I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Ro
bo
t
ics a
nd
Aut
o
m
a
t
io
n
(
I
J
RA)
Vo
l.
1
5
,
No
.
3
,
Sep
tem
b
er
20
2
6
,
p
p
.
6
9
8
~
7
0
8
I
SS
N:
2722
-
2
5
8
6
,
DOI
:
1
0
.
1
1
5
9
1
/i
jr
a
.
v
1
5
i
3
.
pp
6
9
8
-
708
698
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
r
a
.
ia
esco
r
e.
co
m
G
ra
ph
-
g
uid
ed co
ntras
tive t
ra
nsfo
r
mer archit
ectu
re
for rob
ust
a
nd ex
pla
ina
ble
n
etwork in
trusio
n
detec
tion
Arc
ha
na
J
a
y
a
pa
l
1
,
K
a
ma
la
kk
a
nn
a
n So
m
a
s
un
da
ra
m
2
,
Ar
un
K
um
a
r
Ra
m
a
m
o
o
rt
hy
3
1
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
V
e
l
s I
n
st
i
t
u
t
e
o
f
S
c
i
e
n
c
e
,
Te
c
h
n
o
l
o
g
y
a
n
d
A
d
v
a
n
c
e
d
S
t
u
d
i
e
s
(
V
I
S
TA
S
)
,
C
h
e
n
n
a
i
,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
A
p
p
l
i
c
a
t
i
o
n
s,
S
c
h
o
o
l
o
f
C
o
m
p
u
t
i
n
g
S
c
i
e
n
c
e
s,
V
e
l
s I
n
st
i
t
u
t
e
o
f
S
c
i
e
n
c
e
,
Te
c
h
n
o
l
o
g
y
a
n
d
A
d
v
a
n
c
e
d
S
t
u
d
i
e
s
(
V
I
S
TA
S
)
,
C
h
e
n
n
a
i
,
I
n
d
i
a
3
D
i
g
i
t
a
l
F
o
r
e
n
si
c
s
a
n
d
C
y
b
e
r
S
e
c
u
r
i
t
y
,
U
n
i
v
e
r
si
t
y
o
f
S
o
u
t
h
W
a
l
e
s
,
Tr
e
f
o
r
e
s
t
,
U
n
i
t
e
d
K
i
n
g
d
o
m
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Feb
2
5
,
2
0
2
6
R
ev
is
ed
J
u
n
5
,
2
0
2
6
Acc
ep
ted
J
u
l 2
0
,
2
0
2
6
In
tru
si
o
n
d
e
tec
ti
o
n
s
y
ste
m
s
(IDS)
a
re
v
e
ry
in
str
u
m
e
n
tal
in
p
ro
tec
ti
n
g
c
o
n
tem
p
o
ra
ry
n
e
two
r
k
i
n
fra
stru
c
tu
re
s
a
g
a
in
st
th
e
e
v
e
r
-
a
d
v
a
n
c
in
g
c
y
b
e
ra
tt
a
c
k
s.
C
o
n
v
e
n
ti
o
n
a
l
sig
n
a
tu
re
-
b
a
se
d
a
n
d
m
a
c
h
in
e
lea
rn
i
n
g
-
e
n
a
b
led
IDS
so
l
u
ti
o
n
s
f
re
q
u
e
n
tl
y
h
a
v
e
d
if
ficu
lt
y
wh
e
n
it
c
o
m
e
s
t
o
h
ig
h
fa
lse
-
p
o
siti
v
e
ra
tes
,
in
a
b
il
it
y
to
fle
x
ib
l
y
a
d
a
p
t
t
o
n
o
v
e
l
a
tt
a
c
k
s,
a
n
d
t
h
e
lac
k
o
f
s
u
p
p
o
rt
f
o
r
c
o
m
p
lex
traffic
d
y
n
a
m
ics
.
Ne
w
d
e
e
p
lea
rn
in
g
a
rc
h
it
e
c
tu
re
s
h
a
v
e
b
e
tt
e
r
d
e
tec
ti
o
n
p
r
o
p
e
rti
e
s,
y
e
t
a
re
li
m
it
e
d
b
y
fe
a
tu
re
o
v
e
rlap
,
tem
p
o
ra
li
t
y
,
a
n
d
lac
k
o
f
e
x
te
n
siv
e
n
e
ss
t
o
g
e
n
e
ra
li
z
a
ti
o
n
in
c
h
a
n
g
in
g
n
e
two
rk
c
o
n
d
i
ti
o
n
s.
T
o
o
v
e
rc
o
m
e
th
e
se
issu
e
s,
th
is
p
a
p
e
r
p
re
se
n
ts
a
n
e
w
g
ra
p
h
-
g
u
id
e
d
c
o
n
tras
ti
v
e
tran
sfo
rm
e
r
-
b
a
se
d
in
tru
si
o
n
d
e
tec
ti
o
n
sy
ste
m
(G
CT
-
IDS)
wh
ich
a
ims
a
t
imp
ro
v
i
n
g
d
e
tec
ti
o
n
a
c
c
u
ra
c
y
a
n
d
ro
b
u
st
n
e
ss
a
n
d
p
re
se
rv
in
g
re
a
l
-
ti
m
e
fe
a
sib
il
it
y
.
T
h
e
fra
m
e
wo
rk
c
o
m
b
i
n
e
s
fe
a
tu
re
i
n
tera
c
ti
o
n
b
y
g
ra
p
h
m
o
d
e
li
n
g
,
c
o
n
tras
ti
v
e
re
p
re
se
n
tatio
n
lea
rn
i
n
g
,
a
n
d
a
s
p
a
rse
se
lf
-
a
tt
e
n
ti
o
n
tran
sfo
rm
e
r
to
e
ffe
c
ti
v
e
ly
lea
rn
g
l
o
b
a
l
traffic
re
latio
n
sh
ip
s
a
n
d
b
e
h
a
v
io
ra
l
v
a
riat
io
n
s.
Th
e
CS
E
-
CIC
-
IDS2
0
1
8
d
a
ta
is
u
se
d
t
o
tes
t
t
h
e
p
r
o
p
o
se
d
m
e
th
o
d
i
n
re
a
l
n
e
two
r
k
c
o
n
d
i
ti
o
n
s.
K
ey
w
o
r
d
s
:
C
o
n
tr
asti
v
e
r
ep
r
esen
tatio
n
lear
n
in
g
Gr
ap
h
-
g
u
id
ed
lear
n
in
g
I
n
tr
u
s
io
n
d
etec
tio
n
s
y
s
tem
Netwo
r
k
s
ec
u
r
ity
T
r
an
s
f
o
r
m
e
r
n
etwo
r
k
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
:
Kam
alak
k
a
n
n
a
n
So
m
asu
n
d
a
r
a
m
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
A
p
p
licatio
n
s
,
Sch
o
o
l o
f
C
o
m
p
u
tin
g
Scien
ce
s
,
Vels I
n
s
titu
te
o
f
Scien
ce
,
T
ec
h
n
o
lo
g
y
an
d
Ad
v
a
n
ce
d
St
u
d
ies (
VI
STAS)
Pallav
ar
am
,
C
h
en
n
ai
E
m
ail
:
k
an
n
an
.
s
cs@
v
is
tas.ac
.
i
n
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
ac
ce
ler
ated
g
r
o
wth
o
f
d
ig
ital
co
m
m
u
n
icatio
n
,
clo
u
d
c
o
m
p
u
tin
g
,
I
n
ter
n
et
-
of
-
T
h
in
g
s
(
I
o
T
)
,
an
d
r
o
b
o
tic
cy
b
e
r
-
p
h
y
s
ical
s
y
s
tem
s
(
C
P
S)
h
as
s
ig
n
if
ican
tly
in
cr
ea
s
ed
th
e
co
m
p
lex
ity
an
d
v
u
ln
e
r
ab
ilit
y
o
f
m
o
d
e
r
n
n
etwo
r
k
in
f
r
astru
ct
u
r
es.
I
n
r
o
b
o
tics
-
d
r
iv
en
en
v
ir
o
n
m
en
ts
s
u
ch
as
au
to
n
o
m
o
u
s
v
eh
icles,
in
d
u
s
tr
ial
r
o
b
o
ts
,
an
d
s
m
ar
t
h
ea
lth
ca
r
e
s
y
s
tem
s
,
co
n
tin
u
o
u
s
d
ata
ex
c
h
an
g
e
b
etwe
en
s
en
s
o
r
s
,
ac
tu
ato
r
s
,
an
d
c
o
n
tr
o
l
u
n
its
in
tr
o
d
u
ce
s
n
ew
attac
k
s
u
r
f
ac
es.
C
o
n
s
eq
u
en
tly
,
cy
b
e
r
th
r
ea
ts
s
u
ch
as
d
is
tr
ib
u
ted
d
en
ial
o
f
s
er
v
ice
(
DDo
S),
b
r
u
te
-
f
o
r
ce
attac
k
s
,
b
o
tn
ets,
an
d
in
f
iltra
tio
n
attac
k
s
co
n
tin
u
e
t
o
ev
o
l
v
e,
p
o
s
in
g
s
er
io
u
s
r
is
k
s
to
d
ata
co
n
f
id
en
tiality
,
in
teg
r
ity
,
av
ailab
ilit
y
,
an
d
o
p
er
atio
n
al
s
af
ety
o
f
r
o
b
o
tic
s
y
s
tem
s
[
1
]
.
I
n
tr
u
s
io
n
d
etec
tio
n
s
y
s
tem
s
(
I
DS)
h
av
e
th
er
ef
o
r
e
b
ec
o
m
e
ess
en
tial
co
m
p
o
n
en
ts
in
s
ec
u
r
in
g
r
o
b
o
tic
an
d
n
etwo
r
k
ed
en
v
ir
o
n
m
en
ts
b
y
en
ab
lin
g
co
n
tin
u
o
u
s
m
o
n
ito
r
in
g
an
d
ti
m
ely
id
en
tific
atio
n
o
f
m
alicio
u
s
ac
tiv
ities
[
2
]
.
Un
lik
e
p
r
ev
e
n
tiv
e
m
ec
h
a
n
is
m
s
,
I
DS
f
o
cu
s
o
n
d
etec
tin
g
ab
n
o
r
m
al
o
r
u
n
au
th
o
r
ized
b
eh
a
v
io
r
p
o
s
t
-
co
m
p
r
o
m
is
e
an
d
i
n
itiatin
g
ap
p
r
o
p
r
iate
r
esp
o
n
s
e
ac
tio
n
s
to
m
itig
ate
p
o
ten
tial d
am
ag
e
[
3
]
.
E
ar
ly
I
DS
s
o
lu
tio
n
s
p
r
im
ar
ily
r
elied
o
n
s
ig
n
atu
r
e
-
b
ased
d
etec
tio
n
,
wh
er
e
tr
af
f
ic
p
atter
n
s
a
r
e
m
atch
ed
ag
ain
s
t
k
n
o
wn
attac
k
s
ig
n
atu
r
es.
W
h
ile
ef
f
ec
tiv
e
f
o
r
k
n
o
wn
th
r
ea
ts
,
th
ese
s
y
s
tem
s
f
ail
to
d
etec
t
ze
r
o
-
d
ay
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
Gra
p
h
-
g
u
i
d
ed
co
n
tr
a
s
tive
tr
a
n
s
fo
r
mer a
r
ch
itectu
r
e
fo
r
r
o
b
u
s
t a
n
d
ex
p
l
a
in
a
b
le
…
(
A
r
ch
a
n
a
J
a
y
a
p
a
l
)
699
attac
k
s
an
d
r
eq
u
ir
e
f
r
eq
u
e
n
t
m
an
u
al
u
p
d
ates
[
4
]
,
[
5
]
.
An
o
m
aly
-
b
ased
I
DS
ap
p
r
o
ac
h
es,
l
ev
er
ag
in
g
s
tatis
tical
tech
n
iq
u
es
a
n
d
m
ac
h
in
e
lear
n
in
g
alg
o
r
ith
m
s
s
u
ch
as
d
ec
is
io
n
tr
ee
s
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
es,
k
-
n
ea
r
est
n
eig
h
b
o
r
s
,
an
d
r
a
n
d
o
m
f
o
r
e
s
ts
,
o
f
f
er
im
p
r
o
v
e
d
ad
ap
tab
i
lity
.
Ho
wev
er
,
th
ey
o
f
ten
s
tr
u
g
g
le
with
h
ig
h
-
d
im
en
s
io
n
al
d
ata,
f
ea
tu
r
e
g
e
n
er
aliza
tio
n
,
an
d
class
im
b
al
an
ce
—
ch
allen
g
es
th
at
ar
e
f
u
r
th
er
am
p
lifie
d
in
r
o
b
o
tics
en
v
ir
o
n
m
e
n
ts
with
h
e
ter
o
g
en
eo
u
s
s
en
s
o
r
d
ata
s
tr
ea
m
s
[
6
]
,
[
7
]
.
R
ec
en
t
ad
v
an
ce
m
en
ts
in
d
ee
p
lear
n
in
g
h
a
v
e
en
h
an
ce
d
I
DS
ca
p
ab
ilit
ies
th
r
o
u
g
h
au
to
m
at
ed
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
tem
p
o
r
al
m
o
d
elin
g
.
Mo
d
els
s
u
ch
as
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
s
(
C
NNs)
an
d
r
ec
u
r
r
en
t
n
eu
r
al
n
e
two
r
k
s
(
R
NNs),
in
clu
d
in
g
L
STM
a
n
d
B
iLST
M,
h
av
e
s
h
o
wn
s
tr
o
n
g
p
e
r
f
o
r
m
an
ce
in
ca
p
tu
r
in
g
s
eq
u
en
tial
tr
af
f
ic
p
atter
n
s
[
8
]
.
Ho
wev
er
,
th
ese
ap
p
r
o
ac
h
es
s
u
f
f
er
f
r
o
m
h
ig
h
c
o
m
p
u
tatio
n
a
l
o
v
er
h
ea
d
,
lim
ited
in
ter
p
r
etab
ilit
y
,
an
d
s
en
s
itiv
ity
to
n
o
is
y
o
r
r
ed
u
n
d
a
n
t
in
p
u
ts
.
I
n
r
o
b
o
tics
a
p
p
licatio
n
s
,
wh
er
e
r
ea
l
-
tim
e
d
ec
is
io
n
-
m
ak
in
g
is
cr
itical,
s
u
ch
lim
itatio
n
s
ca
n
d
eg
r
ad
e
s
y
s
tem
r
esp
o
n
s
iv
en
ess
an
d
r
eliab
ilit
y
.
Mo
r
eo
v
er
,
r
ec
u
r
r
en
t
ar
c
h
itectu
r
es
o
f
te
n
s
tr
u
g
g
le
with
lo
n
g
-
r
an
g
e
d
ep
en
d
en
cy
m
o
d
elin
g
in
co
m
p
lex
,
d
y
n
am
ic
tr
af
f
i
c
en
v
ir
o
n
m
en
ts
[
9
]
.
T
r
ad
itio
n
al
f
ea
tu
r
e
s
elec
tio
n
tech
n
iq
u
es
l
ik
e
p
r
in
cip
al
c
o
m
p
o
n
en
t
an
aly
s
is
(
PC
A)
,
m
u
tu
al
in
f
o
r
m
atio
n
,
a
n
d
r
ec
u
r
s
iv
e
f
ea
tu
r
e
elim
in
atio
n
r
e
d
u
ce
d
i
m
en
s
io
n
ality
b
u
t
f
ail
to
ca
p
tu
r
e
i
n
ter
-
f
ea
tu
r
e
r
elatio
n
s
h
ip
s
an
d
s
tr
u
ctu
r
al
d
e
p
en
d
en
cies in
h
e
r
en
t in
r
o
b
o
tic
co
m
m
u
n
icatio
n
n
etwo
r
k
s
[
1
0
]
,
[
1
1
]
.
T
o
ad
d
r
ess
th
ese
lim
itatio
n
s
,
th
is
p
ap
er
p
r
o
p
o
s
es
a
g
r
ap
h
-
g
u
id
e
d
c
o
n
tr
asti
v
e
tr
an
s
f
o
r
m
er
-
b
ased
in
tr
u
s
io
n
d
etec
tio
n
s
y
s
tem
(
G
C
T
-
I
DS)
tailo
r
ed
f
o
r
i
n
tellig
en
t
an
d
r
o
b
o
tic
n
etwo
r
k
e
n
v
ir
o
n
m
en
ts
[
1
2
]
.
T
h
e
p
r
o
p
o
s
ed
f
r
a
m
ewo
r
k
lev
er
ag
e
s
g
r
ap
h
-
b
ased
m
o
d
elin
g
to
ca
p
tu
r
e
s
tr
u
ctu
r
al
r
elatio
n
s
h
ip
s
am
o
n
g
f
ea
tu
r
es
a
n
d
co
m
m
u
n
icatio
n
n
o
d
es,
en
a
b
lin
g
m
o
r
e
co
n
tex
t
-
awa
r
e
f
ea
tu
r
e
r
ep
r
esen
tatio
n
.
C
o
n
tr
ast
iv
e
r
ep
r
esen
tatio
n
lear
n
in
g
is
em
p
lo
y
ed
t
o
g
en
e
r
ate
r
o
b
u
s
t
an
d
d
is
cr
im
in
ativ
e
em
b
ed
d
in
g
s
th
at
en
h
a
n
ce
g
e
n
e
r
aliza
tio
n
to
u
n
s
ee
n
attac
k
p
atter
n
s
,
p
ar
ticu
lar
ly
in
d
y
n
am
ic
r
o
b
o
tic
s
y
s
tem
s
[
1
3
]
.
Fu
r
th
er
m
o
r
e,
a
lig
h
t
weig
h
t
s
elf
-
atten
tio
n
tr
an
s
f
o
r
m
er
is
u
tili
ze
d
to
m
o
d
el
g
lo
b
al
d
ep
e
n
d
en
cies
ac
r
o
s
s
tr
af
f
ic
f
lo
ws
with
o
u
t
r
ely
in
g
o
n
co
m
p
u
tatio
n
ally
ex
p
en
s
iv
e
r
ec
u
r
r
en
t
m
ec
h
a
n
is
m
s
.
T
h
is
in
teg
r
ated
ar
ch
itectu
r
e
im
p
r
o
v
es
d
etec
tio
n
ac
cu
r
ac
y
,
r
o
b
u
s
tn
ess
,
s
ca
lab
ilit
y
,
an
d
in
ter
p
r
etab
ilit
y
,
m
ak
in
g
GC
T
-
I
DS
h
ig
h
ly
s
u
itab
le
f
o
r
s
ec
u
r
in
g
n
ex
t
-
g
en
er
atio
n
r
o
b
o
tic
an
d
cy
b
er
-
p
h
y
s
ical
n
etwo
r
k
in
f
r
ast
r
u
ctu
r
es
[
1
4
]
,
[
1
5
]
.
T
h
e
r
em
ain
d
er
o
f
th
is
p
ap
er
i
s
o
r
g
an
ized
as
f
o
llo
ws.
Sectio
n
2
p
r
esen
ts
a
co
m
p
r
eh
e
n
s
iv
e
r
ev
iew
o
f
r
ec
en
t a
n
d
r
elev
a
n
t stu
d
ies o
n
n
etwo
r
k
I
DS,
with
p
ar
ticu
lar
e
m
p
h
asis
o
n
m
ac
h
in
e
lear
n
in
g
an
d
d
ee
p
lear
n
i
n
g
–
b
ased
ap
p
r
o
ac
h
es,
h
ig
h
lig
h
ti
n
g
th
eir
s
tr
en
g
th
s
,
lim
itatio
n
s
,
an
d
r
esear
ch
g
ap
s
.
Sectio
n
3
d
escr
ib
es
th
e
p
r
o
p
o
s
ed
g
r
ap
h
-
g
u
i
d
ed
co
n
tr
asti
v
e
tr
an
s
f
o
r
m
e
r
–
b
ased
in
tr
u
s
io
n
d
etec
tio
n
m
eth
o
d
o
lo
g
y
in
d
etail,
in
clu
d
in
g
d
ata
p
r
ep
r
o
ce
s
s
in
g
,
f
ea
t
u
r
e
i
n
ter
ac
tio
n
m
o
d
elin
g
,
r
ep
r
esen
tatio
n
lear
n
in
g
,
a
n
d
class
if
icatio
n
m
ec
h
a
n
is
m
s
.
Sectio
n
4
d
is
cu
s
s
es
th
e
ex
p
er
im
en
tal
r
esu
lts
an
d
p
er
f
o
r
m
a
n
ce
ev
alu
atio
n
o
f
th
e
p
r
o
p
o
s
e
d
f
r
am
ew
o
r
k
u
n
d
er
r
ea
lis
tic
n
etwo
r
k
co
n
d
itio
n
s
.
Fin
ally
,
th
e
p
ap
e
r
co
n
clu
d
es
with
k
ey
f
in
d
in
g
s
an
d
o
u
tli
n
es
f
u
tu
r
e
r
esear
ch
d
ir
ec
tio
n
s
f
o
r
f
u
r
t
h
er
en
h
an
cin
g
in
tr
u
s
io
n
d
etec
tio
n
p
e
r
f
o
r
m
a
n
ce
an
d
a
d
ap
tab
ilit
y
.
2.
L
I
T
E
R
AT
U
RE
R
E
VI
E
W
T
h
e
d
ev
elo
p
m
en
t
o
f
in
ter
n
et
u
s
ag
e
p
r
o
d
u
ce
d
v
ar
i
o
u
s
tr
af
f
i
c
with
s
ig
n
s
o
f
n
u
m
er
o
u
s
cy
b
er
attac
k
s
.
T
h
e
latest d
ataset
C
SE
-
C
I
C
-
I
DS2
0
1
8
was u
s
ed
to
ap
p
ly
d
e
ep
lear
n
in
g
m
et
h
o
d
s
in
in
tr
u
s
io
n
d
etec
tio
n
with
th
e
co
n
v
en
tio
n
al
ev
alu
atio
n
m
etr
i
cs.
Six
m
o
d
els,
s
u
ch
as
DNN
,
C
NN,
R
NN,
L
STM
,
C
NN
+
R
NN
an
d
C
NN
+
L
STM
,
wer
e
b
u
ilt
af
ter
p
r
ep
r
o
ce
s
s
in
g
an
d
b
i
n
ar
y
a
n
d
m
u
lti
-
class
ical
cla
s
s
if
icatio
n
an
d
d
etec
tio
n
o
f
b
en
ig
n
tr
af
f
ic
an
d
s
ix
ca
teg
o
r
ies
o
f
at
tack
s
wer
e
p
er
f
o
r
m
ed
[
1
6
]
.
T
h
e
ac
cu
r
ac
y
o
f
all
th
e
m
o
d
els
was
m
o
r
e
th
an
9
8
.
DNN,
C
N
N,
an
d
R
NN
m
o
d
els
wer
e
in
d
iv
id
u
ally
less
in
f
er
en
ce
-
tim
e
-
co
n
s
u
m
i
n
g
an
d
,
th
er
ef
o
r
e,
wer
e
m
o
r
e
ap
p
licab
le
in
p
r
ac
tice
wh
en
it
co
m
es to
I
DS a
p
p
licatio
n
s
.
C
lo
u
d
co
m
p
u
tin
g
f
ac
ilit
ated
th
e
o
n
-
d
em
an
d
-
b
asis
av
ailab
ilit
y
o
f
n
etwo
r
k
a
n
d
co
m
p
u
ter
r
eso
u
r
ce
s
s
u
ch
as
s
to
r
ag
e,
an
d
d
ata
m
an
ag
em
en
t,
an
d
im
p
r
o
v
ed
th
e
ef
f
icien
cy
o
f
th
e
s
y
s
tem
.
Ho
wev
er
,
with
th
ese
ad
v
an
tag
es,
clo
u
d
en
v
ir
o
n
m
e
n
ts
wer
e
ch
ar
ac
ter
ized
b
y
h
u
g
e
s
ec
u
r
ity
is
s
u
es
esp
ec
ially
wh
en
it
co
m
es
to
s
ec
u
r
in
g
r
eso
u
r
ce
s
a
n
d
s
er
v
ices.
I
n
o
r
d
e
r
to
o
v
er
co
m
e
th
e
s
e
is
s
u
es,
I
DS
wer
e
em
b
r
ac
e
d
to
tr
ac
k
n
etwo
r
k
tr
af
f
ic
an
d
d
etec
t
ab
n
o
r
m
al
tr
a
f
f
ic
[
1
7
]
.
A
r
an
d
o
m
f
o
r
est
an
d
f
ea
tu
r
e
en
g
in
ee
r
in
g
-
b
ased
clo
u
d
-
b
ased
m
o
d
el
o
f
in
tr
u
s
io
n
d
etec
tio
n
was
cr
ea
ted
to
en
h
an
ce
th
e
ac
c
u
r
ac
y
o
f
d
etec
tio
n
.
I
n
B
o
t
-
I
o
T
an
d
NSL
-
KDD
d
ataset
s
,
9
8
.
3
% a
n
d
9
9
.
9
9
g
av
e
th
e
h
ig
h
est ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
an
d
r
ec
all; th
u
s
,
in
d
icatin
g
h
i
g
h
p
e
r
f
o
r
m
an
ce
.
T
h
e
I
n
ter
n
et
o
f
T
h
i
n
g
s
also
lin
k
ed
p
h
y
s
ical
o
b
jects
in
d
if
f
er
en
t
f
ield
s
,
in
clu
d
in
g
tr
an
s
p
o
r
tatio
n
,
h
ea
lth
ca
r
e,
ag
r
icu
ltu
r
e,
an
d
d
ef
en
s
e,
an
d
allo
wed
th
em
to
b
e
u
s
ed
in
r
ea
l
-
tim
e
y
et
p
o
s
e
d
a
s
er
io
u
s
s
ec
u
r
ity
th
r
ea
t.
Pre
d
ictiv
e
I
o
T
n
etwo
r
k
s
ec
u
r
ity
was
n
o
t
ef
f
icien
t
w
ith
tr
ad
itio
n
al
s
ig
n
atu
r
e
-
an
d
r
u
le
-
b
ased
m
eth
o
d
s
[
1
8
]
.
Pear
s
o
n
co
r
r
elatio
n
c
o
ef
f
icien
t
-
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
m
o
d
el
was
d
esig
n
ed
a
s
an
I
DS
to
d
etec
t
n
etwo
r
k
an
o
m
alies.
T
h
e
u
s
e
o
f
im
p
o
r
tan
t
lin
ea
r
f
ea
tu
r
es
co
m
b
in
ed
with
C
NN
-
b
ased
lear
n
in
g
was
u
s
ed
in
b
in
ar
y
an
d
m
u
lticlas
s
attac
k
d
etec
tio
n
.
T
esti
n
g
o
n
NSL
-
KDD,
C
I
C
I
DS
-
2
0
1
7
an
d
I
OT
I
D2
0
d
atab
ases
d
em
o
n
s
tr
ated
h
ig
h
p
r
ec
is
io
n
o
f
9
9
.
8
9
an
d
ex
tr
e
m
ely
lo
w
m
is
class
if
icatio
n
.
T
h
e
h
ig
h
I
n
ter
n
et
tr
af
f
ic
r
ates
h
ad
d
iv
er
s
if
ied
an
d
co
m
p
lica
ted
m
alicio
u
s
attac
k
s
an
d
s
in
g
le
-
m
o
d
al
in
tr
u
s
io
n
d
etec
tio
n
m
o
d
els
h
a
d
n
o
t
m
ax
im
ized
th
e
r
ich
f
ea
tu
r
e
th
at
n
etwo
r
k
s
o
f
f
er
,
wh
i
ch
r
esu
lted
in
p
o
o
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
5
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
6
9
8
-
7
0
8
700
p
er
f
o
r
m
an
ce
.
T
h
is
lim
itatio
n
was
o
v
er
co
m
e
b
y
p
r
o
p
o
s
in
g
a
m
u
ltimo
d
al
h
y
b
r
id
p
ar
allel
n
etwo
r
k
in
tr
u
s
io
n
d
etec
tio
n
m
o
d
el
[
1
9
]
.
Statis
ti
ca
l
in
f
o
r
m
atio
n
an
d
r
aw
p
ay
lo
ad
d
ata
wer
e
u
s
ed
to
ex
tr
ac
t
n
etwo
r
k
tr
af
f
ic
f
ea
tu
r
es
th
r
o
u
g
h
s
p
ec
ialis
t
n
eu
r
al
ar
ch
itectu
r
es.
T
h
e
C
NNs
lear
n
th
e
s
p
atial
-
tem
p
o
r
al
f
ea
tu
r
es
v
ia
C
NN
-
L
STM
b
r
an
ch
es
an
d
s
tatis
tic
al
f
ea
tu
r
es
v
ia
C
NN
s
.
C
o
s
M
ar
g
in
Featu
r
e
f
u
s
io
n
with
a
C
o
s
Ma
r
g
in
class
if
ier
o
b
tain
ed
9
9
.
9
8
ac
cu
r
ac
y
o
n
b
e
n
ch
m
ar
k
d
atasets
.
I
t
in
tr
o
d
u
ce
d
a
n
ew
f
r
am
ewo
r
k
o
f
in
tr
u
s
io
n
d
etec
tio
n
,
th
at
i
s
,
it
was
ab
le
to
d
etec
t
cy
b
e
r
a
ttack
s
on
th
e
in
ter
n
et
o
f
v
eh
icles
en
v
i
r
o
n
m
en
t
s
u
c
h
as
D
o
S,
DDo
S,
DR
Do
S,
b
r
u
te
f
o
r
ce
,
b
o
tn
ets,
a
n
d
s
n
if
f
in
g
attac
k
s
.
Netwo
r
k
tr
af
f
ic
was
an
aly
ze
d
b
y
a
m
ac
h
in
e
lear
n
i
n
g
b
ased
s
y
s
tem
to
id
en
tify
ab
n
o
r
m
al
f
l
o
ws
in
th
e
n
etwo
r
k
[
2
0
]
.
T
h
e
m
eth
o
d
o
lo
g
y
in
cl
u
d
ed
p
r
e
p
r
o
ce
s
s
in
g
with
Z
-
s
co
r
e
n
o
r
m
aliza
tio
n
,
f
ea
tu
r
e
s
elec
tio
n
with
a
r
eg
r
ess
io
n
ap
p
r
o
ac
h
to
r
ed
u
c
e
th
e
co
m
p
lex
ity
o
f
t
h
e
alg
o
r
ith
m
s
,
an
d
tr
ain
ed
en
s
em
b
le
m
o
d
els
u
s
in
g
th
e
r
an
d
o
m
f
o
r
est
an
d
b
o
o
s
tin
g
alg
o
r
ith
m
s
.
T
ests
o
n
b
en
ch
m
ar
k
p
r
o
b
lem
s
attain
ed
m
o
r
e
th
a
n
9
9
.
8
ac
cu
r
ac
y
with
lo
w
d
etec
tio
n
laten
cy
,
wh
ich
was b
etter
th
an
cu
r
r
en
t p
r
ac
tic
es
.
3.
P
RO
P
O
SE
D
WO
RK
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
i
s
b
ased
o
n
a
s
y
s
tem
atic
s
eq
u
en
ce
o
f
wo
r
k
th
at
in
clu
d
es
f
l
o
w
-
b
ased
tr
af
f
ic
co
n
s
tr
u
ctio
n
u
s
in
g
th
e
C
SE
-
C
I
C
-
I
DS2
0
1
8
d
ata.
T
o
im
p
r
o
v
e
th
e
q
u
ality
o
f
d
ata
an
d
its
co
n
tex
tu
al
r
ep
r
esen
tatio
n
,
d
ata
clea
n
in
g
,
im
p
u
tatio
n
,
n
o
r
m
aliza
tio
n
,
a
n
d
tem
p
o
r
al
en
r
ich
m
en
t
ar
e
u
s
ed
.
Gr
ap
h
-
b
ased
f
ea
tu
r
e
s
elec
tio
n
f
in
d
s
s
tr
u
ctu
r
ally
in
f
lu
en
tial
f
ea
tu
r
es,
wh
ich
ar
e
co
n
v
er
te
d
in
to
th
e
laten
t
em
b
ed
d
in
g
s
th
r
o
u
g
h
co
n
tr
asti
v
e
Siam
ese
lear
n
in
g
.
A
g
r
ap
h
-
g
u
id
e
d
s
p
ar
s
e
tr
an
s
f
o
r
m
er
o
p
tim
ized
b
y
im
b
alan
ce
awa
r
e
lo
s
s
f
u
n
ctio
n
s
is
u
s
ed
to
c
o
n
d
u
ct
f
i
n
al
in
tr
u
s
io
n
class
if
icatio
n
wh
ich
allo
ws ac
cu
r
ate
an
d
r
ea
l tim
e
d
etec
tio
n
.
3
.
1
.
Da
t
a
s
et
des
cr
iptio
n a
nd
t
ra
f
f
ic
f
lo
w
co
ns
t
ruct
io
n
T
h
e
in
tr
u
s
io
n
d
etec
tio
n
f
r
am
e
wo
r
k
p
r
o
p
o
s
ed
is
test
ed
o
n
th
e
C
SE
-
C
I
C
-
I
DS2
0
1
8
d
ataset
wh
ich
is
am
o
n
g
th
e
m
o
s
t
co
m
p
lete
a
n
d
r
ea
lis
tic
b
en
ch
m
ar
k
d
atasets
th
at
ex
is
t
to
co
n
d
u
ct
r
esear
ch
o
n
n
etwo
r
k
in
tr
u
s
io
n
d
etec
tio
n
[
2
1
]
.
T
h
i
s
d
ata
was
cr
ea
ted
t
o
r
e
p
r
esen
t
r
ea
l
-
life
e
n
ter
p
r
is
e
n
etwo
r
k
co
n
d
itio
n
s
a
n
d
co
m
p
r
is
es
n
o
t
o
n
ly
a
b
r
o
ad
v
ar
iety
o
f
b
e
n
ig
n
tr
af
f
ic
b
u
t
also
v
ar
io
u
s
ty
p
es
o
f
m
o
d
e
r
n
cy
b
er
-
attac
k
s
.
I
t
in
clu
d
es
m
o
r
e
th
an
o
n
e
m
illi
o
n
n
etwo
r
k
f
lo
w
r
ec
o
r
d
s
,
all
o
f
wh
ich
h
av
e
a
r
ich
s
et
o
f
tr
af
f
ic
r
elate
d
attr
ib
u
tes
in
clu
d
in
g
p
ac
k
et
s
tatis
tics
,
p
r
o
to
co
l
b
eh
a
v
io
r
,
an
d
tim
in
g
in
f
o
r
m
atio
n
.
Div
er
s
ity
an
d
s
ize
o
f
th
e
d
ataset
r
e
n
d
er
it
esp
ec
ially
ap
p
r
o
p
r
iate
to
te
s
t
s
o
p
h
is
ticated
m
eth
o
d
o
lo
g
ies
o
f
in
tr
u
s
io
n
d
etec
tio
n
in
th
e
co
n
d
itio
n
s
o
f
r
ea
l
o
p
er
atio
n
s
.
=
{
|
(
,
,
)
=
(
,
,
)
,
|
−
|
≤
}
(
1
)
W
h
er
e
is
th
e
ℎ
b
id
ir
ec
tio
n
al
tr
a
f
f
ic
f
lo
w,
is
th
e
p
ac
k
et
;
,
ar
e
s
o
u
r
ce
an
d
d
esti
n
atio
n
a
d
d
r
ess
es,
is
th
e
p
r
o
to
co
l
t
y
p
e,
is
th
e
p
a
ck
et
tim
estam
p
an
d
ad
ap
tiv
e
tim
e
win
d
o
w.
As
o
p
p
o
s
ed
to
ex
am
in
in
g
r
aw
p
ac
k
ets s
ep
ar
ately
,
th
e
p
r
o
p
o
s
ed
s
o
lu
tio
n
b
r
ea
k
s
d
o
w
n
n
etw
o
r
k
tr
af
f
ic
in
to
two
-
wa
y
f
lo
w
s
ess
io
n
s
.
3
.
2
.
Da
t
a
clea
nin
g
a
nd
inte
g
rit
y
v
a
lid
a
t
io
n
T
h
e
q
u
ality
o
f
d
ata
is
f
u
n
d
am
en
tal
to
en
s
u
r
in
g
th
e
s
u
cc
ess
o
f
an
y
I
DS,
esp
ec
ially
in
th
e
ca
s
e
o
f
b
ig
an
d
h
eter
o
g
e
n
eo
u
s
n
etwo
r
k
tr
af
f
ic
d
ata.
T
h
e
q
u
ality
o
f
th
e
d
ata
co
n
tain
ed
in
t
h
e
C
SE
-
C
I
C
-
I
DS2
0
1
8
d
ataset
ca
n
s
till
in
clu
d
e
u
n
n
ec
ess
ar
y
r
ec
o
r
d
s
,
m
alf
o
r
m
ed
f
lo
ws
o
r
in
co
n
s
is
ten
cies
ad
d
ed
b
y
d
ata
ca
p
tu
r
es
an
d
ag
g
r
eg
atio
n
s
.
I
n
o
r
d
er
to
r
eso
l
v
e
th
ese,
an
ef
f
ec
tiv
e
p
r
o
ce
s
s
in
d
ata
clea
n
in
g
an
d
i
n
teg
r
ity
v
er
if
icatio
n
c
o
n
ce
p
t
is
em
p
lo
y
ed
b
ef
o
r
e
b
ein
g
an
aly
ze
d
f
u
r
th
er
.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
u
s
es
d
en
s
ity
-
b
ased
n
eig
h
b
o
r
h
o
o
d
v
alid
atio
n
to
d
eter
m
i
n
e
r
e
d
u
n
d
an
t
o
r
an
o
m
alo
u
s
f
lo
w
r
ec
o
r
d
s
r
ath
er
t
h
an
tr
a
d
itio
n
al
d
u
p
li
ca
te
r
em
o
v
al
wh
ich
u
s
es
th
e
ex
ac
t
m
atch
i
n
g
o
f
th
e
d
u
p
licates.
T
h
is
s
ch
em
e
an
aly
s
es
th
e
lo
ca
l
n
eig
h
b
o
r
h
o
o
d
s
tr
u
ctu
r
e
o
f
tr
a
f
f
ic
f
lo
ws
in
th
e
f
ea
tu
r
e
s
p
ac
e
an
d
elim
in
ates
r
ec
o
r
d
in
g
s
th
at
h
av
e
an
ab
n
o
r
m
ally
lar
g
e
s
im
ilar
ity
d
en
s
ity
wh
ich
ten
d
s
to
r
ep
r
esen
t
d
u
p
licatio
n
ar
tifa
cts
o
r
lo
g
g
in
g
ar
tifa
cts.
Als
o
,
f
lo
w
r
ec
o
r
d
s
co
n
t
ain
in
g
in
c
o
n
s
is
ten
t
p
r
o
to
co
l
f
ield
s
,
in
v
alid
tim
estam
p
s
o
r
i
n
v
alid
attr
ib
u
te
v
alu
es
ar
e
id
en
tifie
d
b
y
i
n
teg
r
ity
ch
ec
k
s
en
f
o
r
ce
d
b
y
r
u
les
an
d
f
ix
e
d
o
r
elim
in
ate
d
.
T
h
is
v
alid
atio
n
p
r
o
ce
d
u
r
e
g
u
a
r
an
tees
th
at
th
e
r
em
ai
n
d
er
d
at
a
is
tr
u
e
to
th
e
r
ea
l
n
etwo
r
k
ac
tiv
ity
,
s
o
th
er
e
is
le
s
s
b
ias
an
d
n
o
is
e
d
o
es
n
o
t
n
eg
ativ
ely
im
p
ac
t
o
n
th
e
s
u
b
s
eq
u
en
t
lear
n
in
g
p
h
ases
[
2
2
]
–
[
2
5
]
.
(
)
=
∑
(
|
|
−
|
|
≤
)
=
1
(
2
)
W
h
er
e
(
)
is
th
e
lo
ca
l
d
en
s
ity
o
f
s
am
p
le
;
,
ar
e
tr
af
f
ic
f
lo
w
f
ea
t
u
r
e
v
ec
to
r
s
,
is
th
e
n
eig
h
b
o
r
h
o
o
d
r
ad
iu
s
an
d
(
⋅
)
is
th
e
in
d
icato
r
f
u
n
ctio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
Gra
p
h
-
g
u
i
d
ed
co
n
tr
a
s
tive
tr
a
n
s
fo
r
mer a
r
ch
itectu
r
e
fo
r
r
o
b
u
s
t a
n
d
ex
p
l
a
in
a
b
le
…
(
A
r
ch
a
n
a
J
a
y
a
p
a
l
)
701
3
.
3
.
Da
t
a
i
m
pu
t
a
t
io
n a
nd
f
e
a
t
ure
no
rm
a
liza
t
io
n
T
h
e
d
atasets
co
llected
b
y
th
e
n
etwo
r
k
tr
a
f
f
ic
ar
e
o
f
te
n
ch
a
r
ac
ter
ized
b
y
th
e
p
r
esen
ce
o
f
m
is
s
in
g
o
r
in
co
m
p
lete
v
alu
es
b
ec
a
u
s
e
o
f
p
ac
k
et
lo
s
s
,
p
er
f
o
r
m
an
ce
c
o
n
s
tr
ain
ts
o
f
m
o
n
ito
r
i
n
g
,
o
r
s
en
s
o
r
f
ailu
r
es.
Su
ch
r
ec
o
r
d
s
s
h
o
u
ld
n
o
t
b
e
i
g
n
o
r
e
d
o
r
s
im
p
lis
tic
im
p
u
tatio
n
s
tr
ateg
ies
s
h
o
u
ld
b
e
u
s
ed
to
p
r
o
d
u
ce
a
d
is
to
r
ted
d
is
tr
ib
u
tio
n
o
f
f
ea
tu
r
es
a
n
d
p
o
o
r
p
e
r
f
o
r
m
an
ce
o
f
m
o
d
els.
W
h
er
e
th
e
n
u
m
e
r
ical
v
alu
es
ar
e
m
is
s
in
g
in
th
e
p
r
o
p
o
s
ed
wo
r
k
f
l
o
w,
a
K
-
n
e
ar
est
n
eig
h
b
o
r
s
(
KNN)
b
ase
d
s
im
ilar
ity
im
p
u
tatio
n
m
et
h
o
d
is
u
s
ed
.
T
h
is
tech
n
iq
u
e
f
in
d
s
clu
s
ter
s
o
f
t
h
e
tr
af
f
ic
f
lo
ws
th
at
s
h
ar
e
th
e
s
am
e
b
eh
av
io
r
al
p
r
o
p
er
ties
an
d
ap
p
r
o
x
im
ates
m
is
s
in
g
v
alu
es
u
s
in
g
th
e
lo
ca
l
s
tr
u
ctu
r
e
o
f
th
e
d
ata,
h
en
ce
m
ain
tain
s
th
e
in
h
er
en
t
r
e
latio
n
s
h
ip
s
am
o
n
g
attr
ib
u
tes
[
2
6
]
.
Af
ter
im
p
u
tatio
n
,
th
e
n
o
r
m
aliza
tio
n
o
f
f
ea
t
u
r
es
is
u
n
d
er
tak
e
n
in
o
r
d
er
to
p
r
o
v
i
d
e
n
u
m
er
ical
s
tab
ilit
y
an
d
u
n
if
o
r
m
ac
co
u
n
ti
n
g
o
f
f
ea
tu
r
es.
̂
,
=
1
∑
,
∈
(
)
(
3
)
W
h
er
e
̂
,
r
ep
r
esen
ts
th
e
im
p
u
t
ed
v
alu
e
o
f
f
ea
tu
r
e
,
(
)
is
th
e
s
et
o
f
n
ea
r
est
n
eig
h
b
o
r
s
an
d
,
is
f
ea
tu
r
e
o
f
n
eig
h
b
o
r
.
′
=
−
1
(
(
)
)
(
4
)
Her
e,
is
th
e
o
r
ig
in
al
f
ea
tu
r
e
v
alu
e,
(
)
is
th
e
em
p
ir
ical
cu
m
u
lativ
e
d
is
tr
ib
u
tio
n
f
u
n
ctio
n
,
−
1
(
⋅
)
is
th
e
in
v
er
s
e
tar
g
et
q
u
an
tile
f
u
n
ctio
n
an
d
′
is
th
e
n
o
r
m
alize
d
f
ea
t
u
r
e.
Fig
u
r
e
1
illu
s
tr
ates
th
e
ar
ch
itectu
r
e
o
f
th
e
p
r
o
p
o
s
ed
GC
T
-
I
DS f
r
am
ewo
r
k
.
Fig
u
r
e
1
.
GC
T
-
I
DS
s
y
s
tem
ar
ch
itectu
r
e
3
.
4
.
F
e
a
t
ure
enco
din
g
a
nd
t
em
po
ra
l e
nrichm
ent
So
m
e
o
f
th
e
ca
teg
o
r
ical
attr
ib
u
tes
in
th
e
d
ataset,
e.
g
.
,
p
r
o
t
o
co
l
ty
p
e
an
d
s
er
v
ice
id
en
tifie
r
s
ar
e
n
o
t
p
r
o
ce
s
s
ab
le
d
ir
ec
tly
b
y
lear
n
in
g
alg
o
r
ith
m
s
.
I
n
o
r
d
er
to
e
n
co
d
e
th
ese
attr
ib
u
tes
ef
f
icien
tly
th
e
s
u
g
g
ested
m
eth
o
d
o
l
o
g
y
ap
p
lies
tar
g
et
-
g
u
id
ed
s
tatis
tical
en
co
d
in
g
.
T
h
e
en
co
d
i
n
g
m
eth
o
d
s
u
b
s
titu
tes
ca
teg
o
r
ical
v
al
u
es
with
s
tatis
tica
lly
o
b
tain
ed
est
im
ates
d
ep
en
d
in
g
o
n
th
eir
co
n
d
itio
n
al
ass
o
ciatio
n
with
th
e
class
n
am
es
an
d
p
er
m
its
th
e
m
o
d
el
to
r
ec
o
r
d
d
is
cr
im
in
ativ
e
in
f
o
r
m
atio
n
with
litt
le
o
r
n
o
m
aj
o
r
g
r
o
wth
o
f
d
i
m
en
s
io
n
ality
.
T
h
is
m
eth
o
d
will
p
r
ev
en
t
th
e
s
p
a
r
s
ity
an
d
s
ca
lab
ilit
y
ch
allen
g
es
th
at
o
n
e
-
h
o
t
en
co
d
in
g
is
k
n
o
wn
to
b
r
in
g
in
lar
g
e
d
atasets
.
I
n
ad
d
itio
n
to
th
e
u
s
e
o
f
s
tatic
f
ea
tu
r
e
r
ep
r
esen
tatio
n
,
th
e
u
s
e
o
f
tem
p
o
r
al
d
y
n
a
m
ics
is
im
p
o
r
tan
t
in
th
e
s
ep
ar
atio
n
o
f
b
en
i
g
n
b
e
h
av
io
r
an
d
th
e
m
alicio
u
s
ac
tiv
ity
f
o
r
au
to
m
atio
n
.
I
n
o
r
d
er
to
ex
tr
ac
t
th
ese
d
y
n
am
ics,
t
h
e
m
et
h
o
d
o
lo
g
y
e
m
p
lo
y
s
tem
p
o
r
al
en
r
ich
m
en
t
in
ter
m
s
o
f
m
u
lti
-
s
ca
le
s
lid
in
g
win
d
o
w
an
al
y
s
is
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
5
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
6
9
8
-
7
0
8
702
E
n
tr
o
p
y
ch
a
n
g
e,
in
ter
-
a
r
r
iv
al
tim
e
v
o
latilit
y
an
d
b
u
r
s
t
in
te
n
s
ity
ar
e
tem
p
o
r
al
d
escr
ip
to
r
s
ca
lcu
lated
o
n
b
o
th
s
h
o
r
t
ter
m
an
d
lo
n
g
-
ter
m
wi
n
d
o
ws.
T
h
ese
ch
ar
ac
ter
is
tics
allo
w
th
e
s
y
s
tem
to
s
im
u
la
te
ch
an
g
in
g
tr
af
f
ic
p
atter
n
s
an
d
b
ein
g
ab
le
to
d
et
ec
t
s
u
b
tle
tim
e
v
ar
iatio
n
s
lin
k
ed
to
r
ec
o
n
n
aiss
an
ce
,
b
r
u
te
f
o
r
ce
ef
f
o
r
ts
o
r
lo
w
r
ates
attac
k
s
.
T
em
p
o
r
al
e
n
r
ich
m
en
t
is
a
p
ar
ticu
lar
ly
u
s
ef
u
l
a
p
p
r
o
ac
h
in
co
n
tex
tu
alizin
g
th
e
n
etwo
r
k
ac
tiv
ity
,
as we
ll a
s
p
r
o
v
id
in
g
b
etter
d
et
ec
tab
ilit
y
o
f
th
e
in
tr
u
s
io
n
s
ce
n
ar
io
s
o
f
a
m
o
r
e
s
o
p
h
is
ticated
n
atu
r
e.
(
)
=
[
|
=
]
(
5
)
(
)
is
th
e
en
co
d
ed
v
alu
e
f
o
r
ca
teg
o
r
y
;
is
th
e
class
lab
el
an
d
is
th
e
ca
teg
o
r
ical
f
ea
tu
r
e.
=
−
∑
l
og
=
1
(
6
)
is
th
e
en
tr
o
p
y
with
in
tim
e
win
d
o
w
,
is
th
e
p
r
o
b
ab
ilit
y
o
f
p
a
ck
et
ev
en
t
an
d
is
th
e
n
u
m
b
er
o
f
ev
en
ts
in
win
d
o
w.
3
.
5
.
G
ra
ph
-
ba
s
ed
f
ea
t
ure
s
elec
t
io
n
T
h
e
h
ig
h
ly
d
im
e
n
s
io
n
al
f
ea
tu
r
e
s
p
ac
es
ar
e
p
r
o
n
e
to
r
ed
u
n
d
an
cy
an
d
c
o
m
p
u
tin
g
in
ef
f
ici
en
cy
an
d
th
er
ef
o
r
e
r
ed
u
ctio
n
ca
n
b
e
a
v
er
y
cr
u
cial
p
h
ase
o
f
th
e
in
tr
u
s
io
n
d
etec
tio
n
p
r
o
ce
s
s
.
Un
lik
e
th
e
r
elev
an
ce
-
b
ased
o
r
elim
in
atio
n
-
b
ased
s
elec
tio
n
m
eth
o
d
s
,
th
e
f
r
am
ew
o
r
k
i
n
q
u
esti
o
n
u
s
es
a
g
r
ap
h
-
b
ased
f
ea
tu
r
e
s
elec
tio
n
ap
p
r
o
ac
h
.
Her
e,
f
ea
tu
r
es
ar
e
r
ep
r
esen
ted
as
th
e
n
o
d
es
o
f
a
f
ea
tu
r
e
i
n
ter
ac
tio
n
g
r
ap
h
,
f
ea
tu
r
es
th
at
ar
e
s
tatis
t
ically
d
ep
en
d
en
t
o
n
ea
ch
o
th
er
illu
s
tr
ated
as th
e
ed
g
es o
f
th
e
g
r
ap
h
.
(
)
=
∑
⋅
(
)
=
1
(
7
)
W
h
er
e
(
)
is
th
e
ce
n
tr
ality
s
co
r
e
o
f
f
ea
tu
r
e
n
o
d
e
;
is
th
e
ad
jace
n
cy
m
atr
ix
an
d
is
th
e
n
u
m
b
er
o
f
f
ea
tu
r
es.
T
h
is
g
r
ap
h
is
an
al
y
ze
d
u
s
in
g
s
p
ec
tr
al
ce
n
tr
ality
an
aly
s
is
in
o
r
d
e
r
to
d
eter
m
in
e
th
e
s
tr
u
ct
u
r
al
in
f
lu
en
tial
f
ea
tu
r
es
th
at
ar
e
ce
n
tr
al
to
th
e
en
tire
in
te
r
ac
tio
n
n
etwo
r
k
.
Attr
ib
u
tes
th
at
h
av
e
a
h
ig
h
ce
n
tr
ality
ar
e
k
ep
t,
as th
ey
p
la
y
an
im
p
o
r
tan
t r
o
le
in
th
e
g
e
n
er
al
r
e
p
r
esen
ta
tio
n
o
f
tr
af
f
ic
b
eh
a
v
io
r
.
T
h
is
g
r
ap
h
-
b
ased
p
r
o
ce
s
s
o
f
s
elec
tio
n
is
n
o
t
o
n
ly
a
d
im
en
s
io
n
ality
r
ed
u
cti
o
n
alg
o
r
ith
m
b
u
t
also
ten
d
s
to
r
etain
co
m
p
lex
r
elatio
n
s
h
i
p
s
b
etwe
en
f
ea
tu
r
es
th
at
ar
e
f
r
e
q
u
en
tly
ig
n
o
r
ed
b
y
o
th
e
r
s
elec
tio
n
tech
n
iq
u
es.
C
o
n
s
eq
u
en
tly
,
th
e
ch
o
s
en
f
ea
tu
r
e
s
et
is
n
o
t
o
n
ly
s
m
all
b
u
t
also
in
f
o
r
m
ativ
e
an
d
,
th
er
e
f
o
r
e,
en
h
an
ce
s
b
o
t
h
co
m
p
u
tatio
n
al
ef
f
icien
cy
an
d
d
etec
tab
ilit
y
.
3
.
6
.
P
r
o
po
s
ed
g
ra
ph
-
g
uid
ed
co
ntr
a
s
t
iv
e
t
ra
ns
f
o
r
m
er
(
G
C
T
-
I
DS)
T
h
e
last
s
tag
e
o
f
clas
s
if
icatio
n
is
ca
r
r
ied
o
u
t
with
th
e
h
elp
o
f
th
e
p
r
o
p
o
s
ed
g
r
ap
h
-
g
u
id
ed
c
o
n
tr
asti
v
e
tr
an
s
f
o
r
m
er
th
at
is
k
n
o
wn
as
GC
T
-
I
DS.
T
h
e
m
o
d
el
co
m
b
i
n
es
th
e
b
e
n
ef
its
o
f
g
r
ap
h
-
in
s
p
ir
ed
f
ea
tu
r
e
s
elec
tio
n
an
d
co
n
tr
asti
v
e
r
ep
r
esen
tatio
n
lear
n
in
g
an
d
th
e
e
x
p
r
ess
iv
ity
o
f
tr
an
s
f
o
r
m
e
r
-
b
ased
m
o
d
els.
T
r
an
s
f
o
r
m
e
r
en
co
d
er
u
s
es
s
p
ar
s
e
s
elf
-
atten
tio
n
,
an
d
it
is
ab
le
to
lear
n
g
lo
b
al
d
ep
en
d
e
n
cy
ac
r
o
s
s
tr
af
f
ic
em
b
ed
d
in
g
s
at
lo
w
co
m
p
u
tatio
n
.
(
,
,
)
=
(
√
⊙
)
(
8
)
w
h
er
e
,
,
ar
e
q
u
er
y
,
k
ey
,
v
alu
e
m
atr
ices,
is
th
e
k
ey
d
im
en
s
io
n
,
is
th
e
s
p
ar
s
ity
m
ask
an
d
⊙
is
th
e
elem
en
t
-
wis
e
m
u
ltip
licatio
n
.
As
o
p
p
o
s
ed
to
r
ec
u
r
r
en
t
a
r
ch
itectu
r
es,
th
e
tr
a
n
s
f
o
r
m
er
p
r
o
ce
s
s
es
em
b
ed
d
in
g
s
p
ar
allel
to
ea
ch
o
th
er
,
wh
ich
m
ea
n
s
th
at
th
ey
ca
n
b
e
tr
ain
e
d
an
d
in
f
er
r
ed
f
aster
.
T
h
e
atten
tio
n
m
ec
h
an
is
m
is
d
y
n
am
ically
c
alcu
lated
an
d
weig
h
s
v
ar
ied
co
m
p
o
n
en
ts
o
f
th
e
ac
q
u
ir
ed
r
ep
r
esen
tatio
n
s
wh
ich
e
n
ab
les
th
e
m
o
d
el
to
co
n
ce
n
tr
ate
o
n
m
o
s
t
p
er
tin
en
t
p
atter
n
o
f
b
eh
av
io
r
in
in
tr
u
s
io
n
class
if
icatio
n
.
T
h
e
R
an
g
er
o
p
tim
izer
is
u
s
ed
to
p
er
f
o
r
m
m
o
d
el
o
p
ti
m
izatio
n
,
an
d
it
is
a
co
m
b
in
at
io
n
o
f
r
ec
tifie
d
ad
a
p
tiv
e
m
o
m
en
t
esti
m
atio
n
an
d
lo
o
k
ah
ea
d
s
tr
ateg
ies
to
en
h
an
ce
co
n
v
er
g
e
n
ce
s
tab
ilit
y
.
Mo
r
eo
v
er
,
f
o
ca
l
lo
s
s
is
u
s
ed
in
th
e
tr
ain
in
g
p
r
o
ce
s
s
to
d
ea
l
with
th
e
p
r
o
b
lem
o
f
th
e
im
b
alan
ce
b
etwe
en
class
es
b
y
f
o
cu
s
in
g
o
n
th
o
s
e
m
in
o
r
ity
attac
k
s
th
at
ar
e
d
if
f
icu
lt
to
class
if
y
.
C
o
llectiv
ely
,
th
e
ab
o
v
e
d
esig
n
d
ec
is
io
n
s
ca
n
lead
t
o
a
s
tr
o
n
g
an
d
s
ca
lab
le
in
tr
u
s
io
n
d
etec
to
r
m
o
d
el
th
at
ca
n
p
er
f
o
r
m
well
ag
ain
s
t d
y
n
am
ic
a
n
d
v
ar
io
u
s
th
r
ea
ts
in
th
e
n
etwo
r
k
.
=
−
(
1
−
)
l
og
(
)
(
9
)
wh
er
e
is
th
e
p
r
e
d
icted
p
r
o
b
ab
ilit
y
f
o
r
tr
u
e
class
,
is
th
e
class
b
alan
cin
g
f
ac
to
r
,
an
d
is
th
e
f
o
cu
s
in
g
p
ar
am
eter
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
Gra
p
h
-
g
u
i
d
ed
co
n
tr
a
s
tive
tr
a
n
s
fo
r
mer a
r
ch
itectu
r
e
fo
r
r
o
b
u
s
t a
n
d
ex
p
l
a
in
a
b
le
…
(
A
r
ch
a
n
a
J
a
y
a
p
a
l
)
703
Alg
o
r
ith
m
1
.
Gr
ap
h
-
g
u
id
e
d
co
n
tr
asti
v
e
tr
an
s
f
o
r
m
er
-
b
ased
in
t
r
u
s
io
n
d
etec
tio
n
(
GC
T
-
I
DS)
Input:
Raw network traffic packets
=
{
1
,
2
,
…
,
}
, time window
Δ
, number of neighbors
,
graph threshold
, contrastive margin
Output:
Predicted traffic class
̂
∈
{
,
}
1. Flow Construction:
Group packets into bidirectional flows using
=
{
|
(
,
,
)
=
(
,
,
)
,
|
−
|
≤
Δ
}
2. For
each flow feature vector
, compute local density
(
)
=
∑
(
|
|
−
|
|
≤
)
=
1
Remove flows with abnormal density.
3. For
missing feature
in flow
, estimate using
̂
,
=
1
∑
,
∈
(
)
4. Apply
quantile normalization
′
=
−
1
(
(
)
)
5. Encode
categorical feature
using
(
)
=
[
|
=
]
6. Compute
entropy over sliding window
=
−
∑
lo
g
=
1
7. Construct
feature interaction graph
(
,
)
Compute centrality:
(
)
=
∑
⋅
(
)
=
1
Select top
-
central features.
8. Generate
embeddings
,
using Siamese encoder
Optimize contrastive loss
=
⋅
2
+
(
1
−
)
⋅
max
(
0
,
−
)
2
9. Apply
sparse self
-
attention
(
,
,
)
=
(
√
⊙
)
10. Compute
focal loss:
=
−
(
1
−
)
lo
g
(
)
Update parameters using Ranger optimizer.
11. Assign
class label:
̂
=
arg
max
(
|
)
Return
̂
End Algorithm
standards.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
I
n
th
e
p
r
o
p
o
s
ed
GC
T
-
I
DS,
i
t
is
th
e
wo
r
k
in
g
p
r
in
cip
le
b
ased
ar
o
u
n
d
th
e
tr
an
s
f
o
r
m
ati
o
n
o
f
r
aw
n
etwo
r
k
t
r
af
f
ic
i
n
to
s
em
an
tic
ally
r
ich
r
ep
r
esen
tatio
n
s
th
at
ca
n
b
e
u
s
ed
to
p
er
f
o
r
m
ac
c
u
r
ate
an
d
r
ea
l
-
tim
e
d
etec
tio
n
o
f
in
tr
u
s
io
n
s
.
T
o
en
s
u
r
e
th
e
r
eliab
ilit
y
an
d
r
o
b
u
s
tn
ess
o
f
th
e
p
r
o
p
o
s
ed
GC
T
-
I
D
S
m
o
d
el,
ex
ten
s
iv
e
s
tatis
t
ical
v
alid
atio
n
was
p
er
f
o
r
m
ed
.
T
h
e
en
tire
e
x
p
er
im
e
n
tal
p
ip
elin
e,
in
cl
u
d
in
g
d
ata
p
r
ep
r
o
ce
s
s
in
g
,
s
tr
atif
ied
d
ata
s
p
litt
in
g
,
m
o
d
el
tr
ain
i
n
g
,
an
d
ev
alu
atio
n
,
was
ex
ec
u
ted
1
0
in
d
e
p
en
d
e
n
t
tim
es.
I
n
ea
ch
r
u
n
,
r
an
d
o
m
in
itializatio
n
an
d
d
ata
s
h
u
f
f
lin
g
wer
e
a
p
p
lied
wh
ile
m
ain
tai
n
in
g
c
o
n
s
is
ten
t
s
tr
atif
ied
tr
ain
in
g
,
v
ali
d
atio
n
,
an
d
test
in
g
s
p
lits
to
en
s
u
r
e
f
air
n
ess
an
d
r
e
p
r
o
d
u
cib
ilit
y
.
T
h
e
p
e
r
f
o
r
m
an
ce
o
f
th
e
m
o
d
el
was
ev
alu
ate
d
u
s
in
g
s
ta
n
d
ar
d
class
if
icatio
n
m
etr
ics,
in
clu
d
in
g
Acc
u
r
ac
y
,
Pre
cisi
o
n
,
R
ec
all,
an
d
F1
-
s
co
r
e.
T
h
e
r
esu
lts
o
f
th
ese
test
s
co
n
f
ir
m
th
at
th
e
p
er
f
o
r
m
an
ce
im
p
r
o
v
em
e
n
ts
ac
h
iev
e
d
b
y
G
C
T
-
I
DS
ar
e
s
tatis
tically
s
ig
n
if
ican
t
(
p
<
0
.
0
5
)
ac
r
o
s
s
all
ev
al
u
atio
n
m
etr
ics.
T
h
is
d
em
o
n
s
tr
ates
th
at
th
e
p
r
o
p
o
s
ed
m
o
d
el
c
o
n
s
is
ten
tly
o
u
tp
er
f
o
r
m
s
ex
is
tin
g
ap
p
r
o
ac
h
es
an
d
th
at
t
h
e
im
p
r
o
v
em
e
n
ts
ar
e
n
o
t
d
u
e
to
r
an
d
o
m
v
a
r
iatio
n
.
Fig
u
r
e
2
s
h
o
ws
th
e
to
tal
d
etec
tio
n
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
GC
T
-
I
DS
m
o
d
el
i
n
v
ar
i
o
u
s
m
etr
ics
o
f
e
v
alu
a
tio
n
.
Fig
u
r
e
3
d
ep
icts
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
s
u
g
g
ested
GC
T
-
I
DS
m
o
d
el
in
ter
m
s
o
f
th
e
d
etec
tio
n
o
f
v
ar
io
u
s
ty
p
es
o
f
tr
af
f
ic.
T
h
e
f
in
d
in
g
s
r
ev
ea
l
a
h
ig
h
lev
el
o
f
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e
in
b
o
th
b
e
n
ig
n
a
n
d
m
alicio
u
s
class
e
s
th
at
s
h
o
ws
a
b
alan
ce
d
ab
ilit
y
in
d
etec
tio
n
.
T
h
e
f
r
eq
u
e
n
t
attac
k
ty
p
es
th
at
th
e
m
o
d
el
p
er
f
o
r
m
ed
well
in
clu
d
e
Do
S,
DDo
S
an
d
Po
r
tScan
,
wh
ich
in
d
icate
th
at
th
e
m
o
d
el
was
ef
f
ec
tiv
e
in
d
etec
tin
g
a
p
r
ev
alen
t
p
atter
n
o
f
in
tr
u
s
io
n
.
T
h
e
f
r
am
ewo
r
k
its
elf
co
n
tin
u
es
to
r
ec
o
r
d
g
o
o
d
p
er
f
o
r
m
an
ce
o
f
lo
w
-
f
r
e
q
u
en
c
y
an
d
s
tealth
y
attac
k
s
lik
e
in
f
iltra
tio
n
an
d
Slo
wlo
r
is
wh
ich
ar
e
ty
p
ically
c
h
allen
g
in
g
to
tr
ac
e
s
in
ce
th
e
y
h
av
e
s
m
all
b
eh
av
io
r
al
p
r
in
ts
.
T
h
e
r
an
k
in
g
s
o
f
s
u
ch
attac
k
s
ar
e
s
lig
h
tly
lo
wer
y
et
th
ey
ar
e
s
till
well
ab
o
v
e
th
e
r
an
g
e
o
f
ac
ce
p
tab
le
lev
els
wh
ich
p
o
in
t
to
s
tr
o
n
g
g
en
er
aliza
tio
n
.
T
h
is
u
n
iv
er
s
al
b
eh
av
io
r
in
r
esp
o
n
s
e
to
th
e
class
es
m
ay
b
e
ex
p
lain
ed
b
y
c
o
n
tr
asti
v
e
r
e
p
r
esen
tatio
n
lear
n
in
g
m
ec
h
a
n
is
m
,
wh
ich
im
p
r
o
v
es
d
is
cr
im
in
atio
n
b
etwe
en
tr
a
f
f
ic
b
eh
av
io
r
s
,
an
d
tr
an
s
f
o
r
m
er
-
b
a
s
ed
g
lo
b
al
d
ep
en
d
en
cy
m
o
d
elin
g
th
at
in
c
o
r
p
o
r
ates
f
in
e
v
ar
i
atio
n
s
am
id
tr
af
f
ic
f
lo
ws.
Fig
u
r
e
4
is
th
e
co
m
p
ar
is
o
n
b
etwe
en
th
e
p
r
o
p
o
s
ed
GC
T
-
I
DS
an
d
th
e
tr
ad
itio
n
a
l
m
ac
h
in
e
lear
n
in
g
m
o
d
el
an
d
th
e
d
ee
p
lear
n
i
n
g
-
b
ased
in
tr
u
s
io
n
d
etec
tio
n
m
o
d
els.
C
las
s
ical
m
eth
o
d
s
lik
e
th
e
lo
g
is
tic
r
eg
r
ess
io
n
an
d
d
ec
is
io
n
tr
ee
s
h
a
v
e
p
o
o
r
p
er
f
o
r
m
an
ce
b
ec
au
s
e
th
e
y
ca
n
n
o
t
p
r
o
v
id
e
a
n
o
n
lin
ea
r
r
elatio
n
s
h
ip
an
d
tem
p
o
r
al
r
elatio
n
s
h
ip
.
R
an
d
o
m
f
o
r
est
an
d
g
r
a
d
ien
t
b
o
o
s
tin
g
ar
e
e
n
s
em
b
le
m
eth
o
d
s
th
at
ca
n
b
e
u
s
ed
to
en
h
a
n
ce
ac
cu
r
ac
y
,
n
o
n
eth
eless
,
th
ese
m
eth
o
d
s
r
em
ain
in
s
u
f
f
icien
t
wh
en
ap
p
lied
in
co
m
p
lex
an
d
d
y
n
am
ic
tr
af
f
ic
p
atter
n
s
.
Fig
u
r
e
5
a
n
aly
s
es
th
e
s
tr
en
g
th
o
f
GC
T
-
I
DS
to
v
a
r
io
u
s
ty
p
es
o
f
attac
k
s
.
Dete
ct
r
ate
o
f
th
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
is
alwa
y
s
h
ig
h
i
n
t
er
m
s
o
f
b
r
u
te
-
f
o
r
ce
,
d
en
ial
-
of
-
s
er
v
ice,
s
ca
n
n
in
g
,
a
n
d
a
p
p
lic
atio
n
-
lay
er
attac
k
s
,
wh
ich
in
d
icate
s
th
e
f
lex
ib
ilit
y
o
f
th
e
p
r
o
p
o
s
ed
f
r
a
m
ewo
r
k
.
I
t
is
in
ter
esti
n
g
to
n
o
te
th
at
lo
w
r
ate
an
d
s
tealth
y
attac
k
s
lik
e
SQL
in
jectio
n
,
X
SS
o
r
s
lo
w
HT
T
P
attac
k
s
h
av
e
a
d
etec
tio
n
r
ate
o
f
o
v
er
9
7
t
h
at
m
o
s
t
tr
ad
itio
n
al
I
DS
m
o
d
els
ca
n
n
o
t
d
etec
t.
T
h
is
s
tr
en
g
th
ca
n
b
e
cr
ed
ited
to
t
em
p
o
r
al
en
r
ic
h
m
en
t
an
d
co
n
tr
asti
v
e
lear
n
in
g
th
at
h
elp
th
e
s
y
s
tem
to
o
b
s
er
v
e
s
lo
w
an
d
s
u
b
tle
ch
a
n
g
es o
f
n
o
r
m
al
b
eh
av
io
r
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
5
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
6
9
8
-
7
0
8
704
Fig
u
r
e
2
.
I
n
tr
u
s
io
n
d
etec
tio
n
m
etr
ics
Fig
u
r
e
3
.
C
lass
-
wis
e
d
etec
tio
n
p
er
f
o
r
m
an
ce
m
et
r
ics
Fig
u
r
e
4
.
Mo
d
el
p
er
f
o
r
m
a
n
ce
co
m
p
ar
is
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
Gra
p
h
-
g
u
i
d
ed
co
n
tr
a
s
tive
tr
a
n
s
fo
r
mer a
r
ch
itectu
r
e
fo
r
r
o
b
u
s
t a
n
d
ex
p
l
a
in
a
b
le
…
(
A
r
ch
a
n
a
J
a
y
a
p
a
l
)
705
Fig
u
r
e
5
.
Dete
ctio
n
r
ates b
y
at
tack
ty
p
e
T
h
e
f
ac
t
t
h
at
th
e
d
is
tr
ib
u
ted
attac
k
s
h
av
e
a
h
ig
h
d
etec
tio
n
r
ate
is
y
et
an
o
th
e
r
in
d
icatio
n
th
at
th
e
tr
an
s
f
o
r
m
er
g
lo
b
al
atten
tio
n
m
ec
h
an
is
m
is
ab
le
to
ca
p
tu
r
e
co
r
r
elate
d
p
atter
n
s
at
th
e
tr
a
f
f
ic
f
lo
ws.
I
n
g
en
er
al,
T
ab
le
1
g
iv
es
th
e
p
er
f
o
r
m
a
n
c
e
o
f
GC
T
-
I
DS
in
co
m
p
ar
is
o
n
with
o
th
er
m
ac
h
in
e
lear
n
in
g
an
d
d
ee
p
lear
n
in
g
m
o
d
els
in
ter
m
s
o
f
tr
ai
n
in
g
tim
e,
test
in
g
,
an
d
m
em
o
r
y
u
s
e.
Alth
o
u
g
h
th
e
tr
a
d
itio
n
al
m
o
d
els
h
av
e
lo
wer
co
m
p
u
tatio
n
al
co
s
t,
th
ey
d
o
n
o
t h
av
e
g
o
o
d
d
etec
tio
n
ac
cu
r
a
cy
.
T
h
e
ad
v
an
ce
d
co
m
p
u
tatio
n
al
o
v
er
h
ea
d
o
f
d
ee
p
lear
n
in
g
m
o
d
els
in
co
m
p
ar
i
s
o
n
to
s
im
p
le
ar
ch
itectu
r
es
an
d
p
r
o
ce
s
s
in
g
in
a
s
eq
u
en
ce
in
cr
ea
s
es
th
e
co
m
p
u
tatio
n
al
o
v
er
h
ea
d
s
.
T
ab
le
2
s
h
o
ws
th
e
p
e
r
f
o
r
m
an
c
e
o
f
GC
T
-
I
DS
in
ter
m
s
o
f
tem
p
o
r
al
g
en
er
aliza
tio
n
in
th
e
v
ar
io
u
s
tim
e
s
eg
m
en
ts
an
d
tr
af
f
ic
lo
ad
s
it
u
atio
n
s
.
T
h
e
m
o
d
el
h
as
a
h
i
g
h
lev
el
o
f
ac
cu
r
ac
y
an
d
r
ec
all
d
u
r
in
g
m
o
r
n
in
g
,
af
ter
n
o
o
n
,
ev
e
n
in
g
,
an
d
n
ig
h
t t
im
es,
wh
ich
is
h
ig
h
ly
tem
p
o
r
a
l stab
le.
T
ab
le
1
.
C
o
m
p
u
tatio
n
al
p
e
r
f
o
r
m
an
ce
ev
alu
atio
n
M
o
d
e
l
Tr
a
i
n
i
n
g
T
i
me
(
s)
Te
st
i
n
g
Ti
m
e
(
s)
M
e
m
o
r
y
U
sa
g
e
(
M
B
)
R
a
n
d
o
m F
o
r
e
s
t
1
1
8
21
6
4
0
G
r
a
d
i
e
n
t
B
o
o
st
i
n
g
1
8
5
29
7
1
0
X
G
B
o
o
st
2
1
0
34
8
2
0
C
N
N
2
6
5
42
9
1
0
LSTM
2
9
5
48
9
8
0
B
i
LST
M
3
2
2
53
1
0
4
0
Tr
a
n
sf
o
r
mer
3
0
8
46
1
0
2
0
A
u
t
o
e
n
c
o
d
e
r
-
C
N
N
2
8
0
44
9
6
0
H
y
b
r
i
d
D
L
3
3
5
58
1
1
0
0
G
C
T
-
I
D
S
3
1
5
49
1
0
5
5
T
ab
le
2
.
T
em
p
o
r
al
g
en
er
aliza
t
io
n
p
er
f
o
r
m
a
n
ce
Ti
me
S
e
g
me
n
t
A
c
c
u
r
a
c
y
(
%)
R
e
c
a
l
l
(
%)
M
o
r
n
i
n
g
9
9
.
2
1
9
9
.
0
5
A
f
t
e
r
n
o
o
n
9
9
.
0
8
9
8
.
9
4
Ev
e
n
i
n
g
9
9
.
1
2
9
8
.
9
7
N
i
g
h
t
9
8
.
9
7
9
8
.
8
3
W
e
e
k
d
a
y
9
9
.
1
5
9
9
.
0
1
W
e
e
k
e
n
d
9
9
.
0
4
9
8
.
8
9
P
e
a
k
Tr
a
f
f
i
c
9
8
.
8
8
9
8
.
7
1
Off
-
P
e
a
k
9
9
.
2
4
9
9
.
1
2
H
i
g
h
L
o
a
d
9
8
.
9
2
9
8
.
7
6
Lo
w
L
o
a
d
9
9
.
3
1
9
9
.
1
8
5.
CO
NCLU
SI
O
N
T
h
e
p
ap
e
r
in
tr
o
d
u
ce
d
a
GC
T
-
I
DS
wh
ich
is
aim
ed
at
o
v
e
r
co
m
in
g
th
e
s
h
o
r
tco
m
i
n
g
s
o
f
t
h
e
cu
r
r
e
n
t
I
DS
m
o
d
els
in
m
an
ag
in
g
co
m
p
lex
,
h
ig
h
-
d
im
en
s
io
n
al
an
d
ch
an
g
in
g
n
etwo
r
k
tr
af
f
ic.
T
h
e
in
ter
ac
tio
n
m
o
d
elin
g
o
f
g
r
a
p
h
-
b
ased
f
ea
tu
r
es,
co
n
tr
asti
v
e
r
ep
r
esen
tatio
n
lear
n
in
g
an
d
s
p
ar
s
e
s
elf
-
atten
tio
n
tr
an
s
f
o
r
m
er
s
ar
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
5
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
6
9
8
-
7
0
8
706
in
co
r
p
o
r
ated
in
to
t
h
e
p
r
o
p
o
s
e
d
f
r
am
ewo
r
k
,
wh
ich
is
ef
f
ec
tiv
ely
ca
p
ab
le
o
f
ca
p
tu
r
in
g
g
lo
b
al
d
ep
en
d
en
cies
as
well
as
wea
k
an
o
m
alies
o
f
b
eh
av
io
r
.
An
e
x
p
er
im
e
n
t
o
n
th
e
C
SE
-
C
I
C
-
I
DS2
0
1
8
d
ataset
p
r
o
v
es
th
at
GC
T
-
I
DS
p
er
f
o
r
m
s
b
etter
th
an
th
e
t
r
ad
itio
n
al
m
ac
h
in
e
lear
n
i
n
g
a
n
d
d
ee
p
lear
n
in
g
-
b
ased
I
DS
m
eth
o
d
s
in
ter
m
s
o
f
d
etec
tio
n
ac
cu
r
ac
y
(
9
9
.
1
2
)
.
T
h
e
f
in
d
in
g
s
af
f
ir
m
th
at
th
e
p
r
o
p
o
s
ed
m
o
d
el
p
r
o
v
id
es
a
s
o
lid
,
s
ca
lab
le
an
d
ex
p
lain
ab
le
s
o
lu
tio
n
t
o
th
e
cu
r
r
en
t
n
etwo
r
k
s
ec
u
r
ity
s
ettin
g
s
.
I
ts
ca
p
ab
ilit
y
to
s
u
s
tain
g
o
o
d
p
er
f
o
r
m
a
n
ce
in
d
etec
tio
n
o
f
v
ar
io
u
s
f
o
r
m
s
o
f
attac
k
s
an
d
tim
e
v
ar
iatio
n
f
ea
tu
r
es
r
en
d
er
it
ap
p
licab
le
in
th
e
r
ea
l
wo
r
ld
.
T
h
e
cu
r
r
en
t
s
tu
d
y
ca
n
b
e
e
x
p
an
d
ed
b
y
f
u
tu
r
e
r
esear
ch
to
in
c
lu
d
e
f
e
d
er
ated
lear
n
in
g
to
p
r
o
m
o
te
d
is
tr
ib
u
ted
in
tr
u
s
io
n
d
etec
tio
n
a
n
d
p
r
iv
ac
y
-
s
en
s
itiv
e
in
tr
u
s
io
n
d
etec
tio
n
.
T
h
e
im
p
lem
e
n
tatio
n
o
f
o
n
lin
e
lear
n
i
n
g
m
ec
h
an
is
m
s
co
u
ld
also
in
cr
ea
s
e
f
lex
ib
ilit
y
to
n
ew
th
r
ea
ts
.
Als
o
,
th
e
f
r
am
ewo
r
k
ca
n
b
e
a
p
p
lied
in
th
e
co
n
tex
t
o
f
m
u
lti
-
m
o
d
al
s
ec
u
r
ity
d
ata
a
n
d
ed
g
e
-
b
ased
d
ep
lo
y
m
en
t
ap
p
licatio
n
s
,
wh
ich
ar
e
p
r
o
s
p
ec
t
s
in
d
ev
elo
p
i
n
g
th
e
n
ex
t
-
g
en
e
r
atio
n
I
DS sy
s
tem
s
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Ar
ch
an
a
J
ay
ap
al
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Kam
alak
k
an
n
a
n
So
m
asu
n
d
ar
am
✓
✓
✓
✓
✓
✓
✓
✓
✓
Ar
u
n
Ku
m
ar
R
am
am
o
o
r
th
y
✓
✓
✓
✓
✓
✓
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
Data
a
v
aila
b
i
lit
y
is
n
o
t
a
p
p
li
ca
b
le
t
o
t
h
is
p
ap
er
as
n
o
n
e
w
d
a
t
a
w
er
e
cr
ea
te
d
o
r
an
al
y
z
e
d
i
n
t
h
is
s
t
u
d
y
.
R
E
FE
R
E
N
C
E
S
[
1
]
I
.
M
a
h
m
o
u
d
i
,
D
.
E
.
B
o
u
b
i
c
h
e
,
S
.
A
t
h
man
i
,
H
.
T
o
r
a
l
-
C
r
u
z
,
a
n
d
F
.
I
.
C
h
a
n
-
P
u
c
,
“
T
o
w
a
r
d
g
e
n
e
r
a
t
i
v
e
A
I
-
b
a
sed
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
sy
st
e
ms f
o
r
t
h
e
I
n
t
e
r
n
e
t
o
f
V
e
h
i
c
l
e
s
(
I
o
V
)
,
”
Fu
t
u
re
I
n
t
e
rn
e
t
,
v
o
l
.
1
7
,
n
o
.
7
,
p
.
3
1
0
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
f
i
1
7
0
7
0
3
1
0
.
[
2
]
B
.
X
i
a
n
g
,
R
.
Z
h
e
n
g
,
K
.
Zh
a
n
g
,
C
.
Li
,
a
n
d
J.
Z
h
e
n
g
,
“
F
F
T
-
R
D
N
e
t
:
A
t
i
me
-
f
r
e
q
u
e
n
c
y
-
d
o
ma
i
n
-
b
a
s
e
d
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
m
o
d
e
l
f
o
r
I
o
T
sec
u
r
i
t
y
,
”
S
e
n
s
o
rs
,
v
o
l
.
2
5
,
n
o
.
1
5
,
p
.
4
5
8
4
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
5
1
5
4
5
8
4
.
[
3
]
M
.
M
.
M
a
h
m
o
u
d
,
Y
.
O
.
Y
o
u
ssef,
a
n
d
A
.
A
.
A
b
d
e
l
-
H
a
mi
d
,
“
X
I
2
S
-
I
D
S
:
A
n
e
x
p
l
a
i
n
a
b
l
e
i
n
t
e
l
l
i
g
e
n
t
2
-
st
a
g
e
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
sy
st
e
m,
”
F
u
t
u
re
I
n
t
e
r
n
e
t
,
v
o
l
.
1
7
,
n
o
.
1
,
p
.
2
5
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
f
i
1
7
0
1
0
0
2
5
.
[
4
]
H
.
P
e
n
g
,
C
.
W
u
,
a
n
d
Y
.
X
i
a
o
,
“
F
D
-
I
D
S
:
F
e
d
e
r
a
t
e
d
l
e
a
r
n
i
n
g
w
i
t
h
k
n
o
w
l
e
d
g
e
d
i
s
t
i
l
l
a
t
i
o
n
f
o
r
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
i
n
n
o
n
-
I
I
D
I
o
T
e
n
v
i
r
o
n
m
e
n
t
s,
”
S
e
n
s
o
rs
,
v
o
l
.
2
5
,
n
o
.
1
4
,
p
.
4
3
0
9
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
s2
5
1
4
4
3
0
9
.
[
5
]
M
.
A
l
R
a
w
a
j
b
e
h
,
A
.
J.
M
a
r
i
a
S
o
o
sai
,
L.
K
.
R
a
m
a
sam
y
,
a
n
d
F
.
K
h
a
n
,
“
Tr
u
st
w
o
r
t
h
y
a
d
a
p
t
i
v
e
A
I
f
o
r
r
e
a
l
-
t
i
m
e
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
i
n
i
n
d
u
st
r
i
a
l
I
o
T
sec
u
r
i
t
y
,
”
I
n
t
e
r
n
e
t
o
f
T
h
i
n
g
s
,
v
o
l
.
6
,
n
o
.
3
,
p
.
5
3
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
i
o
t
6
0
3
0
0
5
3
.
[
6
]
N
.
Te
r
a
w
i
,
H
.
I
.
A
sh
q
a
r
,
O
.
D
a
r
w
i
s
h
,
A
.
A
l
so
b
e
h
,
P
.
Z
a
h
a
r
i
e
v
,
a
n
d
Y
.
Ta
sh
t
o
u
s
h
,
“
En
h
a
n
c
e
d
d
e
t
e
c
t
i
o
n
o
f
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
sy
st
e
m
i
n
c
l
o
u
d
n
e
t
w
o
r
k
s
u
si
n
g
t
i
me
-
a
w
a
r
e
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s,
”
C
o
m
p
u
t
e
rs
,
v
o
l
.
1
4
,
n
o
.
7
,
p
.
2
8
2
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
c
o
m
p
u
t
e
r
s
1
4
0
7
0
2
8
2
.
[
7
]
F
.
A
l
s
e
r
h
a
n
i
,
“
I
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
a
n
d
r
e
a
l
-
t
i
me
a
d
a
p
t
i
v
e
s
e
c
u
r
i
t
y
i
n
me
d
i
c
a
l
I
o
T
u
s
i
n
g
a
c
y
b
e
r
-
p
h
y
s
i
c
a
l
sy
s
t
e
m
d
e
si
g
n
,
”
S
e
n
s
o
rs
,
v
o
l
.
2
5
,
n
o
.
1
5
,
p
.
4
7
2
0
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
5
1
5
4
7
2
0
.
[
8
]
E.
A
l
a
l
w
a
n
y
,
B
.
A
l
sh
a
r
i
f
,
Y
.
A
l
o
t
a
i
b
i
,
A
.
A
l
f
a
h
a
i
d
,
I
.
M
a
h
g
o
u
b
,
a
n
d
M
.
I
l
y
a
s,
“
S
t
a
c
k
i
n
g
e
n
sem
b
l
e
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
r
e
a
l
-
t
i
me
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
i
n
I
o
M
T
e
n
v
i
r
o
n
m
e
n
t
s,”
S
e
n
s
o
rs
,
v
o
l
.
2
5
,
n
o
.
3
,
p
.
6
2
4
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
s2
5
0
3
0
6
2
4
.
[
9
]
A
.
A
l
a
b
d
u
l
a
t
i
f
,
“
A
n
o
v
e
l
e
n
s
e
m
b
l
e
o
f
d
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
f
o
r
c
y
b
e
r
sec
u
r
i
t
y
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
w
i
t
h
e
x
p
l
a
i
n
a
b
l
e
a
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
,
”
A
p
p
l
i
e
d
S
c
i
e
n
c
e
s (S
w
i
t
z
e
rl
a
n
d
)
,
v
o
l
.
1
5
,
n
o
.
1
4
,
p
.
7
9
8
4
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
a
p
p
1
5
1
4
7
9
8
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
Gra
p
h
-
g
u
i
d
ed
co
n
tr
a
s
tive
tr
a
n
s
fo
r
mer a
r
ch
itectu
r
e
fo
r
r
o
b
u
s
t a
n
d
ex
p
l
a
in
a
b
le
…
(
A
r
ch
a
n
a
J
a
y
a
p
a
l
)
707
[
1
0
]
M
.
B
.
B
a
n
k
ó
e
t
a
l
.
,
“
A
d
v
a
n
c
e
me
n
t
s
i
n
m
a
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:
R
e
se
a
r
c
h
t
r
e
n
d
s
a
n
d
c
h
a
l
l
e
n
g
e
s,
”
Al
g
o
r
i
t
h
m
s
,
v
o
l
.
1
8
,
n
o
.
4
,
p
.
2
0
9
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
a
1
8
0
4
0
2
0
9
.
[
1
1
]
N
.
S
a
v
a
n
o
v
i
ć
e
t
a
l
.
,
“
I
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
i
n
h
e
a
l
t
h
c
a
r
e
4
.
0
I
n
t
e
r
n
e
t
o
f
T
h
i
n
g
s
sy
s
t
e
ms
v
i
a
met
a
h
e
u
r
i
st
i
c
s
o
p
t
i
m
i
z
e
d
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
,
”
S
u
st
a
i
n
a
b
i
l
i
t
y
(
S
w
i
t
z
e
rl
a
n
d
)
,
v
o
l
.
1
5
,
n
o
.
1
6
,
p
.
1
2
5
6
3
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
s
u
1
5
1
6
1
2
5
6
3
.
[
1
2
]
D
.
M
u
s
l
e
h
,
M
.
A
l
o
t
a
i
b
i
,
F
.
A
l
h
a
i
d
a
r
i
,
A
.
R
a
h
m
a
n
,
a
n
d
R
.
M
.
M
o
h
a
mm
a
d
,
“
I
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
sy
s
t
e
m
u
si
n
g
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
w
i
t
h
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms
i
n
I
o
T,
”
J
o
u
r
n
a
l
o
f
S
e
n
s
o
r
a
n
d
Ac
t
u
a
t
o
r
N
e
t
w
o
rks
,
v
o
l
.
1
2
,
n
o
.
2
,
M
a
r
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
j
s
a
n
1
2
0
2
0
0
2
9
.
[
1
3
]
A
.
K
u
m
a
r
a
n
d
J.
A
.
G
u
t
i
e
r
r
e
z
,
“
I
mp
a
c
t
o
f
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
o
n
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
t
h
e
p
r
o
t
e
c
t
i
o
n
o
f
c
r
i
t
i
c
a
l
i
n
f
r
a
s
t
r
u
c
t
u
r
e
,
”
I
n
f
o
rm
a
t
i
o
n
(
S
w
i
t
zer
l
a
n
d
)
,
v
o
l
.
1
6
,
n
o
.
7
,
p
.
5
1
5
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
i
n
f
o
1
6
0
7
0
5
1
5
.
[
1
4
]
H
.
A
t
t
o
u
e
t
a
l
.
,
“
T
o
w
a
r
d
s
a
n
i
n
t
e
l
l
i
g
e
n
t
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
sy
s
t
e
m
t
o
d
e
t
e
c
t
mal
i
c
i
o
u
s
a
c
t
i
v
i
t
i
e
s
i
n
c
l
o
u
d
c
o
m
p
u
t
i
n
g
,
”
Ap
p
l
i
e
d
S
c
i
e
n
c
e
s
(
S
w
i
t
zer
l
a
n
d
)
,
v
o
l
.
1
3
,
n
o
.
1
7
,
p
.
9
5
8
8
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
a
p
p
1
3
1
7
9
5
8
8
.
[
1
5
]
Z.
W
a
n
g
,
J
.
Li
,
S
.
Y
a
n
g
,
X
.
L
u
o
,
D
.
L
i
,
a
n
d
S
.
M
a
h
m
o
o
d
i
,
“
A
l
i
g
h
t
w
e
i
g
h
t
I
o
T
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
mo
d
e
l
b
a
s
e
d
o
n
i
mp
r
o
v
e
d
B
ER
T
-
of
-
T
h
e
se
u
s
,
”
Ex
p
e
r
t
S
y
st
e
m
s
w
i
t
h
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
2
3
8
,
p
.
1
2
2
0
4
5
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
sw
a
.
2
0
2
3
.
1
2
2
0
4
5
.
[
1
6
]
Y
.
C
.
W
a
n
g
,
Y
.
C
.
H
o
u
n
g
,
H
.
X
.
C
h
e
n
,
a
n
d
S
.
M
.
Ts
e
n
g
,
“
N
e
t
w
o
r
k
a
n
o
ma
l
y
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
b
a
se
d
o
n
d
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
,
”
S
e
n
so
rs
,
v
o
l
.
2
3
,
n
o
.
4
,
p
.
2
1
7
1
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
3
0
4
2
1
7
1
.
[
1
7
]
H
.
A
t
t
o
u
,
A
.
G
u
e
z
z
a
z
,
S
.
B
e
n
k
i
r
a
n
e
,
M
.
A
z
r
o
u
r
,
a
n
d
Y
.
F
a
r
h
a
o
u
i
,
“
C
l
o
u
d
-
b
a
s
e
d
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
a
p
p
r
o
a
c
h
u
s
i
n
g
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s,
”
B
i
g
D
a
t
a
M
i
n
i
n
g
a
n
d
An
a
l
y
t
i
c
s
,
v
o
l
.
6
,
n
o
.
3
,
p
p
.
3
1
1
–
3
2
0
,
S
e
p
.
2
0
2
3
,
d
o
i
:
1
0
.
2
6
5
9
9
/
B
D
M
A
.
2
0
2
2
.
9
0
2
0
0
3
8
.
[
1
8
]
M
.
B
h
a
v
sar,
K
.
R
o
y
,
J
.
K
e
l
l
y
,
a
n
d
O
.
O
l
u
s
o
l
a
,
“
A
n
o
ma
l
y
-
b
a
s
e
d
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
s
y
st
e
m
f
o
r
I
o
T
a
p
p
l
i
c
a
t
i
o
n
,
”
D
i
sco
v
e
r
I
n
t
e
r
n
e
t
o
f
T
h
i
n
g
s
,
v
o
l
.
3
,
n
o
.
1
,
p
.
5
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s4
3
9
2
6
-
023
-
0
0
0
3
4
-
5.
[
1
9
]
S
.
S
h
i
,
D
.
H
a
n
,
a
n
d
M
.
C
u
i
,
“
A
mu
l
t
i
mo
d
a
l
h
y
b
r
i
d
p
a
r
a
l
l
e
l
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
m
o
d
e
l
,
”
C
o
n
n
e
c
t
i
o
n
S
c
i
e
n
c
e
,
v
o
l
.
3
5
,
n
o
.
1
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
8
0
/
0
9
5
4
0
0
9
1
.
2
0
2
3
.
2
2
2
7
7
8
0
.
[
2
0
]
M
.
S
.
K
o
r
i
u
m
,
M
.
S
a
b
e
r
,
A
.
B
e
a
t
t
i
e
,
A
.
N
a
r
a
y
a
n
a
n
,
S
.
S
a
h
o
o
,
a
n
d
P
.
H
.
J.
N
a
r
d
e
l
l
i
,
“
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
c
y
b
e
r
a
t
t
a
c
k
s
i
n
t
h
e
I
n
t
e
r
n
e
t
o
f
V
e
h
i
c
l
e
s
e
n
v
i
r
o
n
m
e
n
t
,
”
A
d
H
o
c
N
e
t
w
o
rks
,
v
o
l
.
1
5
3
,
p
.
1
0
3
3
3
0
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
d
h
o
c
.
2
0
2
3
.
1
0
3
3
3
0
.
[
2
1
]
S
o
l
a
r
m
a
i
n
f
r
a
m
e
,
“
I
D
S
i
n
t
r
u
s
i
o
n
C
S
V
d
a
t
a
se
t
,
”
K
a
g
g
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/
s
o
l
a
r
mai
n
f
r
a
m
e
/
i
d
s
-
i
n
t
r
u
s
i
o
n
-
c
s
v
(
a
c
c
e
s
se
d
J
u
n
.
1
8
,
2
0
2
5
)
.
[
2
2
]
A
.
P
.
M
u
n
i
y
a
n
d
i
e
t
a
l
.
,
“
I
n
t
e
l
l
i
g
e
n
t
se
c
u
r
i
t
y
sy
s
t
e
m
f
o
r
p
r
e
v
e
n
t
i
n
g
D
D
o
S
a
t
t
a
c
k
s
f
o
r
6
G
e
n
a
b
l
e
d
W
B
S
N
u
s
i
n
g
i
mp
r
o
v
e
g
r
e
y
w
o
l
f
o
p
t
i
m
i
z
a
t
i
o
n
,
”
I
EEE
T
r
a
n
sa
c
t
i
o
n
s
o
n
C
o
n
s
u
m
e
r
El
e
c
t
r
o
n
i
c
s
,
v
o
l
.
7
0
,
n
o
.
3
,
p
p
.
5
7
7
5
–
5
7
8
2
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
T
C
E.
2
0
2
4
.
3
4
1
6
5
4
9
.
[
2
3
]
H
.
M
.
B
e
l
a
c
h
e
w
,
M
.
Y
.
B
e
y
e
n
e
,
A
.
B
.
D
e
st
a
,
B
.
T.
A
l
e
mu
,
S
.
S
.
M
u
s
a
,
a
n
d
A
.
J.
M
u
h
a
m
me
d
,
“
D
e
s
i
g
n
a
r
o
b
u
st
D
D
o
S
a
t
t
a
c
k
d
e
t
e
c
t
i
o
n
a
n
d
m
i
t
i
g
a
t
i
o
n
s
c
h
e
me
i
n
S
D
N
-
e
d
g
e
-
I
o
T
b
y
l
e
v
e
r
a
g
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
,
”
I
E
EE
Ac
c
e
ss
,
v
o
l
.
1
3
,
p
p
.
1
0
1
9
4
–
1
0
2
1
4
,
2
0
2
5
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
5
.
3
5
2
6
6
9
2
.
[
2
4
]
Y
.
M
i
r
s
k
y
,
T.
D
o
i
t
s
h
m
a
n
,
Y
.
El
o
v
i
c
i
,
a
n
d
A
.
S
h
a
b
t
a
i
,
“
K
i
t
s
u
n
e
:
A
n
e
n
semb
l
e
o
f
a
u
t
o
e
n
c
o
d
e
r
s
f
o
r
o
n
l
i
n
e
n
e
t
w
o
r
k
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
,
”
i
n
2
5
t
h
A
n
n
u
a
l
N
e
t
w
o
rk
a
n
d
D
i
st
r
i
b
u
t
e
d
S
y
st
e
m
S
e
c
u
r
i
t
y
S
y
m
p
o
si
u
m
,
N
D
S
S
2
0
1
8
,
2
0
1
8
.
d
o
i
:
1
0
.
1
4
7
2
2
/
n
d
ss.
2
0
1
8
.
2
3
2
0
4
.
[
2
5
]
Y
.
F
u
,
D
.
Li
u
,
J.
C
h
e
n
,
a
n
d
L.
H
e
,
“
S
e
c
r
e
t
a
r
y
b
i
r
d
o
p
t
i
m
i
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m
:
a
n
e
w
m
e
t
a
h
e
u
r
i
st
i
c
f
o
r
s
o
l
v
i
n
g
g
l
o
b
a
l
o
p
t
i
m
i
z
a
t
i
o
n
p
r
o
b
l
e
ms,
”
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
R
e
v
i
e
w
,
v
o
l
.
5
7
,
n
o
.
5
,
p
p
.
1
–
1
0
2
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
0
4
6
2
-
0
2
4
-
1
0
7
2
9
-
y.
[
2
6
]
Y
.
E.
K
i
m
,
Y
.
S
.
K
i
m,
a
n
d
H
.
K
i
m,
“
Ef
f
e
c
t
i
v
e
f
e
a
t
u
r
e
se
l
e
c
t
i
o
n
m
e
t
h
o
d
s
t
o
d
e
t
e
c
t
I
o
T
D
D
o
S
a
t
t
a
c
k
i
n
5
G
c
o
r
e
n
e
t
w
o
r
k
,
”
S
e
n
s
o
rs
,
v
o
l
.
2
2
,
n
o
.
1
0
,
p
.
3
8
1
9
,
M
a
y
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
s2
2
1
0
3
8
1
9
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Ar
c
h
a
n
a
J
a
y
a
p
a
l
re
se
a
rc
h
sc
h
o
lar
a
t
Ve
ls
I
n
stit
u
te
o
f
S
c
ien
c
e
,
Tec
h
n
o
lo
g
y
,
a
n
d
Ad
v
a
n
c
e
d
S
tu
d
ies
,
C
h
e
n
n
a
i,
Ta
m
il
Na
d
u
,
In
d
ia.
S
h
e
h
a
s
e
x
p
e
rti
se
in
d
e
e
p
lea
rn
in
g
i
n
tru
si
o
n
d
e
tec
ti
o
n
s
y
ste
m
s.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
a
rc
h
a
n
a
a
rti
stry
a
c
a
d
e
m
y
@g
m
a
il
.
c
o
m
.
K
a
m
a
l
a
k
k
a
n
n
a
n
S
o
m
a
sun
d
a
r
a
m
re
c
e
iv
e
d
h
is
M
.
S
c
.
c
o
m
p
u
t
e
r
sc
ien
c
e
d
e
g
re
e
fro
m
Bh
a
ra
th
id
a
sa
n
Un
i
v
e
rsity
,
M
.
P
h
i
l.
c
o
m
p
u
ter
sc
ien
c
e
d
e
g
re
e
fro
m
P
e
riy
a
r
Un
iv
e
rsit
y
,
a
n
d
P
h
.
D.
i
n
c
o
m
p
u
ter
sc
ien
c
e
fro
m
Ve
ls
In
stit
u
te
o
f
S
c
ien
c
e
,
Tec
h
n
o
lo
g
y
a
n
d
Ad
v
a
n
c
e
d
S
t
u
d
ies
(VIST
AS),
Tam
il
Na
d
u
,
I
n
d
ia.
He
is
c
u
rre
n
tl
y
wo
r
k
in
g
a
s
a
p
ro
fe
ss
o
r
in
t
h
e
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
Ap
p
li
c
a
ti
o
n
s,
S
c
h
o
o
l
o
f
Co
m
p
u
t
in
g
S
c
ien
c
e
s,
Ve
ls
In
stit
u
te
o
f
S
c
ien
c
e
,
Tec
h
n
o
lo
g
y
a
n
d
A
d
v
a
n
c
e
d
S
t
u
d
ies
(VIST
A
S
),
C
h
e
n
n
a
i,
Tam
il
Na
d
u
,
I
n
d
i
a
,
wh
ich
is
a
we
ll
-
k
n
o
w
n
u
n
i
v
e
rsity
.
He
h
a
s
2
1
y
e
a
rs
o
f
tea
c
h
in
g
e
x
p
e
rien
c
e
a
t
b
o
t
h
th
e
UG
a
n
d
P
G
lev
e
ls.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
b
ig
d
a
ta
a
n
a
ly
t
ics
,
c
lo
u
d
c
o
m
p
u
ti
n
g
,
a
n
d
b
l
o
c
k
c
h
a
in
tec
h
n
o
l
o
g
y
.
He
h
a
s
p
ro
d
u
c
e
d
th
ir
tee
n
P
h
.
D.
re
se
a
rc
h
sc
h
o
lars
.
He
h
a
s
p
u
b
li
sh
e
d
m
o
re
th
a
n
9
0
re
se
a
rc
h
a
rti
c
les
i
n
v
a
rio
u
s
in
ter
n
a
ti
o
n
a
l
jo
u
rn
a
ls
su
c
h
a
s
S
CI,
WOS
,
S
c
o
p
u
s,
a
n
d
UG
C
-
re
fe
rre
d
jo
u
r
n
a
ls.
He
h
a
s
p
u
b
li
sh
e
d
1
0
p
a
te
n
ts.
He
is
a
re
v
iew
e
r
fo
r
v
a
rio
u
s
jo
u
rn
a
ls.
He
is
se
rv
in
g
a
s
a
c
o
-
p
r
in
c
ip
a
l
in
v
e
stig
a
t
o
r
in
a
De
p
a
rtme
n
t
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
(DS
T)
sp
o
n
s
o
re
d
p
r
o
jec
t.
He
re
c
e
iv
e
d
th
e
b
e
st
y
o
u
n
g
sc
ien
ti
s
t
a
wa
rd
,
b
e
st
fa
c
u
lt
y
a
wa
rd
,
a
n
d
b
e
st
re
se
a
rc
h
su
p
e
rv
iso
r
a
wa
rd
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
k
a
n
n
a
n
.
sc
s@
v
istas
.
a
c
.
in
.
Evaluation Warning : The document was created with Spire.PDF for Python.