI
O
N
T
h
e
p
r
o
p
o
s
ed
I
DS
b
ased
o
n
B
iLST
M
an
d
SMOT
E
s
h
o
we
d
ex
ce
llen
t
r
esu
lts
in
d
etec
tin
g
n
etwo
r
k
in
tr
u
s
io
n
s
in
th
e
NSL
-
KDD
d
ataset,
wh
ich
is
a
class
-
im
b
alan
ce
d
d
ataset.
T
h
e
m
eth
o
d
h
as
an
ac
cu
r
ac
y
o
f
9
9
.
1
2
%,
a
p
r
ec
is
io
n
o
f
9
8
.
9
5
%,
a
r
ec
all
o
f
9
9
.
0
5
%,
an
d
an
F1
-
s
co
r
e
o
f
9
8
.
9
6
%
wh
ich
o
u
t
p
er
f
o
r
m
s
tr
ad
itio
n
al
m
eth
o
d
s
s
u
ch
as
RF
an
d
GR
U
-
b
ased
m
eth
o
d
s
.
B
iLST
M
s
h
av
e
s
u
b
s
tan
tially
im
p
r
o
v
ed
t
h
e
m
o
d
el'
s
ab
ilit
y
to
lear
n
co
m
p
lex
tem
p
o
r
al
d
y
n
a
m
ics
in
n
etwo
r
k
tr
a
f
f
ic,
wh
i
ch
is
a
k
ey
elem
e
n
t
in
th
e
c
o
n
tex
t
o
f
in
tr
u
s
io
n
d
etec
tio
n
in
s
eq
u
en
tial
d
ata.
T
h
e
u
s
e
o
f
SMOT
E
p
lay
ed
a
k
e
y
r
o
le
to
r
eso
lv
e
t
h
e
im
b
alan
c
e
o
f
attac
k
class
es,
m
o
r
e
r
ep
r
esen
tativ
e
o
f
m
in
o
r
i
ty
class
es
wh
en
tr
ain
in
g
,
an
d
p
r
ev
en
ted
f
alse
n
eg
ativ
e
er
r
o
r
s
f
r
o
m
o
cc
u
r
r
in
g
.
T
h
e
m
o
d
el
is
o
u
tp
er
f
o
r
m
i
n
g
,
b
u
t
r
eq
u
ir
es
a
lo
t
o
f
co
m
p
u
ta
tio
n
al
r
eso
u
r
ce
s
to
b
e
d
e
p
lo
y
ed
in
th
e
r
eso
u
r
c
e
co
n
s
tr
ain
ed
e
n
v
ir
o
n
m
en
ts
.
A
d
d
itio
n
ally
,
alth
o
u
g
h
it
h
as
b
ee
n
r
ig
o
r
o
u
s
ly
test
ed
u
s
in
g
th
e
NSL
-
KDD
d
ataset,
th
er
e
is
a
n
ee
d
to
f
u
r
t
h
er
test
i
ts
ad
ap
tab
ilit
y
with
r
eg
ar
d
s
to
m
o
d
e
r
n
cy
b
e
r
th
r
ea
ts
o
n
co
n
t
em
p
o
r
ar
y
d
atasets
.
T
o
s
u
m
m
ar
ize,
th
is
I
DS
ap
p
e
ar
s
to
b
e
o
n
e
o
f
th
e
m
ajo
r
s
o
lu
tio
n
s
in
n
etwo
r
k
s
ec
u
r
ity
,
e
s
p
ec
ially
f
o
r
SDNs
an
d
th
e
latest
DL
tech
n
o
lo
g
ie
s
an
d
p
o
wer
f
u
l
p
r
e
-
p
r
o
ce
s
s
in
g
ca
p
ab
ilit
ies.
Fu
tu
r
e
d
ir
ec
tio
n
s
will
in
v
o
lv
e
th
e
r
ea
l
-
tim
e
d
ep
lo
y
m
en
t,
f
u
r
th
e
r
test
in
g
with
a
lar
g
er
d
ataset,
an
d
co
m
b
in
in
g
th
e
h
y
b
r
id
m
o
d
els
with
atten
tio
n
m
ec
h
an
is
m
s
f
o
r
en
h
an
ce
d
ad
ap
tab
ilit
y
an
d
d
etec
tio
n
ca
p
a
b
ilit
ies.
T
h
e
f
in
d
in
g
s
o
f
t
h
is
r
esear
ch
s
h
o
w
th
at
B
iLST
M
m
o
d
els ca
n
b
e
a
v
ita
l c
o
m
p
o
n
en
t o
f
th
e
f
u
tu
r
e
o
f
c
y
b
er
s
ec
u
r
ity
f
r
am
ewo
r
k
s
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
We
h
av
e
not
r
ec
eiv
e
d
f
u
n
d
in
g
f
o
r
th
is
wo
r
k
f
r
o
m
an
y
g
o
v
er
n
m
en
t
an
d
non
-
g
o
v
er
n
m
e
n
t
o
r
g
a
n
i
z
a
t
i
o
n
s
.
T
h
is
is
a
s
elf
-
f
u
n
d
ed
r
esear
ch
w
o
r
k
.
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
Pra
jwalasim
h
a
Sin
d
u
g
atta
Nag
ar
aja
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Nav
y
a
R
ajash
ek
ar
a
✓
✓
✓
✓
✓
Pu
s
h
p
a
B
an
g
alo
r
e
R
am
esh
✓
✓
✓
✓
✓
Dr
u
v
a
Ku
m
ar
Sid
d
ar
aju
✓
✓
✓
✓
✓
San
th
o
s
h
Ku
m
ar
R
am
ac
h
an
d
r
ag
o
wd
a
✓
✓
✓
✓
✓
T
r
u
p
ti Sh
r
ip
a
d
T
ag
ar
e
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
I
NF
O
RM
E
D
CO
NS
E
N
T
W
e
h
av
e
o
b
tain
ed
in
f
o
r
m
ed
c
o
n
s
en
t f
r
o
m
all
in
d
iv
id
u
als in
c
lu
d
ed
in
t
h
is
s
tu
d
y
.
I
NF
O
RM
E
D
CO
NS
E
N
T
T
h
is
s
ec
tio
n
n
o
t
ap
p
licab
le,
as
th
is
s
tu
d
y
d
id
n
o
t
in
v
o
lv
e
h
u
m
an
p
ar
ticip
a
n
ts
o
r
p
er
s
o
n
ally
id
en
tifia
b
le
in
f
o
r
m
atio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
4
,
Au
g
u
s
t 2
0
2
6
:
3
7
0
3
-
3
7
1
1
3710
E
T
H
I
CAL AP
P
RO
V
AL
T
h
is
s
ec
tio
n
n
o
t
ap
p
licab
le,
a
s
th
is
s
tu
d
y
d
id
n
o
t
in
v
o
lv
e
h
u
m
an
o
r
an
im
al
s
u
b
jects
an
d
u
s
ed
o
n
ly
non
-
s
en
s
itiv
e
co
m
p
u
tatio
n
al
o
r
p
u
b
lic
d
ata.
DATA AV
AI
L
AB
I
L
I
T
Y
Data
av
ailab
ilit
y
is
not
ap
p
licab
le
to
th
is
p
ap
e
r
as
no
n
ew
d
at
a
wer
e
cr
ea
ted
or
a
n
aly
ze
d
in
t
h
is
s
t
u
d
y
.
RE
F
E
R
E
NC
E
S
[
1
]
V
.
H
n
a
m
t
e
a
n
d
J.
H
u
ssa
i
n
,
“
D
C
N
N
B
i
LSTM
:
a
n
e
f
f
i
c
i
e
n
t
h
y
b
r
i
d
d
e
e
p
l
e
a
r
n
i
n
g
-
b
a
se
d
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
s
y
st
e
m,
”
T
e
l
e
m
a
t
i
c
s
a
n
d
I
n
f
o
rm
a
t
i
c
s R
e
p
o
rt
s
,
v
o
l
.
1
0
,
Ju
n
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
t
e
l
e
r
.
2
0
2
3
.
1
0
0
0
5
3
.
[
2
]
M
.
C
u
i
,
J.
C
h
e
n
,
X
.
Q
i
u
,
W
.
L
v
,
H
.
Q
i
n
,
a
n
d
X
.
Z
h
a
n
g
,
“
M
u
l
t
i
-
c
l
a
ss
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
sy
s
t
e
m i
n
S
D
N
b
a
se
d
o
n
h
y
b
r
i
d
B
i
LST
M
mo
d
e
l
,
”
C
l
u
s
t
e
r
C
o
m
p
u
t
i
n
g
,
v
o
l
.
2
7
,
n
o
.
7
,
p
p
.
9
9
3
7
–
9
9
5
6
,
O
c
t
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
5
8
6
-
0
2
4
-
0
4
4
7
7
-
5.
[
3
]
S
.
B
i
,
J
.
W
a
n
g
,
J
.
S
o
n
g
,
P
.
L
i
,
a
n
d
L.
Li
,
“
R
e
se
a
r
c
h
o
n
t
h
e
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
mo
d
e
l
f
o
r
p
o
w
e
r
i
n
t
e
r
n
e
t
o
f
t
h
i
n
g
s
c
o
mb
i
n
i
n
g
d
e
e
p
b
e
l
i
e
f
n
e
t
w
o
r
k
a
n
d
B
i
LST
M
,
”
J
o
u
rn
a
l
o
f
C
y
b
e
r
S
e
c
u
r
i
t
y
a
n
d
M
o
b
i
l
i
t
y
,
v
o
l
.
1
4
,
n
o
.
3
,
p
p
.
6
5
3
–
6
7
2
,
A
u
g
.
2
0
2
5
,
d
o
i
:
1
0
.
1
3
0
5
2
/
j
c
sm
2
2
4
5
-
1
4
3
9
.
1
4
3
6
.
[
4
]
J.
G
a
o
,
“
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
me
t
h
o
d
c
o
m
b
i
n
i
n
g
C
N
N
a
n
d
B
i
LS
TM
i
n
c
l
o
u
d
c
o
mp
u
t
i
n
g
e
n
v
i
r
o
n
me
n
t
,
”
C
o
m
p
u
t
a
t
i
o
n
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
N
e
u
ro
s
c
i
e
n
c
e
,
v
o
l
.
2
0
2
2
,
p
p
.
1
–
1
1
,
A
p
r
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
5
5
/
2
0
2
2
/
7
2
7
2
4
7
9
.
[
5
]
L.
Z
h
a
n
g
,
J
.
H
u
a
n
g
,
Y
.
Z
h
a
n
g
,
a
n
d
G
.
Z
h
a
n
g
,
“
I
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
m
o
d
e
l
o
f
C
N
N
-
B
i
LS
TM
a
l
g
o
r
i
t
h
m
b
a
se
d
o
n
mea
n
c
o
n
t
r
o
l
,
”
i
n
2
0
2
0
I
EE
E
1
1
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
S
o
f
t
w
a
re
E
n
g
i
n
e
e
r
i
n
g
a
n
d
S
e
r
v
i
c
e
S
c
i
e
n
c
e
,
O
c
t
.
2
0
2
0
,
p
p
.
2
2
–
27
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
S
ESS
4
9
9
3
8
.
2
0
2
0
.
9
2
3
7
6
5
6
.
[
6
]
A
.
A
l
q
a
h
t
a
n
i
,
“
O
p
t
i
m
i
z
e
d
d
e
e
p
a
u
t
o
e
n
c
o
d
e
r
a
n
d
B
i
LS
TM
f
o
r
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
i
n
I
o
Ts
-
F
o
g
c
o
m
p
u
t
i
n
g
,
”
Mu
l
t
i
m
e
d
i
a
T
o
o
l
s
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
8
4
,
n
o
.
8
,
p
p
.
4
9
0
7
–
4
9
4
3
,
M
a
r
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
1
0
4
2
-
0
2
4
-
1
8
9
1
9
-
0.
[
7
]
H
.
B
e
n
a
h
me
d
e
t
a
l
.
,
“
H
B
i
LD
-
I
D
S
:
a
n
e
f
f
i
c
i
e
n
t
h
y
b
r
i
d
B
i
LST
M
-
D
N
N
m
o
d
e
l
f
o
r
r
e
a
l
-
t
i
me
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
i
n
I
o
M
T
n
e
t
w
o
r
k
s,”
I
n
f
o
rm
a
t
i
o
n
,
v
o
l
.
1
6
,
n
o
.
8
,
A
u
g
.
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
i
n
f
o
1
6
0
8
0
6
6
9
.
[
8
]
A
.
M
.
A
l
h
a
ss
a
n
,
“
S
e
l
f
-
a
d
a
p
t
i
v
e
l
i
g
h
t
w
e
i
g
h
t
a
t
t
e
n
t
i
o
n
mo
d
u
l
e
-
b
a
se
d
B
i
LS
TM
m
o
d
e
l
f
o
r
e
f
f
e
c
t
i
v
e
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
,
”
Ar
a
b
i
a
n
J
o
u
rn
a
l
f
o
r
S
c
i
e
n
c
e
a
n
d
En
g
i
n
e
e
r
i
n
g
,
v
o
l
.
5
0
,
n
o
.
1
5
,
p
p
.
1
1
5
1
3
–
1
1
5
3
8
,
A
u
g
.
2
0
2
5
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
3
3
6
9
-
0
2
4
-
0
9
4
7
6
-
7.
[
9
]
M
.
A
d
e
l
,
A
.
M
a
h
e
r
,
A
.
M
o
h
a
m
me
d
,
a
n
d
M
.
A
.
El
a
z
e
e
m,
“
R
o
b
u
s
t
B
i
LSTM
-
b
a
s
e
d
m
u
l
t
i
-
c
l
a
ss
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
f
o
r
I
o
T
n
e
t
w
o
r
k
s
u
si
n
g
T
o
N
-
I
o
T
d
a
t
a
se
t
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
T
e
l
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
,
v
o
l
.
5
,
n
o
.
2
,
p
p
.
1
–
1
4
,
N
o
v
.
2
0
2
5
,
d
o
i
:
1
0
.
2
1
6
0
8
/
i
j
t
.
2
0
2
5
.
4
2
6
7
0
8
.
1
1
3
4
.
[
1
0
]
B
.
O
m
a
r
o
v
,
Z.
S
a
i
l
a
u
k
y
z
y
,
A
.
B
i
g
a
l
i
y
e
v
a
,
A
.
K
e
r
e
y
e
v
,
L.
N
a
i
z
a
b
a
y
e
v
a
,
a
n
d
A
.
D
a
u
t
b
a
y
e
v
a
,
“
O
n
e
d
i
m
e
n
s
i
o
n
a
l
C
o
n
v
-
B
i
LST
M
n
e
t
w
o
r
k
w
i
t
h
a
t
t
e
n
t
i
o
n
me
c
h
a
n
i
sm
f
o
r
I
o
T
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
,
”
C
o
m
p
u
t
e
rs
,
Ma
t
e
r
i
a
l
s
& C
o
n
t
i
n
u
a
,
v
o
l
.
7
7
,
n
o
.
3
,
p
p
.
3
7
6
5
–
3
7
8
1
,
2
0
2
3
,
d
o
i
:
1
0
.
3
2
6
0
4
/
c
mc
.
2
0
2
3
.
0
4
2
4
6
9
.
[
1
1
]
C
.
Zh
a
n
g
,
J.
L
i
,
N
.
W
a
n
g
,
a
n
d
D
.
Z
h
a
n
g
,
“
R
e
se
a
r
c
h
o
n
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
m
e
t
h
o
d
b
a
se
d
o
n
t
r
a
n
sf
o
r
m
e
r
a
n
d
C
N
N
-
B
i
LS
TM
i
n
i
n
t
e
r
n
e
t
o
f
t
h
i
n
g
s
,
”
S
e
n
s
o
rs
,
v
o
l
.
2
5
,
n
o
.
9
,
A
p
r
.
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
5
0
9
2
7
2
5
.
[
1
2
]
M
.
J
o
u
h
a
r
i
a
n
d
M
.
G
u
i
z
a
n
i
,
“
L
i
g
h
t
w
e
i
g
h
t
C
N
N
-
B
i
LST
M
b
a
s
e
d
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
s
y
s
t
e
ms f
o
r
r
e
s
o
u
r
c
e
-
c
o
n
st
r
a
i
n
e
d
I
o
T
d
e
v
i
c
e
s
,
”
i
n
2
0
2
4
I
n
t
e
rn
a
t
i
o
n
a
l
Wi
r
e
l
e
s
s
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
,
M
a
y
2
0
2
4
,
p
p
.
1
5
5
8
–
1
5
6
3
,
d
o
i
:
1
0
.
1
1
0
9
/
I
W
C
M
C
6
1
5
1
4
.
2
0
2
4
.
1
0
5
9
2
3
5
2
.
[
1
3
]
G
.
S
.
v
i
d
h
y
a
a
n
d
R
.
N
a
g
a
r
a
j
a
n
,
“
A
n
o
v
e
l
b
i
d
i
r
e
c
t
i
o
n
a
l
LST
M
m
o
d
e
l
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
i
n
S
D
N
-
I
o
T
n
e
t
w
o
r
k
,
”
C
o
m
p
u
t
i
n
g
,
v
o
l
.
1
0
6
,
n
o
.
8
,
p
p
.
2
6
1
3
–
2
6
4
2
,
A
u
g
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
6
0
7
-
0
2
4
-
0
1
2
9
5
-
w.
[
1
4
]
I
.
M
.
S
e
l
i
m
a
n
d
R
.
A
.
S
a
d
e
k
,
“
D
A
E
-
B
i
LST
M
:
a
F
o
g
-
b
a
se
d
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
mo
d
e
l
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
I
o
T,
”
J
o
u
r
n
a
l
o
f
T
h
e
o
re
t
i
c
a
l
a
n
d
Ap
p
l
i
e
d
I
n
f
o
rm
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
1
0
1
,
n
o
.
5
,
2
0
2
3
.
[
1
5
]
T.
L
.
Y
a
sara
t
h
n
a
a
n
d
N
.
-
A
.
L
.
-
K
h
a
c
,
“
A
S
EA
D
O
S
-
S
D
N
-
I
o
T:
a
n
o
v
e
l
S
D
N
-
I
o
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
d
a
t
a
se
t
a
n
d
f
r
a
mew
o
r
k
,
”
I
n
t
e
r
n
e
t
o
f
T
h
i
n
g
s
,
v
o
l
.
3
6
,
M
a
r
.
2
0
2
6
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
o
t
.
2
0
2
6
.
1
0
1
8
9
1
.
[
1
6
]
A
.
A
l
a
l
m
a
i
e
,
P
.
N
a
n
d
a
,
a
n
d
X
.
H
e
,
“
Ze
r
o
t
r
u
st
n
e
t
w
o
r
k
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
sy
st
e
m
(
N
I
D
S
)
u
si
n
g
a
u
t
o
e
n
c
o
d
e
r
f
o
r
a
t
t
e
n
t
i
o
n
-
b
a
se
d
C
N
N
-
B
i
LS
TM
,
”
i
n
2
0
2
3
A
u
st
r
a
l
a
s
i
a
n
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
We
e
k
,
Ja
n
.
2
0
2
3
,
p
p
.
1
–
9
,
d
o
i
:
1
0
.
1
1
4
5
/
3
5
7
9
3
7
5
.
3
5
7
9
3
7
6
.
[
1
7
]
J.
Y
i
n
,
B
.
H
o
u
,
J.
D
a
i
,
a
n
d
Y
.
Zu
,
“
A
C
N
N
-
B
i
LS
TM
met
h
o
d
b
a
se
d
o
n
a
t
t
e
n
t
i
o
n
mec
h
a
n
i
sm
f
o
r
c
l
a
ss
-
i
m
b
a
l
a
n
c
e
d
a
b
n
o
r
m
a
l
t
r
a
f
f
i
c
d
e
t
e
c
t
i
o
n
,
”
i
n
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
e
r
Vi
s
i
o
n
a
n
d
D
e
e
p
L
e
a
rn
i
n
g
,
Ja
n
.
2
0
2
4
,
p
p
.
1
–
6
,
d
o
i
:
1
0
.
1
1
4
5
/
3
6
5
3
7
8
1
.
3
6
5
3
8
0
7
.
[
1
8
]
I
.
Jeb
r
i
l
e
t
a
l
.
,
“
D
e
e
p
l
e
a
r
n
i
n
g
b
a
se
d
D
D
o
S
a
t
t
a
c
k
d
e
t
e
c
t
i
o
n
i
n
i
n
t
e
r
n
e
t
o
f
t
h
i
n
g
s:
a
n
o
p
t
i
m
i
z
e
d
C
N
N
-
B
i
LS
TM
a
r
c
h
i
t
e
c
t
u
r
e
w
i
t
h
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
a
n
d
r
e
g
u
l
a
r
i
z
a
t
i
o
n
t
e
c
h
n
i
q
u
e
s,
”
I
n
f
o
c
o
m
m
u
n
i
c
a
t
i
o
n
s
J
o
u
r
n
a
l
,
v
o
l
.
1
6
,
n
o
.
1
,
p
p
.
2
–
1
1
,
2
0
2
4
,
d
o
i
:
1
0
.
3
6
2
4
4
/
I
C
J.2
0
2
4
.
1
.
1
.
[
1
9
]
J.
S
e
h
g
a
l
,
R
.
K
u
mar,
a
n
d
R
.
K
.
B
e
d
i
,
“
A
d
a
p
t
i
v
e
a
n
d
e
x
p
l
a
i
n
a
b
l
e
S
L
W
A
M
-
B
i
LST
M
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
sy
st
e
m
w
i
t
h
d
r
i
f
t
a
w
a
r
e
n
e
ss
a
n
d
c
r
o
ss
-
d
a
t
a
s
e
t
e
v
a
l
u
a
t
i
o
n
,
”
i
n
2
0
2
6
7
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
M
o
b
i
l
e
C
o
m
p
u
t
i
n
g
a
n
d
S
u
s
t
a
i
n
a
b
l
e
I
n
f
o
rm
a
t
i
c
s
,
J
a
n
.
2
0
2
6
,
p
p
.
9
5
6
–
9
6
2
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
M
C
S
I
6
7
2
8
3
.
2
0
2
6
.
1
1
4
1
2
7
2
5
.
[
2
0
]
M
.
S
.
A
t
a
a
,
E.
E.
S
a
n
a
d
,
a
n
d
R
.
A
.
El
-
k
h
o
r
i
b
i
,
“
I
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
i
n
s
o
f
t
w
a
r
e
d
e
f
i
n
e
d
n
e
t
w
o
r
k
u
si
n
g
d
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
e
s,”
S
c
i
e
n
t
i
f
i
c
Re
p
o
r
t
s
,
v
o
l
.
1
4
,
n
o
.
1
,
N
o
v
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
3
8
/
s
4
1
5
9
8
-
0
2
4
-
7
9
0
0
1
-
1.
[
2
1
]
Y
.
F
u
,
Y
.
D
u
,
Z.
C
a
o
,
Q
.
L
i
,
a
n
d
W
.
X
i
a
n
g
,
“
A
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
s
i
o
n
d
e
t
e
c
t
i
o
n
w
i
t
h
i
mb
a
l
a
n
c
e
d
d
a
t
a
,
”
El
e
c
t
r
o
n
i
c
s
,
v
o
l
.
1
1
,
n
o
.
6
,
M
a
r
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
e
l
e
c
t
r
o
n
i
c
s
1
1
0
6
0
8
9
8
.
[
2
2
]
P
.
K
u
m
a
r
,
G
.
P
.
G
u
p
t
a
,
a
n
d
R
.
Tr
i
p
a
t
h
i
,
“
To
w
a
r
d
d
e
si
g
n
o
f
a
n
i
n
t
e
l
l
i
g
e
n
t
c
y
b
e
r
a
t
t
a
c
k
d
e
t
e
c
t
i
o
n
s
y
st
e
m
u
si
n
g
h
y
b
r
i
d
f
e
a
t
u
r
e
r
e
d
u
c
e
d
a
p
p
r
o
a
c
h
f
o
r
I
o
T
n
e
t
w
o
r
k
s
,
”
Ar
a
b
i
a
n
J
o
u
r
n
a
l
f
o
r
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
4
6
,
n
o
.
4
,
p
p
.
3
7
4
9
–
3
7
7
8
,
A
p
r
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
3
3
6
9
-
020
-
0
5
1
8
1
-
3.
[
2
3
]
D
.
G
u
m
u
s
b
a
s
,
T.
Y
l
d
r
m
,
A
.
G
e
n
o
v
e
se,
a
n
d
F
.
S
c
o
t
t
i
,
“
A
c
o
m
p
r
e
h
e
n
si
v
e
su
r
v
e
y
o
f
d
a
t
a
b
a
ses
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
met
h
o
d
s
f
o
r
c
y
b
e
r
se
c
u
r
i
t
y
a
n
d
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
s
y
s
t
e
ms
,
”
I
EE
E
S
y
st
e
m
s
J
o
u
r
n
a
l
,
v
o
l
.
1
5
,
n
o
.
2
,
p
p
.
1
7
1
7
–
1
7
3
1
,
J
u
n
.
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
JS
Y
S
T
.
2
0
2
0
.
2
9
9
2
9
6
6
.
[
2
4
]
M
.
Jo
u
h
a
r
i
,
H
.
B
e
n
a
d
d
i
,
a
n
d
K
.
I
b
r
a
h
i
mi
,
“
Ef
f
i
c
i
e
n
t
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
:
c
o
mb
i
n
i
n
g
X
2
f
e
a
t
u
r
e
s
e
l
e
c
t
i
o
n
w
i
t
h
C
N
N
-
B
i
LST
M
o
n
t
h
e
U
N
S
W
-
N
B
1
5
d
a
t
a
s
e
t
,
”
i
n
2
0
2
4
1
1
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
W
i
re
l
e
ss
N
e
t
w
o
rks
a
n
d
M
o
b
i
l
e
C
o
m
m
u
n
i
c
a
t
i
o
n
s
,
Ju
l
.
2
0
2
4
,
p
p
.
1
–
6
,
d
o
i
:
1
0
.
1
1
0
9
/
W
I
N
C
O
M
6
2
2
8
6
.
2
0
2
4
.
1
0
6
5
8
0
9
9
.
[
2
5
]
B
.
L
i
,
J.
Li
,
a
n
d
M
.
J
i
a
,
“
A
D
F
C
N
N
-
B
i
LST
M
:
a
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
b
a
se
d
o
n
a
t
t
e
n
t
i
o
n
a
n
d
d
e
f
o
r
m
a
b
l
e
c
o
n
v
o
l
u
t
i
o
n
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
e
n
so
r
s
,
v
o
l
.
2
5
,
n
o
.
5
,
F
e
b
.
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
s2
5
0
5
1
3
8
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dee
p
lea
r
n
in
g
in
tr
u
s
io
n
d
etec
tio
n
fo
r
s
o
ftw
a
r
e
-
d
efin
ed
…
(
P
r
a
jw
a
la
s
imh
a
S
in
d
u
g
a
tta
N
a
g
a
r
a
ja
)
3711
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Dr
.
Pra
jwa
la
si
m
h
a
S
i
n
d
u
g
a
tt
a
Na
g
a
r
a
ja
is r
e
se
a
rc
h
a
ss
o
c
iate
a
t
Da
y
a
n
a
n
d
a
S
a
g
a
r
Un
iv
e
rsity
,
h
e
h
a
s
in
tr
o
d
u
c
e
d
a
n
in
n
o
v
a
ti
v
e
c
h
a
o
ti
c
fra
m
e
wo
rk
ti
tl
e
d
n
-
d
ime
n
si
o
n
a
l
m
u
lt
i
-
li
n
e
a
r
tran
sfo
rm
a
ti
o
n
(n
D
-
M
L
T):
n
o
n
-
l
in
e
a
r
d
y
n
a
m
ics
a
p
p
li
c
a
ti
o
n
to
c
ry
p
t
o
g
ra
p
h
y
.
He
re
c
e
iv
e
d
h
is
P
h
.
D.
i
n
Cry
p
t
o
g
ra
p
h
y
a
n
d
Cy
b
e
r
S
e
c
u
rit
y
fro
m
Visv
e
sv
a
ra
y
a
Tec
h
n
o
l
o
g
ica
l
Un
i
v
e
rsity
.
His
re
se
a
rc
h
c
o
n
tri
b
u
t
io
n
s
fo
c
u
s
o
n
d
e
v
e
lo
p
in
g
se
c
u
re
a
n
d
e
fficie
n
t
c
ry
p
to
g
ra
p
h
ic
fra
m
e
wo
rk
s
fo
r
c
y
b
e
r
se
c
u
rit
y
,
Io
T
e
c
o
sy
ste
m
s
,
a
n
d
n
e
x
t
-
g
e
n
e
ra
ti
o
n
in
tell
ig
e
n
t
sy
ste
m
s.
He
h
a
s
a
c
ti
v
e
ly
c
o
n
tri
b
u
ted
to
t
h
e
a
d
v
a
n
c
e
m
e
n
t
o
f
se
c
u
re
c
o
m
m
u
n
ica
ti
o
n
tec
h
n
o
l
o
g
ies
t
h
ro
u
g
h
re
se
a
rc
h
in
c
h
a
o
ti
c
sy
ste
m
s,
m
a
th
e
m
a
ti
c
a
l
m
o
d
e
li
n
g
,
ML
fo
r
c
y
b
e
r
d
e
fe
n
se
,
a
n
d
q
u
a
n
t
u
m
c
ry
p
t
o
g
ra
p
h
y
.
His
b
ro
a
d
e
r
o
b
jec
ti
v
e
is
to
stre
n
g
th
e
n
d
i
g
it
a
l
se
c
u
rit
y
so
lu
ti
o
n
s
fo
r
so
c
ieta
l
a
n
d
in
d
u
strial
a
p
p
li
c
a
ti
o
n
s
th
r
o
u
g
h
i
n
terd
isc
ip
li
n
a
ry
re
se
a
rc
h
a
n
d
in
n
o
v
a
ti
o
n
.
H
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
p
ra
jwa
las
imh
a
.
sn
1
@g
m
a
il
.
c
o
m
.
Dr
.
Na
v
y
a
R
a
ja
shek
a
r
a
is
a
n
a
ss
istan
t
p
ro
fe
ss
o
r
in
th
e
De
p
a
rt
m
e
n
t
o
f
E
lec
tro
n
ics
a
n
d
Co
m
m
u
n
ica
ti
o
n
E
n
g
i
n
e
e
rin
g
a
t
Da
y
a
n
a
n
d
a
S
a
g
a
r
Un
iv
e
rsity
with
o
v
e
r
7
y
e
a
rs
o
f
tea
c
h
in
g
a
n
d
2
y
e
a
rs
o
f
i
n
d
u
stry
e
x
p
e
rie
n
c
e
.
S
h
e
c
o
m
p
lete
d
h
e
r
P
h
.
D.
i
n
Dig
it
a
l
S
i
g
n
a
l
P
ro
c
e
ss
in
g
i
n
2
0
2
4
a
n
d
h
a
s
p
u
b
li
s
h
e
d
re
se
a
rc
h
p
a
p
e
rs,
c
o
n
fe
re
n
c
e
a
rti
c
les
,
a
n
d
b
o
o
k
c
h
a
p
ters
i
n
re
p
u
ted
in
tern
a
ti
o
n
a
l
jo
u
rn
a
ls
a
n
d
S
p
rin
g
e
r
p
u
b
li
c
a
ti
o
n
s.
He
r
re
se
a
rc
h
i
n
tere
sts
in
c
lu
d
e
si
g
n
a
l
p
r
o
c
e
ss
in
g
,
wire
les
s se
n
so
r
n
e
two
rk
s
,
Io
T
,
d
e
e
p
lea
rn
in
g
,
AI,
b
i
o
m
e
d
ica
l
e
n
g
i
n
e
e
rin
g
,
a
n
d
d
ig
i
tal
e
lec
tro
n
ics
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
n
a
v
y
a
-
e
c
e
@d
su
.
e
d
u
.
in
.
Dr
.
Pu
shp
a
B
a
n
g
a
l
o
r
e
Ra
m
e
s
h
is
a
n
a
ss
istan
t
p
ro
fe
ss
o
r
a
t
Da
y
a
n
a
n
d
a
S
a
g
a
r
Co
ll
e
g
e
o
f
En
g
in
e
e
ri
n
g
wit
h
o
v
e
r
1
8
y
e
a
rs
o
f
a
c
a
d
e
m
ic
a
n
d
2
y
e
a
r
s
o
f
in
d
u
str
y
e
x
p
e
rien
c
e
.
S
h
e
h
o
l
d
s
a
P
h
.
D.
a
n
d
sp
e
c
ialize
s
in
a
n
ten
n
a
s,
5
G
tec
h
n
o
l
o
g
ies
,
a
n
d
R
F
sy
ste
m
s.
S
h
e
h
a
s
p
u
b
li
s
h
e
d
se
v
e
ra
l
S
c
o
p
u
s
-
in
d
e
x
e
d
re
se
a
rc
h
p
a
p
e
rs
a
n
d
h
o
ld
s
a
p
a
ten
t
in
5
G
a
n
ten
n
a
d
e
sig
n
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
p
u
sh
p
a
-
tce
@d
a
y
a
n
a
n
d
a
sa
g
a
r.
e
d
u
.
Dr
.
Dr
u
v
a
K
u
m
a
r
S
i
d
d
a
r
a
ju
is
a
n
a
ss
istan
t
p
r
o
fe
ss
o
r
in
th
e
De
p
a
rtme
n
t
o
f
El
e
c
tro
n
ics
a
n
d
C
o
m
m
u
n
ica
ti
o
n
En
g
i
n
e
e
rin
g
a
t
Da
y
a
n
a
n
d
a
S
a
g
a
r
Co
ll
e
g
e
o
f
E
n
g
i
n
e
e
rin
g
with
a
ro
u
n
d
1
5
y
e
a
rs
o
f
tea
c
h
in
g
e
x
p
e
rien
c
e
.
He
o
b
tai
n
e
d
h
is
B.
E
.
,
M
.
Tec
h
.
,
a
n
d
P
h
.
D.
d
e
g
re
e
s
fro
m
Visv
e
sv
a
ra
y
a
Tec
h
n
o
lo
g
ica
l
U
n
iv
e
rsit
y
a
n
d
h
a
s
p
u
b
li
sh
e
d
se
v
e
ra
l
re
se
a
rc
h
p
a
p
e
rs
i
n
in
tern
a
ti
o
n
a
l
j
o
u
r
n
a
ls
a
n
d
c
o
n
fe
r
e
n
c
e
s.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
VLS
I
,
e
m
b
e
d
d
e
d
sy
ste
m
s,
m
a
n
a
g
e
m
e
n
t,
a
n
d
e
n
trep
re
n
e
u
rsh
ip
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
d
ru
v
a
-
e
c
e
@d
a
y
a
n
a
n
d
a
sa
g
a
r.
e
d
u
.
Dr
.
S
a
n
th
o
sh
K
u
m
a
r
Ra
m
a
c
h
a
n
d
r
a
g
o
wd
a
is
a
n
a
ss
istan
t
p
ro
fe
ss
o
r
i
n
th
e
De
p
a
rtme
n
t
o
f
El
e
c
tro
n
ics
a
n
d
Co
m
m
u
n
ica
ti
o
n
E
n
g
i
n
e
e
rin
g
a
t
Da
y
a
n
a
n
d
a
S
a
g
a
r
C
o
ll
e
g
e
o
f
En
g
i
n
e
e
rin
g
wi
th
o
v
e
r
1
2
y
e
a
rs
o
f
a
c
a
d
e
m
ic
e
x
p
e
rien
c
e
.
He
re
c
e
iv
e
d
h
is
P
h
.
D.
i
n
Wi
re
les
s
Co
m
m
u
n
ica
ti
o
n
fro
m
Da
y
a
n
a
n
d
a
S
a
g
a
r
Un
iv
e
rsit
y
in
2
0
2
4
a
n
d
h
is
M
.
Tec
h
.
i
n
VLS
I
De
sig
n
a
n
d
Emb
e
d
d
e
d
S
y
ste
m
s
fro
m
Visv
e
sv
a
ra
y
a
Tec
h
n
o
l
o
g
ica
l
Un
i
v
e
rsity
.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
wire
les
s
c
o
m
m
u
n
ica
ti
o
n
,
VLS
I
d
e
sig
n
,
e
m
b
e
d
d
e
d
sy
ste
m
s
,
F
P
G
A
-
b
a
se
d
d
e
s
ig
n
,
a
n
d
I
o
T,
wit
h
se
v
e
ra
l
p
u
b
li
c
a
ti
o
n
s
in
re
p
u
te
d
p
e
e
r
-
re
v
iew
e
d
jo
u
r
n
a
ls
a
n
d
c
o
n
fe
r
e
n
c
e
s.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
sa
n
th
o
sh
-
e
c
e
@d
a
y
a
n
a
n
d
a
sa
g
a
r.
e
d
u
.
Dr
.
Tr
u
p
ti
S
h
r
ip
a
d
Ta
g
a
r
e
is
a
n
a
ss
istan
t
p
ro
fe
ss
o
r
in
th
e
De
p
a
rtme
n
t
o
f
El
e
c
tro
n
ics
a
n
d
Co
m
m
u
n
ica
ti
o
n
En
g
i
n
e
e
rin
g
a
t
Da
y
a
n
a
n
d
a
S
a
g
a
r
Co
ll
e
g
e
o
f
E
n
g
in
e
e
rin
g
.
S
h
e
e
a
rn
e
d
h
e
r
P
h
.
D.
i
n
Wi
re
les
s
S
e
n
so
r
Ne
two
r
k
s
fro
m
Da
y
a
n
a
n
d
a
S
a
g
a
r
Un
iv
e
rsit
y
i
n
2
0
2
3
a
n
d
h
a
s
p
u
b
li
sh
e
d
re
se
a
rc
h
p
a
p
e
rs,
c
o
n
fe
re
n
c
e
a
rti
c
les
,
a
n
d
S
p
ri
n
g
e
r
b
o
o
k
c
h
a
p
ters
in
re
p
u
te
d
jo
u
r
n
a
ls
a
n
d
c
o
n
fe
re
n
c
e
s.
He
r
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
wire
les
s
se
n
so
r
n
e
two
rk
s,
wire
les
s
c
o
m
m
u
n
ica
ti
o
n
,
a
n
d
d
ig
it
a
l
e
l
e
c
tro
n
ics
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
tru
p
ti
tag
a
re
-
e
c
e
@d
a
y
a
n
a
n
d
a
sa
g
a
r.
e
d
u
.
Evaluation Warning : The document was created with Spire.PDF for Python.