I
nte
rna
t
io
na
l J
o
urna
l o
f
I
nfo
rm
a
t
ics a
nd
Co
m
m
un
ica
t
io
n T
ec
hn
o
lo
g
y
(
I
J
-
I
CT
)
Vo
l.
1
5
,
No
.
4
,
D
e
c
em
b
er
20
2
6
,
p
p
.
1
7
6
8
~
1
7
7
6
I
SS
N:
2252
-
8
7
7
6
,
DOI
:
1
0
.
1
1
5
9
1
/iji
ct
.
v
1
5
i
4
.
pp
1
7
6
8
-
1
7
7
6
1768
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ict.
ia
esco
r
e.
co
m
Deep
trans
former
architec
tures
f
o
r
real
-
time
e
-
pa
y
m
ent
f
ra
ud
detec
tion unde
r cl
a
ss
imba
la
nce
Ria
d L
o
uk
ili
1
,
F
a
y
ça
l Mes
s
a
o
ud
i
2
,
Ra
o
uy
a
E
l Y
o
ub
i
1
,
M
a
na
l Lo
uk
ili
2
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
,
N
a
t
i
o
n
a
l
S
c
h
o
o
l
o
f
A
p
p
l
i
e
d
S
c
i
e
n
c
e
s,
S
i
d
i
M
o
h
a
me
d
B
e
n
A
b
d
e
l
l
a
h
U
n
i
v
e
r
si
t
y
,
F
e
z
,
M
o
r
o
c
c
o
2
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
,
N
a
t
i
o
n
a
l
S
c
h
o
o
l
o
f
B
u
s
i
n
e
ss a
n
d
M
a
n
a
g
e
m
e
n
t
,
S
i
d
i
M
o
h
a
m
e
d
B
e
n
A
b
d
e
l
l
a
h
U
n
i
v
e
r
s
i
t
y
,
F
e
z
,
M
o
r
o
c
c
o
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
n
1
,
2
0
2
4
R
ev
is
ed
J
u
n
3
,
2
0
2
6
Acc
ep
ted
J
u
l 1
9
,
2
0
2
6
E
-
c
o
m
m
e
rc
e
h
a
s
b
e
c
o
m
e
a
ra
p
id
ly
e
x
p
a
n
d
i
n
g
c
h
a
n
n
e
l,
lea
d
i
n
g
to
m
o
re
e
-
c
o
m
m
e
rc
e
p
a
y
m
e
n
t
tran
sa
c
ti
o
n
s
a
n
d
a
n
i
n
c
re
a
se
d
th
re
a
t
fo
r
a
d
v
a
n
c
e
d
fra
u
d
u
le
n
t
a
c
ti
v
it
ies
.
Ne
w
fra
u
d
s
in
c
lu
d
e
c
ro
ss
-
c
h
a
n
n
e
l
a
tt
a
c
k
s,
a
d
v
e
rsa
rial
b
e
h
a
v
i
o
r
a
n
d
sy
n
th
e
ti
c
i
d
e
n
ti
ti
e
s,
wh
ich
a
re
c
h
a
ll
e
n
g
in
g
t
o
ru
le
-
b
a
se
d
d
e
tec
ti
o
n
.
In
t
h
is
p
a
p
e
r,
a
d
e
e
p
lea
rn
i
n
g
fra
m
e
wo
rk
fo
r
d
e
tec
ti
n
g
e
-
p
a
y
m
e
n
t
fra
u
d
i
n
re
a
l
-
ti
m
e
is
p
ro
p
o
se
d
.
Th
e
m
o
d
e
l
in
teg
ra
tes
se
lf
-
a
tt
e
n
ti
o
n
,
e
x
p
lain
a
b
le
AI,
c
las
s
b
a
lan
c
in
g
u
sin
g
G
AN
s,
a
n
d
a
tran
sfo
r
m
e
r
-
b
a
se
d
a
rc
h
it
e
c
tu
re
to
e
n
a
b
le
tran
sp
a
re
n
t
d
e
c
isio
n
-
m
a
k
i
n
g
.
T
h
e
so
lu
ti
o
n
tac
k
les
m
a
jo
r
iss
u
e
s in
fra
u
d
d
e
tec
ti
o
n
,
s
u
c
h
a
s c
las
s imb
a
lan
c
e
,
e
v
o
l
v
in
g
t
ra
n
sa
c
ti
o
n
tren
d
s,
a
n
d
t
h
e
re
q
u
irem
e
n
t
fo
r
in
terp
re
tab
il
it
y
in
c
o
m
p
l
ian
c
e
with
re
g
u
latio
n
s
li
k
e
th
e
EU
AI
Ac
t.
Ex
p
e
rime
n
ts
o
n
b
e
n
c
h
m
a
rk
d
a
tas
e
ts
d
e
m
o
n
stra
te
g
o
o
d
re
su
lt
s,
wit
h
a
n
a
c
c
u
ra
c
y
o
f
9
9
.
8
2
%
,
a
n
d
a
n
AU
C
-
ROC
o
f
9
9
.
5
6
%
.
Ba
se
d
o
n
th
e
re
su
lt
s,
t
h
e
u
se
o
f
tra
n
sfo
rm
e
r
-
b
a
se
d
m
o
d
e
ls
with
t
h
e
h
e
lp
o
f
s
y
n
t
h
e
ti
c
d
a
ta
g
e
n
e
ra
t
io
n
a
n
d
e
x
p
lain
a
b
i
li
ty
c
a
n
e
n
h
a
n
c
e
th
e
re
li
a
b
il
it
y
a
n
d
tran
sp
a
re
n
c
y
o
f
e
-
p
a
y
m
e
n
t
fra
u
d
d
e
tec
ti
o
n
sy
ste
m
s.
K
ey
w
o
r
d
s
:
C
las
s
im
b
alan
c
e
Dee
p
l
ea
r
n
in
g
E
-
co
m
m
er
ce
s
ec
u
r
ity
E
x
p
lain
ab
le
AI
Fed
er
ated
lear
n
in
g
R
ea
l
-
tim
e
f
r
au
d
d
etec
tio
n
T
r
an
s
f
o
r
m
e
r
ar
ch
itectu
r
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
:
Ma
n
al
L
o
u
k
ili
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
,
Natio
n
al
Sch
o
o
l o
f
B
u
s
in
ess
an
d
Ma
n
ag
em
en
t
Sid
i M
o
h
am
ed
B
en
Ab
d
ella
h
Un
iv
er
s
ity
Fez,
Mo
r
o
cc
o
E
m
ail:
m
an
al.
lo
u
k
ili@
u
s
m
b
a.
ac
.
m
a
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
e
-
c
o
m
m
er
ce
d
o
m
ai
n
is
a
p
ar
t
o
f
t
h
e
in
ter
n
atio
n
al
e
c
o
n
o
m
y
an
d
h
as
r
ev
o
l
u
tio
n
ize
d
th
e
way
g
o
o
d
s
an
d
s
er
v
ices a
r
e
p
u
r
ch
a
s
ed
b
y
p
eo
p
le
[
1
]
.
Acc
o
r
d
in
g
t
o
th
is
g
r
o
wth
,
t
h
e
elec
tr
o
n
ic
p
ay
m
en
t sy
s
tem
s
ar
e
n
o
w
ess
en
tial
to
c
o
n
d
u
ct
b
u
s
in
ess
o
n
lin
e
[
2
]
.
T
h
is
ex
te
n
s
io
n
h
as
a
d
d
itio
n
ally
g
r
o
wn
th
e
am
o
u
n
t
o
f
ex
p
o
s
u
r
e
to
f
r
au
d
,
p
ar
ticu
lar
ly
with
p
a
y
m
en
t
s
y
s
tem
s
th
at
in
co
r
p
o
r
a
te
m
ass
iv
e
am
o
u
n
ts
o
f
f
in
an
c
ial
in
f
o
o
n
a
d
aily
b
asis
[
3
]
.
Acc
o
r
d
in
g
to
a
r
e
p
o
r
t
b
y
I
T
D
an
d
th
e
Natio
n
al
R
etail
Fo
u
n
d
atio
n
,
g
lo
b
al
e
-
c
o
m
m
er
ce
s
ales
will
k
ee
p
r
is
in
g
,
an
d
b
y
2
0
2
5
,
th
e
to
tal
am
o
u
n
t
o
f
s
ales
is
p
r
o
j
ec
ted
to
h
it
m
o
r
e
th
an
$
7
.
5
tr
illi
o
n
an
d
co
u
ld
h
it
o
v
er
$
9
tr
illi
o
n
b
y
2
0
2
7
.
Me
an
wh
ile,
th
e
lo
s
s
f
r
o
m
elec
tr
o
n
ic
p
ay
m
en
t
f
r
a
u
d
is
esti
m
a
ted
to
r
ea
ch
ar
o
u
n
d
USD
4
8
b
illi
o
n
g
lo
b
ally
b
y
2
0
2
5
[
4
]
a
n
d
f
u
r
th
er
escalate
with
th
e
ex
p
an
s
io
n
o
f
elec
tr
o
n
ic
p
ay
m
e
n
t
in
th
e
co
m
in
g
y
ea
r
s
.
T
h
e
u
s
e
o
f
AI
-
d
r
iv
en
s
er
v
ices
in
e
-
co
m
m
er
ce
an
d
th
e
g
r
o
win
g
in
te
r
n
et
u
s
e
an
d
s
m
ar
t
p
h
o
n
e
p
en
etr
atio
n
h
av
e
f
a
cilitated
th
e
ex
p
a
n
s
io
n
o
f
th
e
ec
o
m
m
e
r
ce
m
ar
k
et
[
5
]
,
[
6
]
.
T
h
e
p
ay
m
en
t
m
eth
o
d
s
also
ex
p
an
d
with
th
e
u
s
e
o
f
cr
ed
it/d
eb
it
ca
r
d
s
,
d
ig
ital
wallets,
d
ef
er
r
ed
p
a
y
m
en
ts
,
cr
y
p
to
c
u
r
r
en
c
y
p
ay
m
en
ts
,
an
d
in
s
tan
t
b
an
k
tr
an
s
f
er
s
[
7
]
.
T
h
ese
o
p
tio
n
s
m
a
k
e
it
ea
s
ier
f
o
r
th
e
u
s
er
,
b
u
t
al
s
o
p
r
o
v
id
e
m
o
r
e
o
p
p
o
r
tu
n
ities
f
o
r
an
attac
k
er
to
e
x
p
lo
it a
s
y
s
tem
.
C
y
b
er
cr
im
in
als
a
r
e
i
n
ter
ested
in
th
e
e
-
p
a
y
m
en
t
s
y
s
tem
s
,
d
u
e
to
th
e
ad
o
p
tio
n
o
f
a
f
ast,
p
ar
tiall
y
an
o
n
y
m
o
u
s
an
d
u
n
p
r
e
d
ictab
l
e
p
ay
m
e
n
t
s
y
s
tem
,
wh
ich
is
n
o
t
ea
s
ily
v
er
i
f
iab
le
in
r
ea
l
-
ti
m
e
[
8
]
,
[
9
]
.
Sto
len
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
Dee
p
tr
a
n
s
fo
r
mer a
r
ch
itectu
r
e
s
fo
r
r
ea
l
-
time
e
-
p
a
yme
n
t
fr
a
u
d
d
etec
tio
n
…
(
R
ia
d
Lo
u
kili
)
1769
ca
r
d
s
an
d
p
h
is
h
in
g
attac
k
s
ar
e
n
o
t
all
o
f
f
r
au
d
.
So
m
e
o
f
t
h
e
o
n
g
o
in
g
s
ch
em
es
ar
e
s
y
n
t
h
etic
id
en
tity
f
r
au
d
,
cr
ed
en
tial
s
tu
f
f
in
g
,
m
o
b
ile
p
a
y
m
en
t
API
attac
k
s
,
an
d
m
o
n
e
y
lau
n
d
e
r
in
g
v
ia
cr
y
p
to
cu
r
r
en
cy
m
ix
in
g
s
er
v
ices
[
1
0
]
.
W
h
at
ar
e
o
th
e
r
im
p
ac
ts
o
f
th
ese
ac
tiv
ities
as
well
as
a
d
ir
ec
t
f
in
an
cial
im
p
ac
t?
Fra
u
d
h
as
th
e
p
o
ten
tial
to
d
am
ag
e
cu
s
to
m
er
tr
u
s
t,
h
ar
m
t
h
e
r
ep
u
tatio
n
o
f
o
n
lin
e
p
latf
o
r
m
s
,
an
d
r
aise
co
m
p
lian
ce
co
n
ce
r
n
s
,
esp
ec
ially
in
th
e
co
n
tex
t
o
f
th
e
E
u
r
o
p
ea
n
Un
io
n
AI
Act,
wh
ic
h
p
lace
s
a
g
r
ea
te
r
em
p
h
asis
o
n
tr
a
n
s
p
ar
en
cy
a
n
d
h
u
m
a
n
o
v
er
s
ig
h
t in
f
in
an
cial
d
ec
is
io
n
-
m
ak
in
g
p
r
o
ce
s
s
es [
1
1
]
.
T
h
er
ef
o
r
e,
th
e
p
r
e
v
en
tio
n
o
f
f
r
au
d
is
a
p
a
r
t
o
f
e
-
p
a
y
m
en
t
s
ec
u
r
ity
[
1
2
]
.
W
h
ile
r
u
le
-
b
as
ed
s
y
s
tem
s
ca
n
s
till
b
e
ef
f
ec
tiv
e,
th
ese
ar
e
to
o
s
tatic
to
b
e
r
esp
o
n
s
iv
e
to
f
r
au
d
s
ch
em
e
c
h
an
g
es.
W
h
ile
f
in
an
cial
AI
s
y
s
tem
s
ca
n
b
e
s
u
b
ject
to
a
l
o
t
o
f
f
alse
p
o
s
itiv
es,
m
ay
n
o
t
b
e
awa
r
e
o
f
th
e
latest
attac
k
s
tr
ateg
ies,
an
d
m
ay
o
n
ly
ex
p
lain
a
lim
ited
p
ar
t
o
f
th
e
d
ec
is
io
n
[
1
3
]
.
T
h
ese
r
eq
u
ir
em
en
ts
ar
e
in
c
r
ea
s
in
g
ly
c
r
u
cial
in
a
r
e
g
u
lat
o
r
y
en
v
ir
o
n
m
en
t.
T
h
e
E
U
AI
Act
(
to
b
e
im
p
lem
en
ted
f
r
o
m
A
u
g
u
s
t
2
0
2
4
an
d
f
u
lly
in
f
o
r
c
e
b
y
Au
g
u
s
t
2
0
2
6
)
f
u
r
th
er
s
tr
en
g
th
en
s
th
e
r
eq
u
ir
e
m
en
ts
to
u
s
e
AI
in
s
en
s
itiv
e
f
i
n
an
cial
s
itu
atio
n
s
[
1
4
]
.
As
f
r
au
d
s
ch
em
es
b
ec
o
m
e
m
o
r
e
co
m
p
le
x
,
an
d
c
h
an
g
e
o
v
e
r
tim
e,
d
ee
p
le
ar
n
i
n
g
ca
n
b
e
a
n
ef
f
ec
tiv
e
to
o
l
to
co
m
b
at
th
is
d
if
f
icu
lty
[
1
5
]
.
T
r
a
n
s
f
o
r
m
er
m
o
d
els
ca
n
ca
p
tu
r
e
th
e
lo
n
g
-
ter
m
d
e
p
en
d
en
ce
b
etwe
en
th
e
s
eq
u
en
ce
o
f
tr
an
s
ac
tio
n
s
,
an
d
d
etec
t
s
u
s
p
icio
u
s
tr
an
s
ac
tio
n
s
th
at
m
ay
n
o
t
ap
p
ea
r
in
th
e
in
d
iv
id
u
al
tr
an
s
ac
tio
n
s
[
1
6
]
.
Gr
ap
h
Neu
r
al
Netwo
r
k
s
ca
n
m
o
d
el
th
e
r
elatio
n
s
b
etwe
en
u
s
er
s
,
ac
co
u
n
ts
,
m
er
c
h
an
t
s
,
an
d
tr
a
n
s
ac
tio
n
s
,
an
d
tr
a
n
s
ac
tio
n
s
,
m
ak
in
g
th
e
m
u
s
ef
u
l
i
n
d
etec
tin
g
co
o
r
d
i
n
ated
f
r
a
u
d
[
1
7
]
.
An
o
th
er
m
eth
o
d
is
f
e
d
er
ated
lear
n
in
g
,
wh
e
r
e
an
y
i
n
s
titu
tio
n
ca
n
tr
ain
th
eir
o
wn
m
o
d
el
wi
th
o
u
t
s
h
ar
in
g
th
e
d
ata
ab
o
u
t
its
cu
s
to
m
er
s
[
1
8
]
.
Me
an
wh
ile,
s
ev
er
al
tech
n
iq
u
e
s
h
av
e
em
er
g
ed
f
o
r
g
ai
n
in
g
a
b
etter
u
n
d
e
r
s
tan
d
in
g
o
f
d
ec
is
io
n
s
u
s
ed
f
o
r
f
r
au
d
d
etec
tio
n
s
u
ch
as SHAP,
L
I
ME
,
an
d
atten
tio
n
-
b
ased
in
te
r
p
r
e
tatio
n
[
1
9
]
.
E
-
p
ay
m
e
n
t
f
r
au
d
in
e
-
co
m
m
e
r
ce
is
o
n
e
ar
ea
wh
er
e
d
ee
p
lear
n
in
g
ca
n
b
e
a
p
p
lied
f
o
r
f
r
au
d
d
etec
tio
n
.
W
e
h
av
e
p
r
esen
ted
a
m
o
d
el
b
ased
o
n
a
tr
an
s
f
o
r
m
er
to
lear
n
tr
an
s
ac
tio
n
s
,
a
g
en
er
ativ
e
ad
v
er
s
ar
ial
n
etwo
r
k
to
ad
d
r
ess
class
im
b
alan
ce
,
an
d
e
x
p
lain
ab
le
AI
to
im
p
r
o
v
e
tr
an
s
p
ar
en
cy
a
n
d
c
o
m
p
lia
n
ce
.
2.
RE
L
AT
E
D
WO
RK
Pay
m
en
t
f
r
a
u
d
is
b
ec
o
m
i
n
g
in
cr
ea
s
in
g
l
y
p
o
p
u
la
r
,
b
u
t
s
till
to
o
d
if
f
icu
lt
to
d
et
ec
t
[
2
0
]
.
R
ec
en
t
r
esear
ch
h
as
f
o
cu
s
ed
o
n
m
o
d
els
b
ased
o
n
d
ee
p
lear
n
in
g
th
at
ca
n
lear
n
tr
an
s
ac
tio
n
s
,
d
ea
l
with
im
b
alan
ce
d
d
ata
a
n
d
ex
p
lain
th
e
d
ec
is
io
n
s
m
ad
e
b
y
th
e
m
o
d
el.
T
h
er
e
ar
e
th
r
ee
d
ir
ec
tio
n
s
o
f
r
esear
ch
in
e
-
p
ay
m
en
t
f
r
a
u
d
d
etec
tio
n
t
h
at
ar
e
in
p
ar
ticu
lar
n
ee
d
o
f
atten
tio
n
:
tr
an
s
f
o
r
m
e
r
-
b
ased
ar
c
h
itectu
r
es;
m
o
d
els
b
ased
o
n
class
im
b
alan
ce
an
d
ex
p
lain
ab
le
AI
.
2
.
1
.
T
ra
ns
f
o
rm
er
-
ba
s
ed
a
rc
hite
ct
ures
Fin
an
cial
f
r
au
d
d
etec
tio
n
r
eli
es
in
cr
ea
s
in
g
ly
o
n
tr
an
s
f
o
r
m
e
r
m
o
d
els
b
ec
a
u
s
e
th
ey
ca
n
c
ap
tu
r
e
th
e
d
ep
en
d
e
n
ce
b
etwe
en
tr
an
s
ac
tio
n
s
with
o
u
t
th
e
n
ee
d
o
f
r
ec
u
r
r
en
t
h
id
d
e
n
s
tates.
T
h
es
e
m
o
d
els
a
n
aly
ze
tr
an
s
ac
tio
n
d
ata
with
a
v
ar
iety
o
f
asp
ec
ts
,
s
u
c
h
as
tr
a
n
s
ac
tio
n
am
o
u
n
t,
tim
estam
p
,
m
e
r
ch
a
n
t
ty
p
e,
a
n
d
d
ev
ice
in
f
o
r
m
atio
n
,
to
id
e
n
tify
p
atter
n
s
o
f
f
r
a
u
d
u
len
t
ac
tiv
ity
[
2
1
]
.
Self
-
atten
tio
n
will
p
lace
d
if
f
er
en
t
weig
h
ts
o
n
d
if
f
er
e
n
t
f
ea
tu
r
es
o
r
tr
an
s
ac
tio
n
s
in
o
r
d
er
to
p
r
ed
ict
f
r
au
d
.
Atten
tio
n
s
co
r
es
ca
n
a
ls
o
b
e
u
s
ed
to
ass
is
t
in
in
ter
p
r
et
atio
n
,
d
is
p
lay
in
g
wh
at
p
o
r
tio
n
s
o
f
th
e
tr
an
s
ac
tio
n
h
is
to
r
y
ca
u
s
ed
th
e
f
r
au
d
d
ete
ctio
n
.
Acc
o
r
d
i
n
g
to
th
e
E
U
A
I
Act
f
o
r
s
y
s
tem
s
th
at
in
f
l
u
en
ce
u
s
er
s
o
r
r
e
q
u
ir
e
tr
an
s
p
ar
en
cy
[
1
4
]
,
tr
a
n
s
f
o
r
m
er
-
b
ased
m
o
d
els
o
u
t
p
er
f
o
r
m
tr
ad
itio
n
al
m
ac
h
i
n
e
lear
n
in
g
m
eth
o
d
s
an
d
d
ee
p
lear
n
in
g
m
o
d
els
lik
e
f
ee
d
-
f
o
r
war
d
n
etwo
r
k
s
a
n
d
L
STM
s
in
f
r
au
d
d
etec
tio
n
task
s
[
2
2
]
,
[
2
3
]
.
T
h
eir
ab
ilit
y
to
h
an
d
le
lo
n
g
-
r
a
n
g
e
d
ep
en
d
en
c
ies
m
ak
es
th
em
ef
f
ec
tiv
e
in
s
p
o
ttin
g
f
r
a
u
d
p
atter
n
s
th
at
h
ap
p
en
o
v
e
r
tim
e,
lik
e
ac
co
u
n
t ta
k
e
o
v
er
an
d
m
u
l
ti
-
s
tag
e
m
o
n
ey
lau
n
d
er
in
g
.
2
.
2
.
G
ener
a
t
iv
e
m
o
dels
f
o
r
c
la
s
s
im
ba
la
nce
C
las
s
im
b
alan
ce
r
em
ain
s
o
n
e
o
f
th
e
m
ain
d
if
f
icu
lties
in
f
r
au
d
d
etec
tio
n
,
s
in
ce
f
r
au
d
u
le
n
t tr
an
s
ac
tio
n
s
u
s
u
ally
r
ep
r
esen
t
o
n
ly
a
v
e
r
y
s
m
all
p
r
o
p
o
r
tio
n
o
f
t
h
e
d
ata.
I
n
s
o
m
e
d
atasets
,
f
r
au
d
ca
s
es
ac
co
u
n
t
f
o
r
less
th
an
0
.
1
%
o
f
all
tr
a
n
s
ac
tio
n
s
.
E
ar
lier
ap
p
r
o
ac
h
es
o
f
ten
r
elied
o
n
s
im
p
le
o
v
er
s
am
p
lin
g
,
b
u
t
r
ec
en
t
g
e
n
er
ativ
e
m
o
d
els p
r
o
v
id
e
m
o
r
e
f
lex
i
b
le
s
o
lu
tio
n
s
f
o
r
cr
ea
tin
g
s
y
n
th
eti
c
m
in
o
r
ity
-
class
ex
am
p
les [
2
4
]
-
[
2
6
]
.
GAN
-
b
ased
m
eth
o
d
s
g
en
er
at
e
s
y
n
th
etic
f
r
au
d
s
am
p
les
th
r
o
u
g
h
an
ad
v
er
s
ar
ial
p
r
o
ce
s
s
in
v
o
lv
in
g
a
g
en
er
ato
r
a
n
d
a
d
is
cr
im
in
at
o
r
.
Mo
r
e
r
ec
e
n
t
d
if
f
u
s
io
n
-
b
a
s
ed
ap
p
r
o
ac
h
es
g
e
n
er
ate
d
at
a
th
r
o
u
g
h
g
r
a
d
u
al
d
en
o
is
in
g
,
wh
ich
ca
n
im
p
r
o
v
e
tr
ain
in
g
s
tab
ilit
y
an
d
r
ed
u
ce
p
r
o
b
lem
s
s
u
ch
as
m
o
d
e
c
o
llap
s
e
[
2
5
]
,
[
2
6
]
.
T
h
is
is
im
p
o
r
tan
t
in
f
r
au
d
d
etec
tio
n
b
ec
a
u
s
e
m
o
d
els
m
u
s
t
lea
r
n
d
i
f
f
er
en
t
f
r
a
u
d
b
eh
av
i
o
r
s
r
a
th
er
th
an
o
v
er
f
it
t
o
a
n
ar
r
o
w
s
et
o
f
e
x
am
p
les.
Yin
an
d
Z
h
a
n
g
p
r
o
p
o
s
ed
GA
-
GAN,
wh
ich
u
s
es
g
ated
atten
tio
n
to
b
etter
r
ep
r
esen
t
th
e
n
o
n
-
s
tatio
n
ar
y
an
d
h
eter
o
g
en
eo
u
s
n
atu
r
e
o
f
f
r
au
d
u
len
t
t
r
an
s
ac
tio
n
s
[
2
4
]
.
B
y
weig
h
tin
g
im
p
o
r
tan
t
f
ea
tu
r
es
d
u
r
in
g
g
en
er
atio
n
,
GA
-
GAN
ca
n
p
r
o
d
u
ce
s
y
n
th
etic
f
r
a
u
d
s
am
p
les
th
at
ar
e
clo
s
er
to
r
ea
l
f
r
au
d
u
le
n
t
b
eh
a
v
io
r
an
d
im
p
r
o
v
e
class
if
ier
p
er
f
o
r
m
an
ce
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
4
,
D
e
c
em
b
er
20
2
6
:
1
7
6
8
-
1
7
7
6
1770
2
.
3
.
E
x
pla
ina
ble A
I
i
nte
g
ra
t
io
n
I
n
th
e
f
in
an
cial
s
ec
to
r
wh
er
e
th
e
u
s
e
o
f
AI
s
y
s
tem
s
i
s
g
ain
in
g
tr
ac
tio
n
,
th
e
n
ee
d
f
o
r
ex
p
lain
ab
ilit
y
h
as
g
r
o
wn
i
n
cr
ea
s
in
g
ly
s
ig
n
if
ican
t
in
f
in
an
cial
f
r
au
d
d
etec
tio
n
.
I
n
2
0
2
4
,
t
h
e
E
U
AI
Ac
t
was
ad
o
p
ted
th
at
s
ets
r
eq
u
ir
em
en
ts
f
o
r
th
e
i
n
teg
r
ity
o
f
AI
s
y
s
tem
s
,
r
is
k
m
an
a
g
em
en
t,
d
ata
g
o
v
er
n
an
ce
,
tr
a
n
s
p
ar
en
cy
,
h
u
m
an
o
v
er
s
ig
h
t,
an
d
co
n
f
o
r
m
ity
ass
ess
m
en
t
o
f
AI
s
y
s
tem
s
[
1
1
]
,
[
1
4
]
.
I
n
th
is
c
o
n
tex
t,
f
r
au
d
d
etec
tio
n
m
o
d
els
h
av
e
to
b
e
ac
cu
r
ate
as we
ll a
s
u
n
d
er
s
tan
d
a
b
le
an
d
au
d
itab
le.
SHAP
is
wid
ely
u
s
ed
in
f
in
an
cial
f
r
au
d
d
etec
tio
n
b
ec
au
s
e
it
p
r
o
v
id
es
f
ea
tu
r
e
-
le
v
el
ex
p
lan
atio
n
s
g
r
o
u
n
d
ed
in
c
o
o
p
e
r
ativ
e
g
am
e
th
eo
r
y
.
An
al
y
s
is
o
f
f
ea
tu
r
e
s
elec
tio
n
d
u
r
in
g
p
r
ed
ictio
n
all
o
ws
an
aly
s
ts
to
s
ee
wh
at
elem
en
ts
co
n
tr
i
b
u
te
to
e
ac
h
p
r
e
d
ictio
n
.
U
n
d
er
s
tan
d
in
g
h
o
w
ea
c
h
f
ea
tu
r
e
co
n
tr
ib
u
tes
to
ea
ch
p
r
e
d
ictio
n
im
p
r
o
v
es
in
ter
p
r
eta
b
ilit
y
o
f
m
o
d
el
b
eh
av
i
o
r
an
d
d
ec
is
io
n
m
ak
in
g
[
1
9
]
.
I
n
ad
d
itio
n
to
p
o
s
t
h
o
c
ex
p
lan
atio
n
tech
n
iq
u
es,
s
o
m
e
r
esear
c
h
er
s
h
av
e
attem
p
ted
to
d
esig
n
m
o
d
els
th
at
ar
e
in
tr
in
s
ically
in
ter
p
r
etab
le.
Fo
r
ex
am
p
le,
T
r
an
s
f
o
r
m
er
-
B
DI
co
m
b
in
es
tr
an
s
f
o
r
m
er
-
b
ased
p
r
ed
ictio
n
m
o
d
els
with
s
y
m
b
o
lic
r
ea
s
o
n
in
g
to
p
r
o
d
u
ce
a
d
ec
is
io
n
tr
ail
th
at
is
ea
s
y
to
u
n
d
er
s
tan
d
an
d
m
ain
tain
s
h
ig
h
d
etec
tio
n
r
ate
s
[
2
1
]
.
T
ab
Net
also
u
tili
ze
s
s
eq
u
en
tial
atten
tio
n
in
its
p
r
ed
ictio
n
to
c
h
o
o
s
e
f
ea
tu
r
es
th
at
will
b
e
u
s
ed
in
ea
ch
p
r
ed
ictio
n
,
wh
ich
ca
n
b
e
ea
s
ily
u
n
d
er
s
to
o
d
b
y
an
a
ly
s
ts
[
2
2
]
.
As
b
a
n
k
s
d
em
a
n
d
h
ig
h
p
r
e
d
ictiv
e
p
e
r
f
o
r
m
an
ce
an
d
c
o
m
p
lian
ce
,
ex
p
lain
ab
le
ap
p
r
o
ac
h
es a
r
e
in
cr
ea
s
in
g
ly
im
p
o
r
tan
t to
p
r
o
v
i
d
e
a
r
eliab
le
s
y
s
tem
f
o
r
f
r
au
d
d
etec
tio
n
.
3.
M
E
T
H
O
D
I
n
th
is
s
tu
d
y
,
a
d
ee
p
lear
n
in
g
f
r
am
ewo
r
k
was
in
tr
o
d
u
ce
d
t
o
d
etec
t
e
-
p
ay
m
e
n
t
f
r
au
d
s
in
r
ea
l
-
tim
e
.
I
t
in
clu
d
es
th
r
ee
p
ar
ts
:
th
e
tr
a
n
s
f
o
r
m
er
-
b
ased
m
o
d
el
with
m
u
lti
-
h
ea
d
s
elf
-
atten
tio
n
,
s
y
n
th
etic
d
ata
g
en
er
atio
n
b
ased
o
n
GA
-
GAN
to
tack
le
class
im
b
alan
ce
,
an
d
ex
p
lain
ab
ilit
y
b
ased
o
n
SHAP
to
en
ab
le
ex
p
lain
ab
le
d
ec
is
io
n
s
.
T
h
e
g
en
e
r
al
p
r
o
ce
d
u
r
e
f
o
llo
we
d
is
s
h
o
wn
in
Fig
u
r
e
1
.
Fig
u
r
e
1
.
Pro
p
o
s
ed
d
ee
p
lear
n
i
n
g
f
r
a
m
ewo
r
k
3
.
1
.
Da
t
a
s
et
T
h
e
m
ain
ex
p
er
im
en
ts
u
s
e
th
e
Pay
Sim
f
in
a
n
cial
tr
an
s
a
ctio
n
d
ataset,
wh
ic
h
co
n
tain
s
5
0
,
0
0
0
tr
an
s
ac
tio
n
s
an
d
1
1
f
ea
tu
r
es.
T
h
ese
in
clu
d
e
th
e
tr
an
s
ac
tio
n
s
tep
,
tr
an
s
ac
tio
n
ty
p
e,
am
o
u
n
t,
o
r
ig
in
ato
r
an
d
r
ec
ip
ien
t
id
en
tifie
r
s
,
in
itial
an
d
f
in
al
b
ala
n
ce
s
f
o
r
b
o
t
h
p
ar
ties
,
an
d
th
e
b
in
ar
y
lab
els
is
Fra
u
d
an
d
is
Fl
ag
g
ed
Fra
u
d
.
T
o
ass
ess
g
en
er
aliza
tio
n
,
th
e
m
o
d
el
is
also
e
v
alu
ated
o
n
th
e
E
u
r
o
p
ea
n
C
r
ed
it
C
ar
d
Fra
u
d
Data
s
et,
wh
ich
in
clu
d
es
2
8
4
,
8
0
7
tr
a
n
s
ac
tio
n
s
an
d
3
0
f
ea
tu
r
es,
an
d
th
e
I
E
E
E
-
C
I
S
Fra
u
d
Dete
cti
o
n
Data
s
et,
wh
ich
co
n
tain
s
5
6
1
,
0
1
3
tr
ain
in
g
tr
an
s
ac
tio
n
s
an
d
6
7
f
ea
tu
r
es.
T
h
e
s
e
d
atasets
r
ef
lect
th
e
s
tr
o
n
g
i
m
b
alan
ce
ty
p
ically
o
b
s
er
v
ed
i
n
f
r
a
u
d
d
etec
tio
n
,
with
f
r
au
d
ca
s
es
r
ep
r
esen
tin
g
ab
o
u
t
0
.
1
3
%
o
f
th
e
E
u
r
o
p
ea
n
C
r
ed
it
C
ar
d
Data
s
et
an
d
3
.
5
% o
f
th
e
Pay
Sim
d
ataset
[
2
5
]
.
3
.
2
.
Da
t
a
p
re
pro
ce
s
s
ing
T
h
e
p
u
r
p
o
s
e
o
f
th
is
s
t
ep
i
s
to
r
ed
u
ce
n
o
is
e
an
d
leak
ag
e
r
is
k
an
d
s
tr
en
g
th
e
n
th
e
m
o
d
el
.
No
n
-
in
f
o
r
m
ativ
e
i
d
en
tifie
r
s
li
k
e
s
tep
,
n
a
m
eOr
ig
,
is
Flag
g
ed
F
r
au
d
a
r
e
r
em
o
v
ed
.
T
h
e
tr
an
s
ac
tio
n
ty
p
e
is
u
s
ed
as
lear
n
ed
e
m
b
ed
d
in
g
s
r
at
h
er
th
an
o
n
e
h
o
t
en
co
d
in
g
to
lear
n
th
e
r
elatio
n
s
h
i
p
b
etwe
en
tr
an
s
ac
tio
n
ty
p
es
.
Fo
r
n
u
m
er
ical
v
a
r
iab
les,
we
s
ca
le
n
u
m
er
ical
v
ar
iab
les
u
s
in
g
R
o
b
u
s
tScaler
,
wh
ich
is
less
s
en
s
itiv
e
to
o
u
tlier
s
th
an
s
tan
d
ar
d
n
o
r
m
aliza
tio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
Dee
p
tr
a
n
s
fo
r
mer a
r
ch
itectu
r
e
s
fo
r
r
ea
l
-
time
e
-
p
a
yme
n
t
fr
a
u
d
d
etec
tio
n
…
(
R
ia
d
Lo
u
kili
)
1771
Oth
er
f
ea
tu
r
es
ar
e
ad
d
ed
to
en
co
d
e
tem
p
o
r
al
an
d
b
eh
a
v
io
r
al
asp
ec
ts
.
T
h
ey
en
co
m
p
ass
r
o
llin
g
s
tati
s
t
ics
o
f
tr
an
s
ac
tio
n
lev
els
f
o
r
5
s
tep
an
d
2
4
s
tep
win
d
o
ws,
tim
e
s
in
ce
last
tr
an
s
ac
ti
o
n
p
er
o
r
ig
i
n
ato
r
,
b
alan
ce
ch
a
n
g
e
r
atio
s
an
d
tr
a
n
s
ac
tio
n
v
elo
city
.
T
h
ese
v
ar
ia
b
les
g
iv
e
ad
d
itio
n
al
c
o
n
tex
t
to
th
e
m
o
d
el
to
ass
is
t
with
th
e
d
etec
tio
n
o
f
u
n
u
s
u
al
p
ay
m
en
t
b
eh
av
io
r
.
T
h
e
d
atasets
ar
e
s
p
lit
in
a
s
tr
a
tifie
d
m
an
n
e
r
b
ased
o
n
tim
e,
wh
er
e
7
0
%
is
u
s
ed
f
o
r
tr
ain
in
g
,
1
5
%
f
o
r
v
alid
atio
n
an
d
1
5
%
f
o
r
test
in
g
.
T
h
is
m
ain
tain
s
th
e
class
p
r
o
p
o
r
tio
n
s
an
d
m
ain
tain
s
c
h
r
o
n
o
lo
g
ical
o
r
d
er
,
m
in
im
izin
g
f
u
tu
r
e
in
f
o
r
m
atio
n
leak
ag
e
in
to
t
h
e
tr
ai
n
in
g
a
n
d
m
o
r
e
ac
c
u
r
ately
s
im
u
latin
g
th
e
c
o
n
d
itio
n
s
o
f
d
ep
lo
y
m
e
n
t [
2
7
]
.
3
.
3
.
Cla
s
s
im
ba
la
nce
ha
nd
lin
g
C
las
s
im
b
alan
ce
is
r
eg
ar
d
ed
a
s
a
co
r
e
asp
ec
t
o
f
th
e
m
eth
o
d
o
lo
g
y
,
as
tr
an
s
ac
tio
n
s
ar
e
u
n
lik
ely
to
b
e
f
r
au
d
u
le
n
t.
T
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
is
b
ased
o
n
GA
-
GAN
s
y
n
th
etic
d
ata
g
en
er
atio
n
an
d
co
s
t
-
s
en
s
itiv
e
lear
n
in
g
.
I
n
o
r
d
er
to
g
en
er
ate
m
in
o
r
it
y
class
s
am
p
les
th
at
r
ep
r
esen
t
th
e
s
tr
u
ctu
r
e
o
f
r
ea
l
f
r
a
u
d
u
len
t
tr
an
s
ac
tio
n
s
,
GA
-
GAN
em
p
lo
y
s
g
ated
atten
tio
n
[
2
4
]
.
T
h
e
g
en
er
ato
r
cr
ea
tes
f
ak
e
f
r
au
d
ex
am
p
les
f
r
o
m
r
an
d
o
m
n
o
is
e,
an
d
th
e
d
is
cr
im
in
ato
r
is
tr
ain
ed
to
id
en
tify
g
e
n
u
in
e
f
r
au
d
ex
am
p
les
f
r
o
m
th
e
g
en
er
ated
ex
a
m
p
les.
T
h
e
g
ated
atte
n
tio
n
m
ec
h
an
is
m
h
elp
s
th
e
g
en
er
ato
r
to
f
o
c
u
s
o
n
th
e
m
o
s
t
r
elev
an
t
f
ea
t
u
r
es
r
elate
d
to
f
r
au
d
.
GA
-
GAN
is
tr
ain
ed
f
o
r
1
0
0
ep
o
c
h
s
with
Ad
am
o
p
tim
izer
h
av
in
g
lear
n
in
g
r
ate
o
f
0
.
0
0
0
1
an
d
s
y
n
t
h
etic
s
am
p
les
ar
e
g
en
er
ated
u
n
til
th
e
r
atio
o
f
1
:5
(
f
alsely
l
ab
eled
to
co
r
r
ec
tly
la
b
eled
)
is
ac
h
iev
ed
.
Fo
ca
l
lo
s
s
h
elp
s
f
u
r
th
er
r
ed
u
ce
th
e
m
ajo
r
ity
class
d
o
m
in
an
ce
b
y
u
s
in
g
th
e
f
o
ca
l
lo
s
s
f
u
n
ctio
n
.
I
t
p
lace
s
g
r
ea
ter
em
p
h
asis
o
n
ch
allen
g
in
g
ca
s
es
an
d
h
elp
s
to
m
in
im
is
e
th
e
im
p
ac
t
o
f
ea
s
y
,
leg
itima
te
tr
an
s
ac
tio
n
s
wh
en
tr
ain
in
g
.
T
h
e
L
o
s
s
is
ca
lcu
lated
as:
(
)
=
−
(
1
−
)
l
og
(
)
wh
er
e
p
t
is
th
e
p
r
ed
icted
p
r
o
b
ab
ilit
y
f
o
r
th
e
ac
t
u
al
class
,
α
is
to
s
et
th
e
weig
h
ts
o
n
th
e
class
es,
an
d
γ
is
to
f
o
cu
s
o
n
h
ar
d
ex
am
p
les.
T
h
e
o
p
tim
u
m
v
alu
es
o
f
α
=
0
.
2
5
an
d
γ
=
2
.
0
ar
e
d
eter
m
in
ed
f
r
o
m
th
e
v
alid
atio
n
r
esu
lts
.
3
.
4
.
M
o
del
a
rc
hite
ct
ure
des
ig
n
T
h
e
p
r
o
p
o
s
ed
ar
c
h
itectu
r
e
is
d
iv
id
ed
in
to
th
r
ee
p
ar
ts
,
wh
ic
h
ar
e
illu
s
tr
ated
in
th
e
Fig
u
r
e
2
.
Firstl
y
,
th
e
f
ea
tu
r
e
em
b
ed
d
in
g
lay
e
r
co
n
v
er
ts
ca
teg
o
r
ical
f
ea
tu
r
e
s
in
to
lear
n
e
d
e
m
b
ed
d
in
g
s
a
n
d
m
a
p
s
n
u
m
er
ical
f
ea
tu
r
es
in
to
a
co
m
m
o
n
r
ep
r
es
en
tatio
n
s
p
ac
e.
T
h
is
r
esu
lts
in
a
s
er
ies
o
f
f
ea
t
u
r
e
to
k
en
s
th
at
ca
n
b
e
f
ed
in
to
th
e
tr
an
s
f
o
r
m
er
.
Seco
n
d
,
t
h
e
tr
an
s
f
o
r
m
er
e
n
co
d
e
r
h
as f
o
u
r
lay
er
s
,
eig
h
t a
tten
tio
n
h
ea
d
s
an
d
a
h
i
d
d
en
d
im
e
n
s
io
n
o
f
2
5
6
.
T
h
e
m
u
lti
-
h
ea
d
s
elf
-
atten
tio
n
,
f
ee
d
f
o
r
war
d
tr
a
n
s
f
o
r
m
ati
o
n
,
r
esid
u
al
co
n
n
ec
tio
n
,
an
d
l
ay
er
n
o
r
m
aliza
tio
n
ar
e
ap
p
lied
to
ea
ch
lay
er
.
T
h
e
atten
tio
n
m
ec
h
an
is
m
allo
ws th
e
m
o
d
el
to
lear
n
in
ter
ac
tio
n
s
b
etwe
en
tr
an
s
ac
tio
n
f
ea
tu
r
es.
Fo
r
in
ter
p
r
etab
ilit
y
,
atten
tio
n
weig
h
ts
ar
e
a
v
er
a
g
ed
ac
r
o
s
s
h
ea
d
s
an
d
la
y
er
s
to
o
b
tain
a
g
lo
b
al
f
ea
tu
r
e
-
im
p
o
r
tan
ce
s
ig
n
al.
Fi
n
ally
,
th
e
class
if
icatio
n
h
ea
d
co
n
s
is
ts
o
f
th
r
ee
f
u
lly
c
o
n
n
ec
ted
lay
er
s
with
d
im
e
n
s
io
n
s
[
2
5
6
,
1
2
8
,
6
4
,
1
]
.
GE
L
U
ac
tiv
atio
n
an
d
d
r
o
p
o
u
t
with
a
r
ate
o
f
0
.
3
ar
e
u
s
ed
b
ef
o
r
e
th
e
f
in
al
s
ig
m
o
id
o
u
tp
u
t,
wh
ich
p
r
o
d
u
ce
s
th
e
f
r
au
d
p
r
o
b
a
b
ilit
y
.
T
h
e
m
o
d
el
co
n
tain
s
ab
o
u
t
2
.
8
m
illi
o
n
p
ar
am
eter
s
,
m
ak
in
g
it su
itab
le
f
o
r
r
ea
l
-
tim
e
in
f
er
en
ce
w
h
ile
r
etai
n
in
g
e
n
o
u
g
h
c
ap
ac
ity
to
lear
n
co
m
p
le
x
f
r
au
d
p
atter
n
s
.
3
.
5
.
T
ra
ini
ng
T
h
e
m
o
d
el
is
tr
ain
e
d
u
s
in
g
A
d
am
W
with
a
weig
h
t
d
ec
a
y
o
f
0
.
0
1
a
n
d
a
n
in
itial
lear
n
in
g
r
a
te
o
f
1
e
-
4
.
A
co
s
in
e
an
n
ea
lin
g
s
ch
e
d
u
le
is
ap
p
lied
o
v
er
5
0
ep
o
c
h
s
.
T
h
e
b
atch
s
ize
is
s
et
to
2
5
6
,
a
n
d
ea
r
l
y
s
to
p
p
i
n
g
is
u
s
ed
with
a
p
atien
ce
o
f
s
ev
e
n
ep
o
ch
s
b
ased
o
n
v
alid
atio
n
AUC.
T
h
e
tr
ain
in
g
o
b
jectiv
e
co
m
b
in
es
b
in
ar
y
cr
o
s
s
-
en
tr
o
p
y
with
f
o
ca
l
lo
s
s
weig
h
tin
g
.
Sev
er
al
r
e
g
u
lar
iz
atio
n
tech
n
iq
u
es
ar
e
u
s
ed
to
lim
it
o
v
er
f
itti
n
g
,
in
clu
d
in
g
d
r
o
p
o
u
t
in
t
h
e
class
if
icatio
n
h
ea
d
,
s
to
ch
asti
c
d
ep
th
i
n
th
e
t
r
an
s
f
o
r
m
er
l
ay
er
s
,
an
d
Mix
u
p
au
g
m
en
tatio
n
with
α
=
0
.
2
.
T
r
ain
in
g
is
p
er
f
o
r
m
ed
o
n
an
NV
I
DI
A
A1
0
0
GPU
with
4
0
GB
m
em
o
r
y
a
n
d
ta
k
es
ap
p
r
o
x
im
ately
4
5
m
in
u
tes.
3
.
6
.
E
v
a
lua
t
i
o
n
T
h
e
m
o
d
el
is
ev
alu
ated
u
s
in
g
m
etr
ic
s
s
u
ited
to
im
b
alan
ce
d
c
lass
if
icatio
n
:
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
F1
-
s
co
r
e,
AUC
-
R
OC
,
A
UC
-
P
R
,
an
d
Ma
tth
ews
C
o
r
r
elatio
n
C
o
ef
f
icien
t.
B
alan
ce
d
ac
cu
r
ac
y
an
d
th
e
g
eo
m
etr
ic
m
ea
n
o
f
s
en
s
itiv
ity
an
d
s
p
ec
if
icity
ar
e
also
r
ep
o
r
ted
to
p
r
o
v
id
e
a
m
o
r
e
r
eliab
le
v
iew
o
f
p
er
f
o
r
m
an
ce
u
n
d
er
class
im
b
alan
ce
.
Fo
r
th
ese
p
r
ed
ictio
n
s
to
b
e
u
n
d
er
s
tan
d
ab
l
e,
SHAP
v
alu
es
ar
e
ca
lcu
late
d
.
W
e
u
s
e
b
o
o
ts
tr
ap
r
esam
p
lin
g
to
ass
ess
f
ea
tu
r
e
im
p
o
r
tan
ce
s
tab
ilit
y
an
d
th
e
SHAP
Stab
ili
ty
I
n
d
ex
to
ass
ess
co
n
s
is
ten
cy
o
f
ex
p
lan
atio
n
s
.
T
h
is
is
im
p
o
r
ta
n
t
as
ex
p
lan
atio
n
in
s
tab
ilit
y
m
ay
lead
to
m
is
tr
u
s
t
in
f
r
a
u
d
d
etec
tio
n
s
y
s
tem
s
esp
ec
ially
wh
en
class
-
b
alan
cin
g
is
u
s
ed
[
2
8
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
4
,
D
e
c
em
b
er
20
2
6
:
1
7
6
8
-
1
7
7
6
1772
Fig
u
r
e
2
.
Ar
c
h
itectu
r
e
o
f
th
e
p
r
o
p
o
s
ed
tr
a
n
s
f
o
r
m
e
r
-
b
ased
m
o
d
el
f
o
r
f
r
au
d
d
etec
tio
n
4.
RE
SU
L
T
S
AND
D
I
SCU
SS
I
O
N
R
esu
lts
s
h
o
w
th
at
th
e
m
o
d
el
p
er
f
o
r
m
s
well
in
ea
ch
o
f
th
e
th
r
ee
d
atasets
,
with
th
e
h
ig
h
est
p
er
f
o
r
m
an
ce
o
n
Pay
Sim.
4
.
1
.
M
o
del
p
er
f
o
rma
nce
T
h
e
m
o
d
el
with
a
tr
an
s
f
o
r
m
e
r
h
as
an
ac
cu
r
ac
y
o
f
9
9
.
8
2
%
an
d
an
AUC
-
R
OC
o
f
9
9
.
5
6
%
o
n
th
e
Pay
Sim
d
ataset.
T
h
e
AU
C
-
P
R
is
9
8
.
8
7
%
d
u
e
to
b
iased
d
ata
f
o
r
f
r
au
d
d
etec
tio
n
.
Ma
tth
ews
co
r
r
elatio
n
co
ef
f
icien
t (
MCC
)
was 0
.
9
6
1
,
in
d
icatin
g
g
o
o
d
a
g
r
ee
m
en
t
b
et
wee
n
th
e
p
r
e
d
icted
an
d
r
ea
l la
b
els.
T
h
e
co
n
f
u
s
io
n
m
atr
ix
f
u
r
t
h
er
co
n
f
ir
m
s
th
e
s
tab
ilit
y
o
f
th
e
m
o
d
el
.
I
t
g
en
er
a
ted
1
4
,
8
6
5
tr
u
e
n
eg
ativ
es,
1
2
6
tr
u
e
p
o
s
itiv
es,
7
f
alse
p
o
s
itiv
es
an
d
o
n
ly
2
f
alse
n
e
g
ativ
es.
T
h
is
is
wh
er
e
th
e
m
o
d
el
ca
n
p
r
ed
ict
s
u
cc
ess
f
u
l
tr
an
s
ac
tio
n
s
an
d
f
r
au
d
u
len
t
tr
an
s
ac
tio
n
s
,
an
d
m
in
im
ize
m
is
s
ed
f
r
au
d
ca
s
es
an
d
f
alse f
r
au
d
aler
ts
.
T
h
e
m
o
d
el
s
h
o
wed
g
o
o
d
r
esu
lts
o
n
o
th
er
d
atase
ts
as
well.
I
t
is
co
m
p
etitiv
e
with
r
ec
en
t
r
esu
lts
in
liter
atu
r
e
[
2
4
]
with
a
n
AUC
-
R
OC
o
f
0
.
9
8
4
an
d
an
AUC
-
PR
o
f
0
.
8
9
2
o
n
th
e
E
u
r
o
p
ea
n
cr
ed
it
ca
r
d
f
r
au
d
d
ataset
.
T
h
e
m
o
d
el
ac
h
iev
e
d
an
AUC
-
R
OC
o
f
0
.
9
7
4
o
n
th
e
I
E
E
E
-
C
I
S
d
ataset,
in
d
icatin
g
th
at
th
e
m
o
d
el
ca
n
be
g
en
er
alize
d
to
o
th
er
tr
an
s
ac
tio
n
en
v
ir
o
n
m
en
ts
an
d
f
ea
tu
r
e
s
p
ac
es.
4
.
2
.
Co
m
pa
ra
t
iv
e
a
na
ly
s
is
T
ab
le
1
an
d
Fig
u
r
e
3
p
r
esen
t
a
co
m
p
ar
is
o
n
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
with
r
ec
e
n
t
f
r
a
u
d
d
etec
tio
n
b
aselin
es.
T
h
e
p
r
o
p
o
s
ed
tr
an
s
f
o
r
m
er
m
o
d
el
o
u
tp
er
f
o
r
m
s
th
e
co
m
p
ar
ed
m
e
th
o
d
s
in
ter
m
s
o
f
AUC
-
R
OC
an
d
F1
-
s
co
r
e
an
d
also
p
r
o
v
i
d
es
ex
p
lan
atio
n
s
u
s
in
g
SHAP
v
alu
es.
T
GNN
also
h
as
g
o
o
d
p
er
f
o
r
m
an
ce
,
p
ar
ticu
la
r
ly
in
tr
an
s
ac
tio
n
n
etwo
r
k
s
th
at
h
av
e
h
ig
h
ly
s
tr
u
ctu
r
e
d
g
r
a
p
h
s
,
b
u
t
its
ap
p
licatio
n
is
b
et
ter
s
u
ited
to
f
r
au
d
s
ce
n
ar
io
s
in
v
o
l
v
in
g
a
g
r
ap
h
r
ath
er
th
a
n
r
eg
u
la
r
e
-
c
o
m
m
e
r
ce
tr
an
s
ac
tio
n
s
.
I
n
ter
p
r
etab
il
ity
is
ac
h
iev
e
d
in
T
r
an
s
f
o
r
m
e
r
-
B
DI
an
d
XFL,
wh
ile
th
eir
p
r
ed
ictiv
e
ac
cu
r
a
cy
is
lo
wer
in
th
is
co
m
p
ar
is
o
n
.
GA
-
GAN
an
d
Dif
f
u
s
io
n
-
SDG
ar
e
b
en
e
f
icial
f
o
r
s
y
n
t
h
etic
d
ata
g
e
n
er
atio
n
b
u
t
a
r
e
n
o
t
m
ea
n
t
to
b
e
c
o
m
p
lete
en
d
-
to
-
e
nd
f
r
au
d
d
etec
tio
n
to
o
ls
.
T
ab
le
1
.
Mo
d
el
co
m
p
ar
is
o
n
to
b
aselin
e
m
o
d
els
M
o
d
e
l
AUC
-
R
O
C
F1
-
S
c
o
r
e
Ex
p
l
a
i
n
a
b
l
e
B
a
se
l
i
n
e
D
N
N
0
.
9
7
4
3
0
.
9
4
9
No
Tr
a
n
sf
o
r
mer (
P
r
o
p
o
se
d
)
0
.
9
9
5
6
0
.
9
7
4
Y
e
s (S
H
A
P
)
Tr
a
n
sf
o
r
mer
-
B
D
I
[
2
1
]
0
.
9
5
0
0
0
.
9
2
0
Y
e
s (BDI
)
TG
N
N
[
2
9
]
0
.
9
9
2
4
0
.
9
6
1
Li
mi
t
e
d
GA
-
G
A
N
[
2
4
]
0
.
9
2
9
0
0
.
8
7
5
No
X
F
L
[
3
0
]
0
.
9
4
0
0
0
.
8
9
0
Y
e
s (S
H
A
P
+
LI
M
E)
D
i
f
f
u
si
o
n
-
S
D
G
[
2
5
]
0
.
9
2
0
0
0
.
8
9
0
No
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
Dee
p
tr
a
n
s
fo
r
mer a
r
ch
itectu
r
e
s
fo
r
r
ea
l
-
time
e
-
p
a
yme
n
t
fr
a
u
d
d
etec
tio
n
…
(
R
ia
d
Lo
u
kili
)
1773
Fig
u
r
e
3
.
Per
f
o
r
m
an
c
e
co
m
p
ar
is
o
n
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
with
th
e
latest f
r
au
d
d
etec
tio
n
b
e
n
ch
m
ar
k
4
.
3
.
Dis
cus
s
io
n
T
h
e
ef
f
ec
tiv
en
ess
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
ca
n
b
e
attr
ib
u
te
d
to
th
e
s
y
n
e
r
g
y
a
m
o
n
g
th
e
th
r
ee
k
ey
co
m
p
o
n
en
ts
-
tr
an
s
f
o
r
m
er
lear
n
in
g
,
s
y
n
th
etic
d
ata
g
en
er
ati
o
n
,
an
d
c
o
s
t
-
s
en
s
itiv
e
tr
ain
in
g
.
T
h
e
s
elf
-
atten
tio
n
m
ec
h
an
is
m
en
a
b
les
th
e
m
o
d
el
to
b
etter
g
r
asp
th
e
in
ter
ac
tio
n
s
am
o
n
g
tr
an
s
ac
tio
n
f
ea
t
u
r
es
th
an
tr
ad
itio
n
al
f
ee
d
f
o
r
war
d
m
o
d
el
s
.
GA
-
GAN
en
h
an
ce
s
th
e
r
ep
r
esen
tatio
n
o
f
t
h
e
m
in
o
r
ity
f
r
au
d
ca
s
es
b
y
c
r
ea
tin
g
m
o
r
e
r
ea
lis
tic
s
y
n
th
etic
s
am
p
les,
an
d
f
o
ca
l
lo
s
s
m
itig
ate
s
th
e
im
p
ac
t
o
f
leg
itima
te
t
r
an
s
ac
tio
n
s
d
u
r
i
n
g
th
e
lear
n
in
g
p
r
o
ce
s
s
.
T
h
e
SHAP
an
aly
s
is
al
s
o
h
elp
s
in
th
e
m
o
d
el
in
ter
p
r
etab
ilit
y
.
T
h
e
m
o
s
t
in
f
lu
en
tial
f
ea
tu
r
es
ar
e
in
lin
e
with
p
r
ac
tical
f
r
au
d
an
aly
s
i
s
o
f
tr
an
s
ac
tio
n
am
o
u
n
t,
b
alan
ce
ch
an
g
e
r
atio
s
an
d
tr
an
s
ac
tio
n
v
elo
city
.
An
ex
p
lan
atio
n
is
co
n
s
is
ten
t
ac
r
o
s
s
r
esam
p
lin
g
ex
p
er
im
en
ts
as
in
d
icate
d
b
y
a
s
tab
i
lity
in
d
ex
o
f
0
.
9
4
f
o
r
th
e
SHAP
[
2
8
]
.
T
h
e
m
o
d
e
l
is
al
s
o
ap
p
r
o
p
r
iate
f
o
r
r
ea
l
tim
e
ap
p
licatio
n
s
.
I
t
in
f
er
s
tr
an
s
ac
tio
n
s
in
ju
s
t
o
v
er
3
.
2
m
s
p
er
tr
an
s
ac
tio
n
o
v
er
a
s
in
g
le
C
PU
co
r
e,
wh
ich
is
m
o
r
e
th
an
3
0
0
tr
an
s
ac
tio
n
s
p
er
s
ec
o
n
d
.
T
h
is
is
th
e
lev
el
o
f
laten
cy
th
at
is
ac
ce
p
tab
le
f
o
r
m
o
s
t
e
-
c
o
m
m
er
ce
p
ay
m
en
t
p
latf
o
r
m
s
,
wh
er
e
cu
s
to
m
er
s
s
h
o
u
ld
n
o
t
ex
p
er
ien
ce
a
n
y
d
ela
y
s
in
th
e
p
ay
m
en
t p
r
o
ce
s
s
wh
ile
th
e
s
y
s
tem
d
etec
ts
f
r
au
d
.
5.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
in
tr
o
d
u
ce
s
th
e
d
ee
p
lear
n
in
g
tech
n
iq
u
e
f
o
r
r
ea
l
-
tim
e
e
-
p
ay
m
en
t
f
r
au
d
d
e
tectio
n
in
e
-
co
m
m
er
ce
.
I
t
in
cl
u
d
es
tr
a
n
s
f
o
r
m
er
-
b
ased
le
ar
n
i
n
g
,
GA
-
GAN
s
y
n
th
etic
d
ata
g
e
n
e
r
atio
n
,
an
d
SHAP
ex
p
lain
ab
ilit
y
f
o
r
f
r
a
u
d
d
etec
t
io
n
.
R
esu
lts
f
r
o
m
ex
p
er
im
en
t
s
s
h
o
w
th
at
t
h
e
p
r
o
p
o
s
ed
m
o
d
el
p
er
f
o
r
m
e
d
well
with
ac
cu
r
ac
y
o
f
9
9
.
8
2
%
an
d
an
AUC
o
f
9
9
.
5
6
%
in
Pay
Sim
d
ataset.
T
h
e
m
o
d
els
ar
e
ab
le
t
o
d
etec
t
f
r
au
d
u
len
t
tr
an
s
ac
tio
n
s
with
h
ig
h
ac
cu
r
a
cy
an
d
co
m
p
en
s
ate
f
o
r
class
i
m
b
alan
ce
b
y
r
ed
u
cin
g
f
o
ca
l
l
o
s
s
an
d
g
e
n
er
atin
g
s
y
n
th
etic
d
ata.
T
h
e
lo
w
in
f
er
en
ce
laten
cy
o
f
th
e
m
o
d
el
h
i
g
h
lig
h
ts
its
u
s
e
in
r
ea
l
-
tim
e
in
f
er
en
ce
,
s
u
c
h
as
e
-
co
m
m
er
ce
p
a
y
m
en
t
s
y
s
tem
s
.
On
th
e
g
r
o
u
n
d
,
th
e
p
r
o
p
o
s
ed
m
o
d
el
ca
n
h
el
p
s
ec
u
r
e
d
ig
ital
p
ay
m
en
ts
,
m
in
im
ize
f
in
an
cial
lo
s
s
es,
an
d
in
cr
ea
s
e
co
n
s
u
m
er
t
r
u
s
t
in
o
n
lin
e
tr
an
s
ac
tio
n
s
.
On
e
o
f
th
e
b
ig
g
est
ad
v
an
ta
g
es
o
f
in
tr
o
d
u
cin
g
t
h
e
m
o
d
el
is
e
x
p
lain
ab
ilit
y
,
wh
ich
h
elp
s
th
e
f
r
au
d
a
n
aly
s
t
o
r
d
ec
is
io
n
-
m
a
k
e
r
u
n
d
er
s
tan
d
wh
y
a
s
u
s
p
icio
u
s
tr
an
s
ac
tio
n
.
T
h
is
is
p
ar
ticu
lar
ly
im
p
o
r
tan
t
in
th
e
f
in
an
cial
in
d
u
s
tr
y
,
wh
er
e
tr
an
s
p
ar
en
c
y
,
ac
co
u
n
tab
ilit
y
,
a
n
d
r
e
g
u
lato
r
y
co
m
p
lian
ce
ar
e
b
ec
o
m
i
n
g
in
cr
ea
s
in
g
ly
im
p
o
r
ta
n
t.
Alth
o
u
g
h
t
h
ese
f
in
d
in
g
s
ar
e
e
n
co
u
r
a
g
in
g
,
r
esear
ch
n
ee
d
s
to
b
e
d
o
n
e
f
u
r
th
er
o
n
a
n
u
m
b
er
o
f
ar
ea
s
.
Fo
r
ex
am
p
le,
L
L
M
in
teg
r
atio
n
co
u
ld
b
e
u
s
ed
to
p
r
o
ce
s
s
u
n
s
tr
u
ctu
r
ed
d
ata
s
u
ch
as
tr
an
s
ac
tio
n
d
escr
ip
tio
n
s
,
cu
s
to
m
er
m
ess
ag
es,
o
r
m
er
c
h
an
t
r
ev
iews.
Seco
n
d
,
u
s
in
g
g
r
ap
h
n
eu
r
al
n
etwo
r
k
s
,
it
is
p
o
s
s
ib
le
to
ca
p
tu
r
e
r
elat
io
n
s
h
ip
s
b
etwe
en
u
s
er
s
,
ac
co
u
n
ts
,
m
er
ch
a
n
ts
,
an
d
tr
a
n
s
ac
tio
n
s
,
an
d
th
u
s
id
en
tify
co
o
r
d
in
ate
d
f
r
au
d
s
ch
em
es.
T
h
ir
d
,
f
ed
e
r
ated
le
ar
n
in
g
co
u
ld
allo
w
m
u
ltip
le
f
in
an
cial
in
s
titu
tio
n
s
to
jo
in
tl
y
d
ev
elo
p
a
f
r
au
d
d
etec
tio
n
m
o
d
el
with
o
u
t
s
h
ar
i
n
g
cu
s
to
m
er
in
f
o
r
m
atio
n
.
Fu
r
th
er
r
esear
ch
is
n
ee
d
ed
o
n
co
n
tin
u
o
u
s
lear
n
in
g
to
ad
ap
t
th
e
m
o
d
el
to
th
e
ev
o
lv
in
g
f
r
a
u
d
p
atter
n
o
v
er
tim
e.
Stat
ic
m
o
d
els
will
n
o
lo
n
g
e
r
b
e
u
s
ef
u
l
if
th
ey
ar
e
n
o
t
co
n
tin
u
ally
u
p
d
ated
with
ch
a
n
g
in
g
f
r
au
d
u
len
t
ac
tiv
ity
.
Fo
u
r
th
,
it
is
im
p
o
r
tan
t
to
in
co
r
p
o
r
ate
ex
p
lain
ab
ilit
y
an
d
co
n
s
tr
ain
ts
in
to
th
e
r
ea
l
-
wo
r
ld
s
u
ch
as
d
o
c
u
m
e
n
tatio
n
,
h
u
m
an
o
v
e
r
s
ig
h
t,
s
y
s
tem
s
ca
lab
ilit
y
,
an
d
r
esis
tan
ce
to
ad
v
e
r
s
ar
ial
attac
k
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
4
,
D
e
c
em
b
er
20
2
6
:
1
7
6
8
-
1
7
7
6
1774
ACK
NO
WL
E
DG
M
E
N
T
S
R
iad
L
o
u
k
ili
wo
u
ld
lik
e
to
ex
p
r
ess
h
is
s
in
ce
r
e
g
r
atitu
d
e
to
h
is
s
u
p
er
v
is
o
r
,
Pr
o
f
.
Fay
ça
l
Me
s
s
ao
u
d
i,
f
o
r
h
is
g
u
id
an
ce
an
d
co
n
tin
u
e
d
s
u
p
p
o
r
t
th
r
o
u
g
h
o
u
t
th
is
r
esear
ch
.
He
also
e
x
ten
d
s
h
is
h
ea
r
tf
elt
th
an
k
s
to
h
is
s
is
ter
,
Pro
f
.
Ma
n
al
L
o
u
k
ili,
f
o
r
h
er
en
co
u
r
ag
em
e
n
t,
s
u
p
p
o
r
t,
an
d
v
alu
ab
le
ad
v
ice
th
r
o
u
g
h
o
u
t
h
is
ac
ad
em
ic
jo
u
r
n
e
y.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
T
h
e
au
th
o
r
s
s
tate
th
at
n
o
f
u
n
d
i
n
g
was r
ec
eiv
ed
f
o
r
t
h
is
r
esear
ch
.
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
R
iad
L
o
u
k
ili
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Fay
ça
l M
ess
ao
u
d
i
✓
✓
✓
✓
✓
✓
✓
R
ao
u
y
a
E
l Y
o
u
b
i
✓
✓
✓
✓
✓
Ma
n
al
L
o
u
k
ili
✓
✓
✓
✓
✓
✓
✓
✓
✓
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
T
h
e
au
th
o
r
s
s
tate
th
at
th
ey
h
av
e
n
o
co
n
f
lict o
f
i
n
ter
est
.
I
NF
O
RM
E
D
CO
NS
E
N
T
T
h
is
s
tu
d
y
d
id
n
o
t
in
v
o
lv
e
th
e
d
ir
ec
t
r
ec
r
u
itm
en
t
o
f
h
u
m
an
p
ar
ticip
an
ts
o
r
th
e
c
o
llectio
n
o
f
p
er
s
o
n
ally
id
e
n
tifia
b
le
in
f
o
r
m
atio
n
.
E
T
H
I
CAL AP
P
RO
V
AL
T
h
e
s
tu
d
y
was
co
n
d
u
cted
u
s
i
n
g
ex
is
tin
g
b
en
ch
m
ar
k
d
atas
ets
an
d
d
id
n
o
t
in
v
o
lv
e
ex
p
e
r
im
en
ts
o
n
h
u
m
an
p
ar
ticip
an
ts
o
r
a
n
im
als.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
atasets
u
s
ed
in
t
h
is
s
tu
d
y
ar
e
p
u
b
licly
a
v
ailab
le
f
r
o
m
th
eir
r
esp
ec
tiv
e
r
ep
o
s
ito
r
ies.
T
h
e
Pay
Sim
f
in
an
cial
tr
an
s
ac
tio
n
d
ata
s
et
is
av
ailab
le
at
h
t
tp
s
://g
ith
u
b
.
co
m
/Ed
g
ar
L
o
p
ez
Ph
D/Pay
Sim
an
d
h
ttp
s
://www.
k
ag
g
le.
co
m
/d
atasets
/eala
x
i/p
ay
s
im
1
.
T
h
e
E
u
r
o
p
ea
n
C
r
ed
it
C
ar
d
Fra
u
d
Data
s
et,
p
r
o
v
i
d
ed
b
y
th
e
Ma
ch
i
n
e
L
ea
r
n
in
g
Gr
o
u
p
o
f
U
n
iv
er
s
ité
L
i
b
r
e
d
e
B
r
u
x
elles,
is
av
ailab
le
at
h
ttp
s
://www.
k
ag
g
le.
co
m
/d
atasets
/m
lg
-
u
lb
/cr
ed
itcar
d
f
r
au
d
.
T
h
e
I
E
E
E
-
C
I
S
Fra
u
d
Det
ec
tio
n
Data
s
et
is
av
ailab
le
th
r
o
u
g
h
th
e
o
f
f
icial
I
E
E
E
-
C
I
S
Kag
g
le
co
m
p
etitio
n
at
h
ttp
s
://www.
k
ag
g
le.
co
m
/co
m
p
et
itio
n
s
/ieee
-
f
r
au
d
-
d
etec
tio
n
/d
ata.
No
n
ew
p
r
o
p
r
ietar
y
d
ataset
was c
r
ea
ted
f
o
r
t
h
is
s
tu
d
y
.
RE
F
E
R
E
NC
E
S
[
1
]
O
.
El
A
a
l
o
u
c
h
e
,
F
.
M
e
ss
a
o
u
d
i
,
M
.
Lo
u
k
i
l
i
,
a
n
d
R
.
L
o
u
k
i
l
i
,
“
A
t
e
m
p
o
r
a
l
a
d
a
p
t
i
v
e
f
u
z
z
y
c
l
u
s
t
e
r
i
n
g
f
r
a
m
e
w
o
r
k
f
o
r
d
y
n
a
mi
c
b
e
h
a
v
i
o
r
a
l
se
g
me
n
t
a
t
i
o
n
i
n
e
-
c
o
mm
e
r
c
e
,”
S
t
a
t
i
s
t
i
c
s,
O
p
t
i
m
i
za
t
i
o
n
a
n
d
I
n
f
o
rm
a
t
i
o
n
C
o
m
p
u
t
i
n
g
,
v
o
l
.
1
5
,
n
o
.
4
,
p
p
.
3
0
2
9
–
3
0
4
1
,
M
a
r
.
2
0
2
6
,
d
o
i
:
1
0
.
1
9
1
3
9
/
s
o
i
c
-
2
3
1
0
-
5
0
7
0
-
3
2
4
3
.
[
2
]
R
.
E
l
Y
o
u
b
i
,
F
.
M
e
ss
a
o
u
d
i
,
M
.
L
o
u
k
i
l
i
,
R
.
Lo
u
k
i
l
i
,
a
n
d
O
.
El
A
a
l
o
u
c
h
e
,
“
F
r
o
m
c
l
i
c
k
s
t
o
c
o
n
v
e
r
si
o
n
s:
l
e
v
e
r
a
g
i
n
g
a
p
r
i
o
r
i
a
n
d
b
e
h
a
v
i
o
u
r
a
l
se
g
m
e
n
t
a
t
i
o
n
i
n
e
-
c
o
m
merce
,”
S
t
a
t
i
s
t
i
c
s,
O
p
t
i
m
i
za
t
i
o
n
a
n
d
I
n
f
o
rm
a
t
i
o
n
C
o
m
p
u
t
i
n
g
,
v
o
l
.
1
5
,
n
o
.
1
,
p
p
.
8
4
8
–
8
5
8
,
O
c
t
.
2
0
2
5
,
d
o
i
:
1
0
.
1
9
1
3
9
/
s
o
i
c
-
2
3
1
0
-
5
0
7
0
-
2
8
9
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
Dee
p
tr
a
n
s
fo
r
mer a
r
ch
itectu
r
e
s
fo
r
r
ea
l
-
time
e
-
p
a
yme
n
t
fr
a
u
d
d
etec
tio
n
…
(
R
ia
d
Lo
u
kili
)
1775
[
3
]
M
.
L
o
u
k
i
l
i
,
F
.
M
e
ssa
o
u
d
i
,
a
n
d
R
.
El
Y
o
u
b
i
,
En
h
a
n
c
i
n
g
f
i
n
a
n
c
i
a
l
t
ra
n
s
a
c
t
i
o
n
s
e
c
u
ri
t
y
:
a
d
e
e
p
l
e
a
rn
i
n
g
a
p
p
r
o
a
c
h
f
o
r
e
-
p
a
y
m
e
n
t
f
ra
u
d
d
e
t
e
c
t
i
o
n
.
B
o
c
a
R
a
t
o
n
:
C
h
a
p
m
a
n
a
n
d
H
a
l
l
/
C
R
C
,
2
0
2
4
.
d
o
i
:
1
0
.
1
2
0
1
/
9
7
8
1
0
3
2
6
5
6
8
3
0
-
1
4
.
[
4
]
M
.
Lo
u
k
i
l
i
,
F
.
M
e
ss
a
o
u
d
i
,
O
.
E
l
A
a
l
o
u
c
h
e
,
R
.
El
Y
o
u
b
i
,
a
n
d
R
.
L
o
u
k
i
l
i
,
“
A
d
a
p
t
i
v
e
p
r
i
c
i
n
g
st
r
a
t
e
g
i
e
s
i
n
d
i
g
i
t
a
l
mark
e
t
i
n
g
,
”
S
t
a
t
i
st
i
c
s,
O
p
t
i
m
i
z
a
t
i
o
n
a
n
d
I
n
f
o
rm
a
t
i
o
n
C
o
m
p
u
t
i
n
g
,
v
o
l
.
1
4
,
n
o
.
3
,
p
p
.
1
2
4
4
–
1
2
5
1
,
Ju
l
.
2
0
2
5
,
d
o
i
:
1
0
.
1
9
1
3
9
/
s
o
i
c
-
2
3
1
0
-
5
0
7
0
-
2
2
0
0
.
[5
]
M
.
A
n
s
h
a
r
i
a
n
d
Y
.
A
l
a
s,
“
S
mart
p
h
o
n
e
s
h
a
b
i
t
s,
n
e
c
e
ssi
t
i
e
s,
a
n
d
b
i
g
d
a
t
a
c
h
a
l
l
e
n
g
e
s,”
J
o
u
r
n
a
l
o
f
H
i
g
h
T
e
c
h
n
o
l
o
g
y
Ma
n
a
g
e
m
e
n
t
Re
se
a
rc
h
,
v
o
l
.
2
6
,
n
o
.
2
,
p
p
.
1
7
7
–
1
8
5
,
2
0
1
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
h
i
t
e
c
h
.
2
0
1
5
.
0
9
.
0
0
5
.
[
6
]
B
.
U
l
,
R
.
F
.
,
A
.
M
e
h
r
a
j
,
A
.
A
h
m
a
d
,
a
n
d
S
.
A
ssa
d
,
“
A
c
o
m
p
e
n
d
i
o
u
s
st
u
d
y
o
f
o
n
l
i
n
e
p
a
y
me
n
t
s
y
st
e
ms:
p
a
st
d
e
v
e
l
o
p
me
n
t
s,
p
r
e
se
n
t
i
mp
a
c
t
,
a
n
d
f
u
t
u
r
e
c
o
n
si
d
e
r
a
t
i
o
n
s
,”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
Ad
v
a
n
c
e
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
8
,
n
o
.
5
,
2
0
1
7
,
d
o
i
:
1
0
.
1
4
5
6
9
/
i
j
a
c
sa
.
2
0
1
7
.
0
8
0
5
3
2
.
[
7
]
M
.
Q
a
sa
i
me
h
,
N
.
A
.
H
a
l
e
m
a
h
,
R
.
R
a
w
a
sh
d
e
h
,
R
.
S
.
A
l
-
Q
a
ssas
,
a
n
d
A
.
Q
u
sef,
“
S
y
s
t
e
m
a
t
i
c
r
e
v
i
e
w
o
f
e
-
c
o
mm
e
r
c
e
s
e
c
u
r
i
t
y
i
ss
u
e
s
a
n
d
c
u
st
o
m
e
r
sat
i
sf
a
c
t
i
o
n
i
m
p
a
c
t
,
”
i
n
2
0
2
2
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
E
n
g
i
n
e
e
ri
n
g
&
MIS
(
I
C
EM
I
S
)
,
I
EEE,
Ju
l
.
2
0
2
2
,
p
p
.
1
–
8
.
d
o
i
:
1
0
.
1
1
0
9
/
I
C
E
M
I
S
5
6
2
9
5
.
2
0
2
2
.
9
9
1
4
3
9
3
.
[
8
]
A
.
S
.
N
a
i
r
a
n
d
P
.
K
a
n
n
a
n
,
“
D
i
g
i
t
a
l
p
a
y
me
n
t
me
t
h
o
d
s:
c
h
a
l
l
e
n
g
e
s
a
n
d
o
p
p
o
r
t
u
n
i
t
i
e
s
,
”
J
o
u
rn
a
l
o
f
N
a
m
i
b
i
a
n
S
t
u
d
i
e
s
,
v
o
l
.
3
7
,
p
p
.
3
6
7
–
3
7
6
,
2
0
2
3
.
[
9
]
Z.
A
l
k
h
a
l
i
l
,
C
.
H
e
w
a
g
e
,
L.
N
a
w
a
f
,
a
n
d
I
.
K
h
a
n
,
“
P
h
i
sh
i
n
g
a
t
t
a
c
k
s
:
a
r
e
c
e
n
t
c
o
mp
r
e
h
e
n
s
i
v
e
st
u
d
y
a
n
d
a
n
e
w
a
n
a
t
o
m
y
,
”
Fro
n
t
i
e
rs
i
n
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
3
,
M
a
r
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
8
9
/
f
c
o
m
p
.
2
0
2
1
.
5
6
3
0
6
0
.
[
1
0
]
R
.
El
Y
o
u
b
i
,
F
.
M
e
ss
a
o
u
d
i
,
M
.
L
o
u
k
i
l
i
,
a
n
d
R
.
L
o
u
k
i
l
i
,
“
F
r
o
m
c
l
i
c
k
t
o
c
h
e
c
k
o
u
t
:
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
r
e
a
l
-
t
i
m
e
f
r
a
u
d
d
e
t
e
c
t
i
o
n
i
n
e
-
p
a
y
me
n
t
s
y
st
e
ms
,
”
S
t
a
t
i
st
i
c
s
,
O
p
t
i
m
i
z
a
t
i
o
n
a
n
d
I
n
f
o
rm
a
t
i
o
n
C
o
m
p
u
t
i
n
g
,
v
o
l
.
1
4
,
n
o
.
6
,
p
p
.
3
3
9
8
–
3
4
0
8
,
S
e
p
.
2
0
2
5
,
d
o
i
:
1
0
.
1
9
1
3
9
/
s
o
i
c
-
2
3
1
0
-
5
0
7
0
-
2
8
9
1
.
[
1
1
]
Eu
r
o
p
e
a
n
P
a
r
l
i
a
m
e
n
t
a
n
d
C
o
u
n
c
i
l
o
f
t
h
e
Eu
r
o
p
e
a
n
U
n
i
o
n
,
“
R
e
g
u
l
a
t
i
o
n
(
EU
)
2
0
2
4
/
1
6
8
9
l
a
y
i
n
g
d
o
w
n
h
a
r
mo
n
i
se
d
r
u
l
e
s
o
n
a
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
,
”
O
f
f
i
c
i
a
l
J
o
u
rn
a
l
o
f
t
h
e
E
u
ro
p
e
a
n
U
n
i
o
n
,
2
0
2
4
,
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s:
/
/
e
u
r
-
l
e
x
.
e
u
r
o
p
a
.
e
u
/
e
l
i
/
r
e
g
/
2
0
2
4
/
1
6
8
9
/
o
j
/
e
ng
[
1
2
]
M
.
L
o
u
k
i
l
i
,
F
.
M
e
ss
a
o
u
d
i
,
a
n
d
H
.
A
z
i
r
a
r
,
“
E
-
p
a
y
m
e
n
t
f
r
a
u
d
d
e
t
e
c
t
i
o
n
i
n
e
-
c
o
mm
e
r
c
e
u
s
i
n
g
s
u
p
e
r
v
i
se
d
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms
,
”
i
n
A
d
v
a
n
c
e
s
i
n
Em
e
r
g
i
n
g
F
i
n
a
n
c
i
a
l
T
e
c
h
n
o
l
o
g
y
a
n
d
D
i
g
i
t
a
l
M
o
n
e
y
,
v
o
l
.
1
7
,
n
o
.
5
,
B
o
c
a
R
a
t
o
n
:
C
R
C
P
r
e
ss,
2
0
2
4
,
p
p
.
2
7
–
3
5
.
d
o
i
:
1
0
.
1
2
0
1
/
9
7
8
1
0
3
2
6
6
7
4
7
8
-
3.
[
1
3
]
A
.
N
e
sv
i
j
e
v
sk
a
i
a
,
S
.
O
u
i
l
l
a
d
e
,
P
.
G
u
i
l
m
i
n
,
a
n
d
J.
D
.
Z
u
c
k
e
r
,
“
T
h
e
a
c
c
u
r
a
c
y
v
e
r
s
u
s
i
n
t
e
r
p
r
e
t
a
b
i
l
i
t
y
t
r
a
d
e
-
o
f
f
i
n
f
r
a
u
d
d
e
t
e
c
t
i
o
n
mo
d
e
l
,
”
D
a
t
a
a
n
d
P
o
l
i
c
y
,
v
o
l
.
3
,
n
o
.
7
,
p
.
e
1
2
,
J
u
l
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
7
/
d
a
p
.
2
0
2
1
.
3
.
[
1
4
]
P
w
C
,
“
EU
’
s
A
I
A
c
t
:
W
h
a
t
r
e
g
u
l
a
t
o
r
s
sh
o
u
l
d
k
n
o
w
1
4
,
”
P
r
i
c
e
w
a
t
e
r
h
o
u
s
e
C
o
o
p
e
r
s
.
A
c
c
e
sse
d
:
M
a
y
2
8
,
2
0
2
6
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
w
w
w
.
p
w
c
.
c
o
m
/
u
s
/
e
n
/
s
e
r
v
i
c
e
s/
c
o
n
su
l
t
i
n
g
/
c
y
b
e
r
sec
u
r
i
t
y
-
r
i
sk
-
r
e
g
u
l
a
t
o
r
y
/
l
i
b
r
a
r
y
/
t
e
c
h
-
r
e
g
u
l
a
t
o
r
y
-
p
o
l
i
c
y
-
d
e
v
e
l
o
p
me
n
t
s/
e
u
-
ai
-
a
c
t
.
h
t
ml
[
1
5
]
R
.
A
v
a
c
h
a
r
ma
l
,
“
Le
v
e
r
a
g
i
n
g
su
p
e
r
v
i
sed
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms
f
o
r
e
n
h
a
n
c
e
d
a
n
o
ma
l
y
d
e
t
e
c
t
i
o
n
i
n
a
n
t
i
-
mo
n
e
y
l
a
u
n
d
e
r
i
n
g
(
A
M
L)
t
r
a
n
sac
t
i
o
n
mo
n
i
t
o
r
i
n
g
s
y
st
e
ms
:
a
c
o
m
p
a
r
a
t
i
v
e
a
n
a
l
y
s
i
s
o
f
p
e
r
f
o
r
man
c
e
a
n
d
e
x
p
l
a
i
n
a
b
i
l
i
t
y
,
”
A
f
ri
c
a
n
J
o
u
rn
a
l
O
f
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
S
u
s
t
a
i
n
a
b
l
e
D
e
v
e
l
o
p
m
e
n
t
,
v
o
l
.
1
,
n
o
.
2
,
p
p
.
6
8
–
8
5
,
2
0
2
1
.
[
1
6
]
M
.
S
a
i
d
i
,
F
.
M
e
ss
a
o
u
d
i
,
R
.
L
o
u
k
i
l
i
,
B
.
E
l
g
h
o
mr
i
,
E.
I
.
A
l
-
W
r
e
i
k
a
t
,
a
n
d
M
.
Lo
u
k
i
l
i
,
“
H
y
b
r
i
d
n
e
u
r
a
l
a
r
c
h
i
t
e
c
t
u
r
e
s
f
o
r
e
-
c
o
mm
e
r
c
e
p
e
r
s
o
n
a
l
i
z
a
t
i
o
n
:
b
a
l
a
n
c
i
n
g
a
c
c
u
r
a
c
y
a
n
d
d
i
v
e
r
s
i
t
y
t
h
r
o
u
g
h
GNN
-
t
r
a
n
sf
o
r
mer
i
n
t
e
g
r
a
t
i
o
n
,
”
i
n
2
0
2
6
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
f
o
r
S
u
st
a
i
n
a
b
l
e
En
g
i
n
e
e
ri
n
g
a
n
d
I
n
n
o
v
a
t
i
o
n
(
AI
S
EI
)
,
I
EEE,
A
p
r
.
2
0
2
6
,
p
p
.
6
9
4
–
6
9
9
.
d
o
i
:
1
0
.
1
1
0
9
/
A
I
S
EI
6
8
6
2
8
.
2
0
2
6
.
1
1
5
7
2
8
1
5
.
[
1
7
]
M
.
L
o
u
k
i
l
i
,
F
.
M
e
ssa
o
u
d
i
,
a
n
d
M
.
E
l
G
h
a
z
i
,
“
P
e
r
s
o
n
a
l
i
z
i
n
g
P
r
o
d
u
c
t
R
e
c
o
mm
e
n
d
a
t
i
o
n
s
u
si
n
g
C
o
l
l
a
b
o
r
a
t
i
v
e
F
i
l
t
e
r
i
n
g
i
n
O
n
l
i
n
e
R
e
t
a
i
l
:
A
M
a
c
h
i
n
e
L
e
a
r
n
i
n
g
A
p
p
r
o
a
c
,
”
i
n
2
0
2
3
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
I
n
f
o
r
m
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
(
I
C
I
T
)
,
I
EEE,
A
u
g
.
2
0
2
3
,
p
p
.
1
9
–
2
4
.
d
o
i
:
1
0
.
1
1
0
9
/
I
C
I
T5
8
0
5
6
.
2
0
2
3
.
1
0
2
2
6
0
4
2
.
[
1
8
]
M
.
A
b
d
u
l
S
a
l
a
m,
K
.
M
.
F
o
u
a
d
,
D
.
L.
E
l
b
a
b
l
y
,
a
n
d
S
.
M
.
El
s
a
y
e
d
,
“
F
e
d
e
r
a
t
e
d
l
e
a
r
n
i
n
g
m
o
d
e
l
f
o
r
c
r
e
d
i
t
c
a
r
d
f
r
a
u
d
d
e
t
e
c
t
i
o
n
w
i
t
h
d
a
t
a
b
a
l
a
n
c
i
n
g
t
e
c
h
n
i
q
u
e
,
”
N
e
u
ra
l
C
o
m
p
u
t
i
n
g
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
3
6
,
n
o
.
1
1
,
p
p
.
6
2
3
1
–
6
2
5
6
,
A
p
r
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
5
2
1
-
023
-
0
9
4
1
0
-
2.
[
1
9
]
P
.
O
.
S
h
o
e
t
a
n
a
n
d
B
.
T
.
F
a
mi
l
o
n
i
,
“
Tr
a
n
sf
o
r
m
i
n
g
f
i
n
t
e
c
h
f
r
a
u
d
d
e
t
e
c
t
i
o
n
w
i
t
h
a
d
v
a
n
c
e
d
a
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
a
l
g
o
r
i
t
h
ms
,
”
Fi
n
a
n
c
e
&
Ac
c
o
u
n
t
i
n
g
R
e
se
a
rc
h
J
o
u
r
n
a
l
,
v
o
l
.
6
,
n
o
.
4
,
p
p
.
6
0
2
–
6
2
5
,
2
0
2
4
,
d
o
i
:
1
0
.
5
1
5
9
4
/
f
a
r
j
.
v
6
i
4
.
1
0
3
6
.
[
2
0
]
M
.
L
o
u
k
i
l
i
,
F
.
M
e
ssa
o
u
d
i
,
a
n
d
M
.
E
l
G
h
a
z
i
,
“
D
e
f
e
n
d
i
n
g
a
g
a
i
n
st
d
i
g
i
t
a
l
t
h
i
e
v
e
r
y
:
a
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
t
o
p
r
e
d
i
c
t
e
-
p
a
y
me
n
t
f
r
a
u
d
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
M
a
n
a
g
e
m
e
n
t
Pra
c
t
i
c
e
,
v
o
l
.
1
7
,
n
o
.
5
,
p
p
.
5
2
2
–
5
3
8
,
2
0
2
4
,
d
o
i
:
1
0
.
1
5
0
4
/
I
JM
P
.
2
0
2
4
.
1
4
0
8
6
1
.
[
2
1
]
P
.
D
u
b
e
y
,
P
.
D
u
b
e
y
,
a
n
d
P
.
N
.
B
o
k
o
r
o
,
“
A
u
n
i
f
i
e
d
t
r
a
n
sf
o
r
mer
–
B
D
I
a
r
c
h
i
t
e
c
t
u
r
e
f
o
r
f
i
n
a
n
c
i
a
l
f
r
a
u
d
d
e
t
e
c
t
i
o
n
:
d
i
st
r
i
b
u
t
e
d
k
n
o
w
l
e
d
g
e
t
r
a
n
sf
e
r
a
c
r
o
ss
d
i
v
e
r
s
e
d
a
t
a
set
s
,”
Fo
r
e
c
a
st
i
n
g
,
v
o
l
.
7
,
n
o
.
2
,
p
.
3
1
,
J
u
n
.
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
f
o
r
e
c
a
st
7
0
2
0
0
3
1
.
[
2
2
]
Y
.
G
o
r
i
s
h
n
i
y
,
I
.
R
u
b
a
c
h
e
v
,
V
.
K
h
r
u
l
k
o
v
,
a
n
d
A
.
B
a
b
e
n
k
o
,
“
R
e
v
i
si
t
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
s
f
o
r
t
a
b
u
l
a
r
d
a
t
a
,
”
i
n
A
d
v
a
n
c
e
s
i
n
N
e
u
ra
l
I
n
f
o
rm
a
t
i
o
n
Pr
o
c
e
ssi
n
g
S
y
s
t
e
m
s
,
2
0
2
1
,
p
p
.
1
8
9
3
2
–
1
8
9
4
3
.
[
2
3
]
I
.
D
.
M
i
e
n
y
e
a
n
d
N
.
Jere,
“
D
e
e
p
l
e
a
r
n
i
n
g
f
o
r
c
r
e
d
i
t
c
a
r
d
f
r
a
u
d
d
e
t
e
c
t
i
o
n
:
a
r
e
v
i
e
w
o
f
a
l
g
o
r
i
t
h
ms,
c
h
a
l
l
e
n
g
e
s,
a
n
d
s
o
l
u
t
i
o
n
s
,”
I
EEE
A
c
c
e
ss
,
v
o
l
.
1
2
,
p
p
.
9
6
8
9
3
–
9
6
9
1
0
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
4
.
3
4
2
6
9
5
5
.
[
2
4
]
J.
G
e
,
L
.
Y
i
n
,
S
.
Z
h
a
n
g
,
a
n
d
X
.
Z
h
a
o
,
“
G
a
t
e
d
a
t
t
e
n
t
i
o
n
b
a
se
d
g
e
n
e
r
a
t
i
v
e
a
d
v
e
r
sar
i
a
l
n
e
t
w
o
r
k
s
f
o
r
i
m
b
a
l
a
n
c
e
d
c
r
e
d
i
t
c
a
r
d
f
r
a
u
d
d
e
t
e
c
t
i
o
n
,
”
Pe
e
rJ
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
1
1
,
p
.
e
2
9
7
2
,
Ju
n
.
2
0
2
5
,
d
o
i
:
1
0
.
7
7
1
7
/
p
e
e
r
j
-
c
s
.
2
9
7
2
.
[
2
5
]
Y
.
P
u
s
h
k
a
r
e
n
k
o
a
n
d
V
.
Za
s
l
a
v
sk
y
i
,
“
S
y
n
t
h
e
t
i
c
d
a
t
a
g
e
n
e
r
a
t
i
o
n
f
o
r
f
r
a
u
d
d
e
t
e
c
t
i
o
n
u
s
i
n
g
d
i
f
f
u
si
o
n
m
o
d
e
l
s
,
”
I
n
f
o
rm
a
t
i
o
n
&
S
e
c
u
r
i
t
y
:
A
n
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
,
v
o
l
.
5
5
,
n
o
.
2
,
p
p
.
1
8
5
–
1
9
8
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
6
1
0
/
i
si
j
.
5
5
3
4
.
[
2
6
]
R
.
R
o
y
,
D
.
Ti
w
a
r
i
,
a
n
d
A
.
P
a
n
d
e
y
,
“
F
r
a
u
d
D
i
f
f
u
s
e
:
d
i
f
f
u
si
o
n
-
a
i
d
e
d
s
y
n
t
h
e
t
i
c
f
r
a
u
d
a
u
g
m
e
n
t
a
t
i
o
n
f
o
r
i
m
p
r
o
v
e
d
f
r
a
u
d
d
e
t
e
c
t
i
o
n
,”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
5
t
h
AC
M
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
AI
i
n
F
i
n
a
n
c
e
,
N
e
w
Y
o
r
k
,
N
Y
,
U
S
A
:
A
C
M
,
N
o
v
.
2
0
2
4
,
p
p
.
9
0
–
9
8
.
d
o
i
:
1
0
.
1
1
4
5
/
3
6
7
7
0
5
2
.
3
6
9
8
6
5
8
.
[
2
7
]
B
.
Li
m,
S
.
A
r
ı
k
,
N
.
L
o
e
f
f
,
a
n
d
T.
P
f
i
st
e
r
,
“
Te
m
p
o
r
a
l
f
u
si
o
n
t
r
a
n
sf
o
r
mers
f
o
r
i
n
t
e
r
p
r
e
t
a
b
l
e
m
u
l
t
i
-
h
o
r
i
z
o
n
t
i
me
se
r
i
e
s
f
o
r
e
c
a
s
t
i
n
g
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
F
o
re
c
a
s
t
i
n
g
,
v
o
l
.
3
7
,
n
o
.
4
,
p
p
.
1
7
4
8
–
1
7
6
4
,
O
c
t
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
j
f
o
r
e
c
a
st
.
2
0
2
1
.
0
3
.
0
1
2
.
[
2
8
]
U
.
Za
f
a
r
a
n
d
F
.
W
u
,
“
M
e
t
h
o
d
o
l
o
g
i
c
a
l
c
h
a
l
l
e
n
g
e
s
i
n
e
x
p
l
a
i
n
a
b
l
e
A
I
f
o
r
f
r
a
u
d
d
e
t
e
c
t
i
o
n
:
a
s
y
s
t
e
m
a
t
i
c
l
i
t
e
r
a
t
u
r
e
r
e
v
i
e
w
,
”
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
Re
v
i
e
w
,
v
o
l
.
5
9
,
n
o
.
4
,
p
.
1
1
5
,
F
e
b
.
2
0
2
6
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
4
6
2
-
0
2
6
-
1
1
5
1
6
-
7.
[
2
9
]
C
.
J
u
,
X
.
M
a
,
a
n
d
B
.
D
o
n
g
,
“
R
e
a
l
-
t
i
me
c
r
o
ss
-
b
o
r
d
e
r
p
a
y
me
n
t
f
r
a
u
d
d
e
t
e
c
t
i
o
n
u
s
i
n
g
t
e
mp
o
r
a
l
g
r
a
p
h
n
e
u
r
a
l
n
e
t
w
o
r
k
s:
a
d
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
,”
A
c
a
d
e
m
i
c
J
o
u
rn
a
l
o
f
S
o
c
i
o
l
o
g
y
a
n
d
M
a
n
a
g
e
m
e
n
t
,
v
o
l
.
3
,
n
o
.
2
,
p
p
.
1
–
1
2
,
M
a
r
.
2
0
2
5
,
d
o
i
:
1
0
.
7
0
3
9
3
/
6
1
6
a
7
3
6
d
.
3
2
3
6
3
9
.
[
3
0
]
S
.
K
.
A
l
j
u
n
a
i
d
,
S
.
J
.
A
l
m
h
e
i
r
i
,
H
.
D
a
w
o
o
d
,
a
n
d
M
.
A
.
K
h
a
n
,
“
S
e
c
u
r
e
a
n
d
Tr
a
n
s
p
a
r
e
n
t
B
a
n
k
i
n
g
:
Ex
p
l
a
i
n
a
b
l
e
A
I
-
d
r
i
v
e
n
f
e
d
e
r
a
t
e
d
l
e
a
r
n
i
n
g
mo
d
e
l
f
o
r
f
i
n
a
n
c
i
a
l
f
r
a
u
d
d
e
t
e
c
t
i
o
n
,
”
J
o
u
r
n
a
l
o
f
R
i
sk
a
n
d
F
i
n
a
n
c
i
a
l
M
a
n
a
g
e
m
e
n
t
,
v
o
l
.
1
8
,
n
o
.
4
,
p
.
1
7
9
,
M
a
r
.
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
j
r
f
m1
8
0
4
0
1
7
9
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
4
,
D
e
c
em
b
er
20
2
6
:
1
7
6
8
-
1
7
7
6
1776
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Ria
d
L
o
u
k
i
li
is
a
so
ftwa
re
e
n
g
in
e
e
r
a
n
d
re
se
a
rc
h
e
r
sp
e
c
ializin
g
i
n
AI,
De
v
Op
s,
a
n
d
M
LO
p
s
wit
h
p
r
o
fe
ss
io
n
a
l
e
x
p
e
rien
c
e
in
4
c
o
u
n
tr
ies
a
c
ro
ss
3
c
o
n
ti
n
e
n
ts.
Va
led
ict
o
rian
in
AI
a
n
d
Da
ta
S
c
ien
c
e
,
fo
c
u
se
d
o
n
M
L,
c
l
o
u
d
c
o
m
p
u
ti
n
g
,
a
n
d
o
p
t
imiz
a
ti
o
n
.
He
is
c
u
rre
n
tl
y
a
Ph
.
D
.
stu
d
e
n
t
at
S
i
d
i
M
o
h
a
m
e
d
Be
n
Ab
d
e
ll
a
h
U
n
iv
e
rsit
y
o
f
F
e
z
in
t
h
e
Artifi
c
ial
In
tell
ig
e
n
c
e
,
Da
ta
S
c
ien
c
e
s
a
n
d
S
y
ste
m
s
Lab
o
ra
to
ry
.
His
m
a
in
P
h
.
D
.
re
se
a
rc
h
in
tere
sts
fo
c
u
s
o
n
Artif
icia
l
In
telli
g
e
n
c
e
with
in
th
e
c
o
n
tex
t
o
f
E
-
Co
m
m
e
rc
e
a
n
d
E
-
M
a
rk
e
ti
n
g
stra
teg
ies
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
ria
d
.
lo
u
k
il
i@
u
sm
b
a
.
a
c
.
m
a
.
Pro
f.
F
a
y
ç
a
l
Mes
sa
o
u
d
i
is
a
F
u
ll
P
ro
fe
ss
o
r
a
t
t
h
e
Na
ti
o
n
a
l
S
c
h
o
o
l
o
f
Bu
si
n
e
ss
a
n
d
M
a
n
a
g
e
m
e
n
t,
S
i
d
i
M
o
h
a
m
e
d
Be
n
Ab
d
e
ll
a
h
U
n
iv
e
rsit
y
,
F
e
z
,
M
o
ro
c
c
o
.
He
is
a
m
e
m
b
e
r
o
f
th
e
Re
se
a
rc
h
Lab
o
ra
to
ry
Artifi
c
i
a
l
In
telli
g
e
n
c
e
,
Da
ta
S
c
ien
c
e
,
a
n
d
Eme
rg
in
g
S
y
ste
m
s.
His
m
a
in
tea
c
h
in
g
a
n
d
re
se
a
rc
h
in
te
re
sts
c
o
n
c
e
rn
a
rti
ficia
l
i
n
telli
g
e
n
c
e
,
d
a
ta
a
n
a
ly
sis,
d
a
tab
a
se
m
a
n
a
g
e
m
e
n
t,
a
n
d
e
-
m
a
rk
e
ti
n
g
.
P
ro
f.
M
e
ss
a
o
u
d
i
is
a
lso
d
e
e
p
l
y
i
n
v
o
lv
e
d
i
n
se
v
e
ra
l
re
se
a
rc
h
p
ro
jec
ts
a
n
d
c
o
n
s
u
lt
a
n
c
y
in
t
h
e
fi
e
ld
o
f
d
ig
i
t
a
l
m
a
rk
e
ti
n
g
stra
teg
ies
a
n
d
b
u
sin
e
ss
a
n
a
l
y
ti
c
s
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
fa
y
c
a
l.
m
e
ss
a
o
u
d
i@u
sm
b
a
.
a
c
.
m
a
.
Ra
o
u
y
a
El
Yo
u
b
i
is
a
Do
c
t
o
r
in
S
i
d
i
M
o
h
a
m
e
d
Be
n
A
b
d
e
ll
a
h
u
n
i
v
e
rsity
o
f
F
e
z
fro
m
th
e
Artifi
c
ial
In
telli
g
e
n
c
e
,
Da
ta
S
c
ien
c
e
s
a
n
d
S
y
ste
m
s
Lab
o
ra
to
ry
.
S
h
e
is
a
lso
a
n
I
T
e
n
g
in
e
e
r
w
h
o
g
ra
d
u
a
ted
fro
m
t
h
e
Na
ti
o
n
a
l
S
c
h
o
o
l
o
f
Ap
p
li
e
d
S
c
ien
c
e
s
o
f
F
e
z
.
He
r
m
a
in
re
se
a
rc
h
in
tere
sts
fo
c
us
on
a
rti
fic
ial
in
telli
g
e
n
c
e
,
m
a
c
h
in
e
lea
rn
in
g
,
a
n
d
d
e
e
p
lea
rn
i
n
g
wi
th
i
n
th
e
c
o
n
tex
t
o
f
e
-
c
o
m
m
e
rc
e
a
n
d
d
i
g
it
a
l
m
a
rk
e
ti
n
g
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
ra
o
u
y
a
.
e
l
y
o
u
b
i@
u
sm
b
a
.
a
c
.
m
a
.
Pro
f.
M
a
n
a
l
Lo
u
k
il
i
is
a
P
ro
fe
ss
o
r
a
n
d
re
se
a
rc
h
e
r
a
t
th
e
N
a
ti
o
n
a
l
S
c
h
o
o
l
o
f
Bu
sin
e
ss
a
n
d
M
a
n
a
g
e
m
e
n
t,
S
i
d
i
M
o
h
a
m
e
d
Be
n
Ab
d
e
ll
a
h
Un
iv
e
rs
it
y
i
n
F
e
z
,
M
o
r
o
c
c
o
,
w
h
e
re
sh
e
a
lso
o
b
tain
e
d
h
e
r
d
o
c
to
ra
te.
S
p
e
c
ializin
g
i
n
a
rti
ficia
l
in
telli
g
e
n
c
e
,
m
a
c
h
in
e
lea
rn
in
g
,
a
n
d
d
ig
it
a
l
m
a
rk
e
ti
n
g
,
s
h
e
c
o
n
tri
b
u
te
s
e
x
ten
siv
e
l
y
t
o
t
h
e
a
rti
ficia
l
i
n
t
e
ll
ig
e
n
c
e
,
d
a
ta
sc
ien
c
e
a
n
d
e
m
e
rg
in
g
s
y
ste
m
s
lab
o
ra
t
o
r
y
.
P
r
o
f.
L
o
u
k
il
i
h
a
s
p
u
b
li
sh
e
d
n
u
m
e
ro
u
s
a
rti
c
les
in
th
e
se
a
re
a
s
a
n
d
is
a
n
a
c
ti
v
e
p
a
rti
c
ip
a
n
t
in
in
tern
a
ti
o
n
a
l
c
o
n
fe
re
n
c
e
s
d
e
d
ica
ted
to
AI
a
n
d
it
s
a
p
p
li
c
a
ti
o
n
s
i
n
m
a
rk
e
ti
n
g
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
m
a
n
a
l.
lo
u
k
il
i@u
sm
b
a
.
a
c
.
m
a
.
Evaluation Warning : The document was created with Spire.PDF for Python.