I
nte
rna
t
io
na
l J
o
urna
l o
f
Rec
o
nfig
ura
ble a
nd
E
m
be
dd
e
d Sy
s
t
e
m
s
(
I
J
R
E
S)
Vo
l.
1
5
,
No
.
2
,
J
u
ly
20
2
6
,
p
p
.
439
~
4
4
9
I
SS
N:
2089
-
4864
,
DOI
:
1
0
.
1
1
5
9
1
/i
j
r
es
.
v
1
5
.
i
2
.
pp
4
3
9
-
449
439
J
o
ur
na
l
ho
m
ep
a
g
e
:
h
ttp
:
//ij
r
es.ia
esco
r
e.
co
m
G
ra
ph n
eura
l ne
t
w
o
rk
-
ba
sed bio
medica
l
m
isinfo
r
ma
tion
detec
tion w
ith
sem
a
nti
c cons
isten
c
y
ana
ly
sis
Siv
a
Dhiev
a
ra
j
,
Ag
us
t
hiy
a
r
Ra
m
u
D
e
p
a
r
t
me
n
t
o
f
C
o
mp
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,
S
R
M
I
n
st
i
t
u
t
e
o
f
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
R
a
ma
p
u
r
a
m C
a
m
p
u
s
,
C
h
e
n
n
a
i
,
I
n
d
i
a
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
No
v
2
8
,
2
0
2
5
R
ev
i
s
ed
Ma
r
2
4
,
2
0
2
6
A
cc
ep
ted
Ma
y
3
1
,
2
0
2
6
Co
n
v
e
n
ti
o
n
a
l
m
isin
f
o
r
m
a
ti
o
n
d
e
tec
ti
o
n
a
p
p
r
o
a
c
h
e
s
p
rim
a
ril
y
r
e
l
y
o
n
tex
tu
a
l
f
e
a
tu
re
s
a
n
d
d
e
e
p
lea
rn
in
g
(DL
)
c
las
sif
i
e
rs,
w
h
ich
o
f
ten
fa
il
to
c
a
p
tu
re
c
o
m
p
lex
r
e
latio
n
sh
i
p
s
a
m
o
n
g
b
io
m
e
d
ica
l
e
n
ti
ti
e
s
a
n
d
th
e
u
n
d
e
rly
in
g
sc
ien
ti
f
ic
c
o
n
tex
t
o
f
h
e
a
lt
h
c
laim
s.
T
o
a
d
d
re
ss
th
is
l
im
it
a
ti
o
n
,
t
h
is
st
u
d
y
p
ro
p
o
se
s
a
g
ra
p
h
n
e
u
ra
l
n
e
tw
o
rk
(
G
NN
)
-
b
a
se
d
b
io
m
e
d
ica
l
m
isin
f
o
r
m
a
ti
o
n
d
e
tec
ti
o
n
f
ra
m
e
w
o
rk
th
a
t
in
teg
ra
tes
k
n
o
w
le
d
g
e
g
ra
p
h
p
r
o
p
a
g
a
ti
o
n
w
it
h
se
m
a
n
ti
c
c
o
n
siste
n
c
y
v
e
rif
i
c
a
ti
o
n
.
In
it
ially
,
k
e
y
b
io
m
e
d
ica
l
e
n
ti
ti
e
s
su
c
h
a
s
d
ise
a
se
s,
trea
t
m
e
n
ts,
a
n
d
b
i
o
lo
g
ica
l
p
ro
c
e
ss
e
s
a
re
e
x
tr
a
c
ted
a
n
d
m
a
p
p
e
d
i
n
to
a
stru
c
tu
re
d
b
i
o
m
e
d
ica
l
k
n
o
w
led
g
e
g
ra
p
h
(BKG
)
to
re
p
re
se
n
t
se
m
a
n
ti
c
re
lati
o
n
sh
i
p
s.
A
g
ra
p
h
a
tt
e
n
ti
o
n
n
e
tw
o
rk
(
GA
T
)
is
th
e
n
e
m
p
lo
y
e
d
t
o
m
o
d
e
l
re
latio
n
a
l
d
e
p
e
n
d
e
n
c
ies
a
n
d
p
ro
p
a
g
a
te
c
o
n
t
e
x
tu
a
l
in
f
o
rm
a
ti
o
n
a
c
ro
ss
c
o
n
n
e
c
ted
e
n
ti
ti
e
s,
e
n
a
b
li
n
g
th
e
d
e
tec
ti
o
n
o
f
h
i
d
d
e
n
in
c
o
n
sis
ten
c
ies
in
b
io
m
e
d
ica
l
c
laim
s.
T
h
e
p
ro
p
o
s
e
d
m
o
d
e
l
is
e
v
a
lu
a
te
d
u
sin
g
b
e
n
c
h
m
a
r
k
b
io
m
e
d
ica
l
m
isin
f
o
r
m
a
ti
o
n
d
a
tas
e
ts,
in
c
lu
d
i
n
g
Re
li
a
b
le
COV
ID
-
1
9
Ne
w
s
Da
tas
e
t,
2
0
2
1
(Re
COV
e
r
y)
,
CO
V
ID
-
1
9
He
a
lt
h
c
a
re
M
isin
f
o
rm
a
ti
o
n
Da
tas
e
t,
2
0
2
0
(Co
A
ID)
,
a
n
d
2
0
1
8
–
2
0
2
0
b
i
o
m
e
d
ica
l
h
e
a
lt
h
n
e
ws
c
o
rp
u
s
(He
a
lt
h
S
t
o
ry
)
.
Ex
p
e
rime
n
tal
re
su
lt
s
d
e
m
o
n
stra
te
th
a
t
th
e
p
ro
p
o
se
d
f
ra
m
e
w
o
rk
a
c
h
iev
e
s
a
n
a
v
e
ra
g
e
d
e
te
c
ti
o
n
a
c
c
u
ra
c
y
o
f
9
6
.
3
%
,
o
u
t
p
e
rf
o
rm
in
g
c
o
n
v
e
n
ti
o
n
a
l
l
o
n
g
sh
o
rt
-
term
m
e
m
o
r
y
(
L
S
T
M
)
,
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
s
(
C
NN
)
,
a
n
d
tran
sf
o
r
m
e
r
-
b
a
se
d
m
o
d
e
ls
in
term
s
o
f
p
re
c
isio
n
,
re
c
a
ll
,
a
n
d
F
1
-
sc
o
re
.
T
h
e
f
in
d
in
g
s
h
ig
h
li
g
h
t
t
h
a
t
in
teg
ra
ti
n
g
stru
c
tu
re
d
b
io
m
e
d
ica
l
k
n
o
w
l
e
d
g
e
w
it
h
g
ra
p
h
-
b
a
se
d
re
a
so
n
in
g
sig
n
if
ica
n
tl
y
e
n
h
a
n
c
e
s
th
e
re
li
a
b
il
it
y
a
n
d
in
terp
re
tab
il
it
y
o
f
m
isin
f
o
r
m
a
ti
o
n
d
e
tec
ti
o
n
sy
ste
m
s.
K
ey
w
o
r
d
s
:
B
io
m
ed
ical
n
e
w
s
C
o
n
v
o
lu
tio
n
al
n
e
u
r
al
n
e
t
w
o
r
k
Dete
cti
on
Gr
ap
h
n
eu
r
al
n
et
w
o
r
k
Ma
ch
i
n
e
lear
n
i
n
g
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
:
Siv
a
D
h
ie
v
ar
aj
Dep
ar
t
m
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
an
d
A
p
p
licatio
n
s
,
SR
M
I
n
s
tit
u
te
o
f
Scie
n
ce
an
d
T
ec
h
n
o
l
o
g
y
R
a
m
ap
u
r
a
m
C
a
m
p
u
s
C
h
e
n
n
ai,
T
am
il Na
d
u
,
I
n
d
ia
E
m
ail:
s
iv
ad
@
s
r
m
is
t.e
d
u
.
in
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
cir
cu
latio
n
o
f
b
io
m
ed
ical
in
f
o
r
m
a
tio
n
th
r
o
u
g
h
o
n
li
n
e
p
latf
o
r
m
s
h
as
g
r
o
w
n
at
an
u
n
p
r
ec
ed
en
ted
r
ate,
lar
g
el
y
d
r
iv
en
b
y
t
h
e
p
u
b
lic’
s
i
n
cr
ea
s
i
n
g
r
elian
ce
o
n
d
ig
ital
s
o
u
r
ce
s
f
o
r
h
ea
lt
h
ad
v
ice,
tr
ea
t
m
en
t
r
ec
o
m
m
e
n
d
atio
n
s
,
a
n
d
s
cie
n
ti
f
ic
u
p
d
ates.
W
h
ile
t
h
i
s
r
a
p
id
d
is
s
e
m
i
n
atio
n
o
f
k
n
o
w
led
g
e
h
a
s
en
co
u
r
ag
ed
g
r
ea
ter
a
w
ar
e
n
ess
a
n
d
ac
ce
s
s
ib
ilit
y
,
it
h
as
also
o
p
en
ed
th
e
d
o
o
r
to
a
p
ar
allel
r
i
s
e
in
m
is
in
f
o
r
m
at
io
n
—
r
an
g
i
n
g
f
r
o
m
ex
ag
g
er
ated
clai
m
s
ab
o
u
t
m
ir
a
cle
cu
r
es
to
d
elib
er
atel
y
f
ab
r
ic
ated
n
ar
r
ativ
es
tar
g
eti
n
g
v
ac
ci
n
es
,
ep
id
e
m
ic
s
,
an
d
clin
ical
r
e
s
ea
r
ch
[
1
]
–
[
4
]
.
T
h
e
co
n
s
eq
u
e
n
ce
s
o
f
s
u
c
h
m
is
in
f
o
r
m
at
io
n
e
x
te
n
d
f
ar
b
e
y
o
n
d
co
n
f
u
s
io
n
o
r
p
u
b
lic
p
an
ic;
t
h
e
y
ca
n
d
ir
ec
tl
y
in
f
l
u
e
n
ce
i
n
d
iv
id
u
al
h
ea
lt
h
c
h
o
ices,
u
n
d
er
m
i
n
e
m
ed
ical
g
u
id
elin
e
s
,
f
u
e
l
r
e
s
is
ta
n
ce
to
p
u
b
lic
h
ea
lth
i
n
ter
v
e
n
tio
n
s
,
an
d
co
n
tr
ib
u
te
to
w
id
esp
r
ea
d
m
i
s
tr
u
s
t
in
s
cie
n
ti
f
ic
in
s
tit
u
tio
n
s
.
T
h
is
is
s
u
e
b
ec
a
m
e
p
ar
ticu
lar
l
y
v
i
s
ib
le
d
u
r
in
g
g
l
o
b
al
h
ea
lth
cr
is
es,
w
h
er
e
m
i
s
lead
in
g
n
e
w
s
o
f
te
n
s
p
r
ea
d
f
aster
th
an
v
er
i
f
ied
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
9
-
4864
I
n
t J
R
ec
o
n
f
i
g
u
r
ab
le
&
E
m
b
ed
d
ed
Sy
s
t
,
Vo
l.
1
5
,
No
.
2
,
J
u
l
y
202
6
:
439
-
4
4
9
440
s
cien
t
if
ic
f
i
n
d
i
n
g
s
.
A
lt
h
o
u
g
h
r
esear
ch
er
s
h
av
e
p
r
o
p
o
s
ed
n
u
m
er
o
u
s
m
ac
h
i
n
e
lear
n
in
g
(
M
L
)
a
n
d
n
at
u
r
al
la
n
g
u
a
g
e
p
r
o
ce
s
s
in
g
(
N
L
P
)
ap
p
r
o
ac
h
es
to
tack
le
f
a
k
e
n
e
w
s
d
etec
t
io
n
,
m
o
s
t
ex
is
tin
g
m
o
d
els
tr
ea
t
m
i
s
i
n
f
o
r
m
atio
n
p
r
im
ar
il
y
as
a
te
x
t
u
al
clas
s
i
f
icatio
n
p
r
o
b
lem
[
5
]
–
[
8
]
.
T
h
ey
r
el
y
h
ea
v
i
l
y
o
n
s
u
r
f
ac
e
-
le
v
el
li
n
g
u
i
s
tic
cu
e
s
,
s
tatis
t
ical
p
atter
n
s
,
an
d
co
n
tex
tu
a
l
e
m
b
ed
d
in
g
s
,
y
et
th
e
y
o
f
te
n
f
ail
to
u
n
d
er
s
t
a
n
d
w
h
e
th
er
th
e
b
io
m
ed
ical
s
tate
m
en
t
s
in
a
te
x
t a
r
e
f
ac
t
u
al
l
y
g
r
o
u
n
d
ed
o
r
s
cien
ti
f
icall
y
p
lau
s
ib
le.
B
io
m
ed
ical
l
an
g
u
ag
e
ca
r
r
ie
s
u
n
iq
u
e
co
m
p
lex
i
ties
—
s
p
ec
i
alize
d
ter
m
i
n
o
lo
g
y
,
n
e
s
ted
r
elatio
n
s
h
ip
s
b
et
w
ee
n
e
n
titi
e
s
,
an
d
clai
m
s
t
h
at
f
r
eq
u
e
n
tl
y
d
ep
en
d
o
n
d
o
m
ai
n
-
s
p
ec
if
ic
k
n
o
w
led
g
e
to
i
n
ter
p
r
et
co
r
r
ec
tly
.
A
s
en
te
n
ce
t
h
at
ap
p
ea
r
s
li
n
g
u
i
s
ti
ca
ll
y
co
h
er
e
n
t
m
a
y
s
ti
ll
co
n
v
e
y
f
als
e
i
n
f
o
r
m
atio
n
i
f
it
s
b
io
m
ed
ical
r
elatio
n
s
h
ip
s
co
n
tr
ad
ict
estab
lis
h
ed
s
cie
n
ti
f
ic
u
n
d
er
s
ta
n
d
i
n
g
.
T
h
is
m
i
s
m
a
tch
ex
p
o
s
e
s
a
cr
itical
w
ea
k
n
e
s
s
i
n
co
n
v
en
tio
n
al
d
ee
p
lear
n
in
g
(
D
L
)
m
o
d
el
s
,
in
clu
d
i
n
g
r
ec
u
r
r
en
t
ar
c
h
itect
u
r
es
an
d
tr
an
s
f
o
r
m
er
-
b
ased
s
y
s
t
e
m
s
,
w
h
ic
h
ex
ce
l
a
t
ca
p
tu
r
in
g
tex
t
u
al
p
atter
n
s
b
u
t
s
tr
u
g
g
le
to
r
ea
s
o
n
o
v
er
in
ter
co
n
n
ec
ted
s
cien
ti
f
ic
co
n
ce
p
ts
[
9
]
–
[
1
2
]
.
A
s
b
io
m
ed
ical
k
n
o
w
led
g
e
i
s
i
n
h
e
r
en
tl
y
r
elatio
n
al
—
li
n
k
i
n
g
d
is
e
ases
to
s
y
m
p
to
m
s
,
tr
ea
t
m
en
t
s
to
o
u
tco
m
es,
g
e
n
es
to
p
ath
w
a
y
s
—
it
s
ac
cu
r
ate
i
n
te
r
p
r
etatio
n
r
eq
u
ir
es
m
o
r
e
th
a
n
tex
t
u
al
f
lu
e
n
c
y
;
it
d
e
m
a
n
d
s
s
t
r
u
ctu
r
ed
r
ea
s
o
n
i
n
g
[
1
3
]
–
[
1
6
]
.
Mo
tiv
ated
b
y
th
is
g
ap
,
r
ec
en
t
r
esear
ch
h
as
b
eg
u
n
ex
p
lo
r
in
g
g
r
ap
h
-
b
ased
r
ep
r
e
s
en
tat
io
n
s
to
h
a
n
d
le
d
o
m
ai
n
-
s
p
ec
i
f
ic
co
m
p
le
x
it
y
.
Gr
ap
h
n
e
u
r
al
n
et
w
o
r
k
s
(
GNN
s
)
,
in
p
ar
ticu
lar
,
h
a
v
e
e
m
er
g
ed
a
s
a
p
o
w
er
f
u
l
to
o
l
f
o
r
m
o
d
elin
g
r
elatio
n
al
d
a
ta,
as
th
e
y
e
n
ab
le
th
e
p
r
o
p
ag
atio
n
o
f
i
n
f
o
r
m
atio
n
ac
r
o
s
s
i
n
ter
co
n
n
ec
ted
n
o
d
es
a
n
d
allo
w
a
m
o
d
el
to
lear
n
h
o
w
m
ea
n
i
n
g
i
s
s
h
ap
ed
n
o
t
j
u
s
t
b
y
i
n
d
i
v
id
u
al
co
n
ce
p
t
s
,
b
u
t
b
y
th
e
ir
r
elatio
n
s
h
ip
s
.
I
n
t
h
e
co
n
tex
t
o
f
b
io
m
ed
ical
m
is
in
f
o
r
m
at
io
n
,
th
i
s
ca
p
ab
ilit
y
b
ec
o
m
e
s
esp
ec
iall
y
v
alu
a
b
le
b
ec
au
s
e
m
a
n
y
m
is
lead
in
g
clai
m
s
h
in
g
e
o
n
in
co
r
r
ec
tl
y
p
air
i
n
g
b
io
m
ed
ica
l
en
ti
ties
,
ci
tin
g
i
m
p
la
u
s
ib
le
ca
u
s
al
co
n
n
ec
tio
n
s
,
o
r
d
r
a
w
i
n
g
co
n
cl
u
s
io
n
s
th
at
co
n
f
lic
t
w
it
h
e
s
tab
lis
h
ed
s
cie
n
t
if
ic
e
v
id
en
ce
.
B
y
m
ap
p
in
g
b
io
m
ed
ical
e
n
titi
e
s
i
n
to
a
s
t
r
u
ct
u
r
ed
k
n
o
w
led
g
e
g
r
ap
h
an
d
en
ab
li
n
g
t
h
e
m
o
d
el
to
r
ea
s
o
n
o
v
er
th
e
s
e
co
n
n
ec
tio
n
s
,
it b
ec
o
m
e
s
p
o
s
s
ib
le
to
ca
p
tu
r
e
i
n
c
o
n
s
is
ten
cie
s
t
h
at
a
p
u
r
el
y
te
x
t
-
d
r
iv
e
n
m
o
d
el
w
o
u
l
d
o
v
er
lo
o
k
.
T
h
e
id
ea
o
f
s
e
m
a
n
tic
co
n
s
i
s
te
n
c
y
ch
ec
k
i
n
g
ad
d
s
an
o
th
er
d
i
m
en
s
io
n
to
th
is
ap
p
r
o
a
ch
b
y
co
m
p
ar
i
n
g
th
e
co
n
te
n
t
o
f
a
n
e
w
s
c
lai
m
a
g
ai
n
s
t
v
er
i
f
ied
b
io
m
ed
ic
al
s
o
u
r
ce
s
,
s
u
c
h
a
s
p
ee
r
-
r
ev
ie
w
ed
s
t
u
d
ies
o
r
cu
r
ated
m
ed
ical
d
atab
ases
.
I
n
s
tead
o
f
r
el
y
in
g
s
o
lel
y
o
n
lin
g
u
i
s
tic
s
i
g
n
als,
th
e
s
y
s
te
m
a
s
s
e
s
s
e
s
w
h
e
th
er
a
clai
m
al
ig
n
s
w
ith
k
n
o
w
n
b
io
m
ed
ical
f
ac
t
s
o
r
d
ev
iate
s
f
r
o
m
estab
lis
h
ed
r
elatio
n
s
h
ip
s
[
1
7
]
–
[
2
2
]
.
T
h
is
in
teg
r
ated
v
ie
w
—
co
m
b
i
n
i
n
g
g
r
ap
h
r
ea
s
o
n
in
g
w
i
th
ev
id
en
ce
-
a
w
ar
e
v
alid
atio
n
—
o
f
f
er
s
a
m
o
r
e
r
o
b
u
s
t
f
o
u
n
d
ati
o
n
f
o
r
e
v
alu
a
tin
g
t
h
e
tr
u
s
t
w
o
r
th
i
n
es
s
o
f
b
io
m
ed
ical
n
e
w
s
.
T
h
e
o
b
j
ec
tiv
e
o
f
t
h
i
s
r
esear
ch
is
to
d
esi
g
n
an
d
e
v
al
u
ate
a
g
r
ap
h
-
b
ased
f
r
a
m
e
w
o
r
k
t
h
at
ca
n
e
f
f
ec
tiv
e
l
y
id
en
ti
f
y
f
r
a
u
d
u
le
n
t
b
io
m
ed
ical
n
e
w
s
b
y
le
v
er
ag
i
n
g
b
o
th
th
e
s
t
r
u
ctu
r
al
p
r
o
p
er
ties
o
f
a
b
io
m
e
d
ical
k
n
o
w
led
g
e
g
r
ap
h
(
B
KG)
an
d
th
e
co
n
tex
t
u
al
u
n
d
er
s
ta
n
d
in
g
p
r
o
v
id
ed
b
y
s
e
m
an
tic
s
i
m
i
lar
it
y
m
ea
s
u
r
es.
U
n
li
k
e
p
r
ev
io
u
s
m
o
d
el
s
th
a
t tr
e
at
m
i
s
in
f
o
r
m
a
tio
n
a
s
a
s
tan
d
alo
n
e
tex
t
cla
s
s
i
f
icat
io
n
tas
k
,
t
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
e
m
b
ed
s
ea
c
h
clai
m
w
it
h
i
n
a
w
id
er
b
io
m
ed
ical
co
n
tex
t,
a
llo
w
in
g
t
h
e
s
y
s
te
m
t
o
d
etec
t
l
o
g
ical
i
n
co
n
s
is
te
n
cie
s
,
i
m
p
la
u
s
ib
le
co
n
ce
p
t
as
s
o
cia
tio
n
s
,
a
n
d
s
e
m
a
n
tic
co
n
tr
ad
ictio
n
s
t
h
at
t
y
p
ica
l N
L
P
m
o
d
els
f
ail
to
ca
p
tu
r
e
[
2
3
]
–
[
2
5
]
.
I
n
ad
d
itio
n
to
im
p
r
o
v
in
g
m
is
i
n
f
o
r
m
atio
n
d
etec
tio
n
ac
cu
r
ac
y
,
th
e
p
r
o
p
o
s
ed
f
r
a
m
e
w
o
r
k
is
d
es
ig
n
ed
w
it
h
s
ca
lab
ilit
y
a
n
d
s
y
s
te
m
-
le
v
el
d
ep
lo
y
m
e
n
t
co
n
s
id
er
atio
n
s
in
m
i
n
d
.
Mo
d
er
n
d
ig
ital
h
ea
lth
m
o
n
ito
r
in
g
en
v
ir
o
n
m
e
n
t
s
i
n
cr
ea
s
i
n
g
l
y
r
el
y
o
n
i
n
tell
ig
e
n
t
co
m
p
u
tat
io
n
a
l
f
r
a
m
e
w
o
r
k
s
th
at
ca
n
o
p
er
ate
w
i
th
i
n
d
is
tr
ib
u
ted
an
d
r
eso
u
r
ce
-
co
n
s
tr
ai
n
ed
in
f
r
astru
ct
u
r
es.
T
h
e
in
te
g
r
atio
n
o
f
GNN
r
ea
s
o
n
i
n
g
w
it
h
s
e
m
an
ti
c
v
alid
atio
n
allo
w
s
th
e
m
o
d
el
to
b
e
ad
ap
ted
f
o
r
d
e
p
lo
y
m
e
n
t
i
n
au
to
m
ated
h
ea
lt
h
in
f
o
r
m
atio
n
m
o
n
ito
r
in
g
s
y
s
te
m
s
a
n
d
ed
g
e
-
a
s
s
i
s
ted
an
al
y
tics
p
lat
f
o
r
m
s
.
Su
c
h
ar
c
h
itect
u
r
es
ca
n
s
u
p
p
o
r
t
r
ea
l
-
ti
m
e
f
ilter
in
g
o
f
b
io
m
ed
ical
co
n
ten
t
in
o
n
li
n
e
h
ea
lt
h
p
o
r
tals
,
m
o
b
ile
h
ea
lth
ap
p
licatio
n
s
,
a
n
d
cli
n
ical
d
ec
i
s
io
n
-
s
u
p
p
o
r
t
en
v
ir
o
n
m
en
ts
.
F
u
r
t
h
e
r
m
o
r
e,
th
e
m
o
d
u
lar
d
esig
n
o
f
th
e
p
r
o
p
o
s
ed
f
r
am
e
w
o
r
k
en
ab
le
s
p
o
ten
tia
l
in
te
g
r
atio
n
w
i
th
r
ec
o
n
f
ig
u
r
ab
le
co
m
p
u
ti
n
g
p
latf
o
r
m
s
an
d
e
m
b
ed
d
ed
A
I
s
y
s
te
m
s
,
w
h
er
e
lig
h
t
w
ei
g
h
t
g
r
ap
h
i
n
f
er
en
ce
m
o
d
u
les
ca
n
as
s
is
t
in
co
n
tin
u
o
u
s
m
o
n
ito
r
in
g
o
f
h
ea
lt
h
-
r
elate
d
i
n
f
o
r
m
atio
n
s
tr
ea
m
s
.
T
h
is
ca
p
ab
ilit
y
a
lig
n
s
w
it
h
e
m
er
g
in
g
r
esear
c
h
d
ir
ec
tio
n
s
i
n
i
n
tel
lig
e
n
t
e
m
b
ed
d
ed
s
y
s
te
m
s
,
w
h
er
e
ad
v
an
ce
d
ML
al
g
o
r
ith
m
s
ar
e
i
n
co
r
p
o
r
ated
in
to
s
ca
lab
le
ar
ch
i
tectu
r
es
to
s
u
p
p
o
r
t
r
eliab
le
an
d
tr
u
s
t
w
o
r
th
y
d
ig
ita
l h
ea
lt
h
ec
o
s
y
s
te
m
s
.
2.
L
I
T
E
R
AT
U
RE
R
E
VI
E
W
R
esear
ch
o
n
a
u
to
m
ated
an
al
y
s
is
o
f
b
io
m
ed
ical
i
n
f
o
r
m
atio
n
h
as
ex
p
a
n
d
ed
r
ap
id
ly
in
r
ec
en
t
y
ea
r
s
,
d
r
iv
en
b
y
th
e
u
r
g
en
t
n
ee
d
to
m
an
a
g
e
m
is
i
n
f
o
r
m
atio
n
,
p
r
o
ce
s
s
u
n
s
tr
u
ct
u
r
ed
clin
ical
tex
t,
a
n
d
s
u
p
p
o
r
t
d
ec
is
io
n
-
m
ak
in
g
d
u
r
i
n
g
p
u
b
lic
h
ea
l
th
c
r
is
es.
Mu
c
h
o
f
th
e
ea
r
l
y
p
r
o
g
r
ess
in
t
h
is
ar
ea
ca
m
e
f
r
o
m
t
h
e
ap
p
licatio
n
o
f
DL
an
d
tr
an
s
f
er
lear
n
i
n
g
m
o
d
els
in
b
io
m
ed
ical
i
m
ag
e
an
d
t
ex
t
a
n
al
y
s
i
s
.
Fo
r
in
s
ta
n
ce
,
d
ee
p
co
n
v
o
lu
tio
n
a
l
ar
ch
itect
u
r
es
co
m
b
in
ed
w
it
h
t
r
an
s
f
er
lear
n
i
n
g
h
av
e
d
e
m
o
n
s
tr
ated
s
tr
o
n
g
p
er
f
o
r
m
a
n
ce
in
d
etec
tin
g
co
m
p
le
x
b
io
m
ed
ical
p
atter
n
s
s
u
ch
as
m
o
n
k
e
y
p
o
x
s
k
i
n
le
s
io
n
s
,
h
i
g
h
li
g
h
ti
n
g
th
e
v
al
u
e
o
f
d
o
m
ai
n
-
ad
ap
ted
n
eu
r
al
m
o
d
els
w
h
e
n
d
ea
lin
g
w
it
h
h
i
g
h
-
d
i
m
e
n
s
io
n
al
m
ed
ical
s
i
g
n
als
[
1
]
.
A
lt
h
o
u
g
h
t
h
is
w
o
r
k
p
r
i
m
ar
il
y
f
o
cu
s
es
o
n
m
ed
ical
i
m
a
g
in
g
r
ath
er
t
h
an
te
x
tu
a
l
m
is
in
f
o
r
m
at
io
n
,
it
r
ef
lects
a
b
r
o
ad
e
r
tr
en
d
in
w
h
ic
h
d
ee
p
n
eu
r
a
l
m
o
d
el
s
o
f
ten
o
u
tp
er
f
o
r
m
tr
ad
itio
n
al
ML
tec
h
n
iq
u
es
w
h
e
n
ap
p
lied
to
c
o
m
p
lex
b
io
m
ed
ical
d
ata.
T
h
is
o
b
s
er
v
atio
n
ali
g
n
s
w
it
h
r
esear
ch
in
N
L
P
,
w
h
er
e
co
m
p
ar
is
o
n
s
b
et
w
ee
n
s
h
allo
w
lear
n
in
g
an
d
DL
m
et
h
o
d
s
s
h
o
w
th
a
t
d
ee
p
ar
ch
itectu
r
es
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
R
ec
o
n
f
i
g
u
r
ab
le
&
E
m
b
ed
d
ed
Sy
s
t
I
SS
N:
2089
-
4864
Gra
p
h
n
eu
r
a
l n
etw
o
r
k
-
b
a
s
ed
b
io
med
ica
l
mis
in
fo
r
ma
tio
n
d
et
ec
tio
n
w
ith
s
ema
n
tic
…
(
S
iva
Dh
ieva
r
a
j
)
441
t
y
p
icall
y
ac
h
ie
v
e
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
i
n
tas
k
s
i
n
v
o
lv
i
n
g
co
n
tex
tu
al
u
n
d
er
s
tan
d
i
n
g
an
d
s
e
m
an
tic
i
n
t
er
p
r
etatio
n
[
2
]
.
Ho
w
ev
er
,
t
h
ese
s
t
u
d
ies
also
h
ig
h
li
g
h
t
t
h
at
DL
alo
n
e
ca
n
n
o
t
f
u
ll
y
ad
d
r
ess
tas
k
s
r
eq
u
ir
in
g
s
tr
u
c
tu
r
ed
r
ea
s
o
n
in
g
o
r
d
o
m
ai
n
k
n
o
w
led
g
e,
w
h
ic
h
b
ec
o
m
es e
s
p
ec
iall
y
i
m
p
o
r
tan
t
w
h
en
e
v
al
u
ati
n
g
b
io
m
ed
ical
clai
m
s
.
As
N
L
P
m
e
th
o
d
s
e
v
o
lv
ed
,
ex
p
lain
ab
ili
t
y
a
n
d
in
ter
p
r
etab
ilit
y
b
ec
a
m
e
i
n
cr
ea
s
i
n
g
l
y
i
m
p
o
r
ta
n
t
f
o
r
h
ea
lt
h
-
o
r
ie
n
ted
ap
p
licatio
n
s
.
R
ec
en
t w
o
r
k
e
m
p
h
a
s
izes
t
h
at
m
an
y
DL
m
o
d
els
i
n
h
ea
lt
h
ca
r
e
clai
m
ex
p
lai
n
ab
ilit
y
b
u
t
f
ail
to
p
r
o
v
id
e
e
x
p
lan
atio
n
s
a
lig
n
ed
w
it
h
g
e
n
u
in
e
b
io
m
ed
ical
r
ea
s
o
n
in
g
[
3
]
.
T
h
is
i
s
s
u
e
is
p
ar
ticu
lar
l
y
r
elev
an
t
i
n
m
i
s
in
f
o
r
m
a
tio
n
d
et
ec
tio
n
,
w
h
er
e
m
is
lead
in
g
b
io
m
ed
ical
n
ar
r
ativ
es
o
f
te
n
m
i
m
ic
l
eg
iti
m
ate
s
c
ien
t
if
ic
lan
g
u
a
g
e
w
h
ile
s
u
b
tl
y
d
is
to
r
tin
g
f
ac
tu
a
l
r
elatio
n
s
h
ip
s
.
S
y
s
te
m
atic
r
ev
ie
w
s
o
f
f
a
k
e
n
e
w
s
d
etec
ti
o
n
r
esear
ch
also
r
ev
ea
l
th
at
al
th
o
u
g
h
DL
tec
h
n
iq
u
es
h
a
v
e
s
ig
n
i
f
ica
n
tl
y
i
m
p
r
o
v
ed
class
i
f
icat
io
n
p
er
f
o
r
m
a
n
c
e,
m
o
s
t
ap
p
r
o
ac
h
es
r
e
m
ain
h
ea
v
il
y
te
x
t
-
b
ased
an
d
r
ar
ely
i
n
co
r
p
o
r
ate
d
o
m
ai
n
-
s
p
ec
if
ic
k
n
o
w
led
g
e
v
er
i
f
icatio
n
m
ec
h
a
n
i
s
m
s
[
4
]
.
A
s
a
r
esu
lt,
m
an
y
s
y
s
te
m
s
ca
n
d
e
tect
s
t
y
li
s
tic
an
o
m
alies
b
u
t
s
tr
u
g
g
le
to
v
er
if
y
w
h
et
h
er
b
io
m
ed
ical
claim
s
alig
n
w
it
h
es
tab
lis
h
ed
s
cien
tific
ev
i
d
en
ce
.
T
h
e
i
m
p
o
r
tan
ce
o
f
d
o
m
ai
n
k
n
o
w
led
g
e
b
ec
a
m
e
e
s
p
ec
ia
ll
y
e
v
id
en
t
d
u
r
i
n
g
th
e
C
O
VI
D
-
1
9
p
an
d
e
m
ic,
w
h
e
n
m
is
i
n
f
o
r
m
atio
n
s
p
r
ea
d
r
ap
id
ly
ac
r
o
s
s
d
ig
i
tal
p
lat
f
o
r
m
s
.
A
r
ti
f
icial
in
te
lli
g
en
ce
(
A
I
)
a
n
d
N
L
P
to
o
ls
w
er
e
w
id
el
y
u
s
ed
to
p
r
o
ce
s
s
b
io
m
e
d
ical
li
ter
atu
r
e,
an
al
y
ze
e
m
er
g
in
g
s
cie
n
ti
f
ic
f
in
d
i
n
g
s
,
an
d
s
u
p
p
o
r
t
in
f
o
r
m
ati
o
n
r
etr
iev
al
d
u
r
in
g
th
e
p
an
d
e
m
ic
[
5
]
.
T
h
ese
ap
p
licatio
n
s
h
i
g
h
li
g
h
ted
th
e
ch
alle
n
g
e
s
o
f
an
al
y
zi
n
g
b
io
m
ed
ic
al
tex
t,
w
h
ic
h
co
n
tai
n
s
s
p
ec
ialized
ter
m
i
n
o
lo
g
y
,
e
v
o
lv
in
g
k
n
o
w
led
g
e
s
tr
u
ct
u
r
es,
an
d
co
m
p
le
x
r
ela
tio
n
s
h
ip
s
b
et
w
ee
n
en
titi
e
s
s
u
c
h
as d
is
ea
s
es,
tr
ea
t
m
en
ts
,
a
n
d
b
io
lo
g
ical
p
r
o
ce
s
s
es.
Si
m
ilar
l
y
,
lar
g
e
-
s
ca
le
m
i
s
i
n
f
o
r
m
atio
n
d
etec
t
io
n
s
y
s
te
m
s
co
m
b
in
in
g
NL
P
,
b
ig
d
a
ta
an
al
y
tics
,
an
d
DL
h
a
v
e
d
e
m
o
n
s
tr
ated
th
e
ab
ilit
y
to
d
etec
t
m
is
i
n
f
o
r
m
atio
n
p
atter
n
s
ac
r
o
s
s
lar
g
e
d
ataset
s
[
6
]
.
Desp
ite
th
ese
ad
v
an
ce
m
en
ts
,
m
o
s
t
o
f
th
e
s
e
f
r
a
m
e
w
o
r
k
s
lac
k
ex
p
lici
t
b
io
m
ed
ical
k
n
o
w
led
g
e
s
tr
u
ctu
r
es th
at
co
u
ld
en
ab
le
d
ee
p
er
v
alid
atio
n
o
f
s
cie
n
ti
f
ic
clai
m
s
.
Sev
er
al
s
tu
d
ie
s
h
av
e
atte
m
p
t
ed
to
en
h
a
n
ce
m
i
s
i
n
f
o
r
m
atio
n
d
etec
tio
n
p
er
f
o
r
m
a
n
ce
b
y
d
ev
elo
p
in
g
h
y
b
r
id
ar
ch
itect
u
r
es
a
n
d
f
ea
t
u
r
e
-
r
ich
lear
n
i
n
g
f
r
a
m
e
w
o
r
k
s
.
Di
s
tr
ib
u
ted
lear
n
i
n
g
ap
p
r
o
ac
h
es
h
av
e
b
ee
n
p
r
o
p
o
s
ed
to
i
m
p
r
o
v
e
s
ca
lab
ili
t
y
a
n
d
ef
f
i
cien
c
y
w
h
e
n
p
r
o
ce
s
s
in
g
lar
g
e
v
o
lu
m
es
o
f
m
i
s
in
f
o
r
m
a
tio
n
d
a
ta
[
7
]
.
Hy
b
r
id
d
ee
p
n
eu
r
al
n
et
w
o
r
k
s
(
DN
N
s
)
t
h
a
t
co
m
b
i
n
e
m
u
ltip
le
ar
c
h
itect
u
r
a
l
co
m
p
o
n
e
n
ts
h
a
v
e
al
s
o
d
e
m
o
n
s
tr
ated
i
m
p
r
o
v
ed
class
i
f
icatio
n
ac
cu
r
ac
y
w
h
en
ap
p
lied
to
s
o
cial
m
ed
ia
m
i
s
in
f
o
r
m
at
io
n
d
etec
tio
n
[
8
]
.
I
n
ad
d
itio
n
,
f
ea
tu
r
e
-
b
ased
DL
m
o
d
els
t
h
at
in
te
g
r
ate
li
n
g
u
is
tic,
s
e
m
an
tic,
a
n
d
co
n
tex
t
u
al
s
ig
n
al
s
h
av
e
s
h
o
w
n
p
r
o
m
i
s
i
n
g
r
esu
lt
s
in
d
etec
tin
g
f
ak
e
n
e
w
s
ac
r
o
s
s
s
o
cial
n
et
w
o
r
k
s
[
9
]
.
E
ar
lier
r
esear
ch
o
n
DL
-
b
ased
m
is
i
n
f
o
r
m
atio
n
d
etec
tio
n
laid
th
e
f
o
u
n
d
atio
n
f
o
r
th
ese
d
ev
elo
p
m
en
ts
b
y
d
e
m
o
n
s
tr
ati
n
g
h
o
w
co
n
v
o
l
u
tio
n
a
l
an
d
r
ec
u
r
r
en
t
n
eu
r
al
n
et
w
o
r
k
s
ca
n
b
e
ad
ap
ted
f
o
r
tex
t
-
b
ased
cr
ed
ib
ilit
y
clas
s
if
icatio
n
tas
k
s
[
1
0
]
.
C
o
llecti
v
el
y
,
th
e
s
e
s
tu
d
ie
s
il
lu
s
tr
ate
th
a
t
n
e
u
r
al
ar
ch
itect
u
r
es
h
a
v
e
m
at
u
r
ed
s
ig
n
i
f
ica
n
tl
y
;
h
o
w
e
v
er
,
th
e
y
s
ti
ll
r
ely
p
r
i
m
ar
il
y
o
n
tex
t
u
al
p
atter
n
s
an
d
r
a
r
el
y
in
co
r
p
o
r
ate
s
tr
u
ctu
r
ed
b
io
m
ed
ical
k
n
o
w
led
g
e.
A
d
v
an
ce
s
in
DNN
o
p
ti
m
iza
tio
n
h
a
v
e
also
co
n
tr
ib
u
ted
to
i
m
p
r
o
v
e
m
e
n
ts
i
n
m
o
d
el
p
er
f
o
r
m
a
n
ce
ac
r
o
s
s
m
u
ltip
le
ap
p
licatio
n
d
o
m
ai
n
s
.
R
ec
en
t
r
esear
c
h
h
a
s
ex
p
lo
r
e
d
s
tr
ateg
ies
f
o
r
i
m
p
r
o
v
i
n
g
co
n
v
o
l
u
tio
n
a
l
n
e
u
r
al
n
et
w
o
r
k
(
C
NN
)
tr
ai
n
i
n
g
ef
f
ici
en
c
y
th
r
o
u
g
h
q
u
a
lit
y
-
a
w
ar
e
d
ataset
o
p
ti
m
izatio
n
,
d
e
m
o
n
s
tr
ati
n
g
h
o
w
r
e
f
i
n
ed
d
ata
s
a
m
p
li
n
g
an
d
tr
ain
i
n
g
s
tr
ate
g
ies
ca
n
en
h
a
n
ce
m
o
d
el
r
eliab
ilit
y
[
1
1
]
.
Si
m
ilar
m
et
h
o
d
o
lo
g
ical
d
e
v
elo
p
m
e
n
ts
ap
p
ea
r
in
c
y
b
e
r
s
ec
u
r
it
y
r
esear
ch
,
w
h
er
e
e
x
p
lai
n
ab
le
DL
f
r
a
m
e
w
o
r
k
s
h
a
v
e
b
ee
n
ap
p
lied
to
d
etec
t b
o
tn
et
tr
af
f
ic
w
h
ile
m
ai
n
tain
in
g
tr
an
s
p
ar
en
c
y
in
m
o
d
el
d
ec
is
io
n
-
m
a
k
i
n
g
[
1
2
]
.
T
h
ese
ef
f
o
r
ts
h
ig
h
li
g
h
t
t
h
e
i
n
cr
ea
s
i
n
g
i
m
p
o
r
tan
ce
o
f
in
ter
p
r
etab
ilit
y
i
n
s
e
n
s
it
iv
e
d
o
m
a
in
s
s
u
c
h
as
h
ea
l
t
h
ca
r
e
an
d
c
y
b
er
s
ec
u
r
it
y
.
A
d
d
itio
n
al
p
r
o
g
r
ess
h
as
b
ee
n
m
ad
e
in
d
esig
n
i
n
g
co
m
p
u
ta
tio
n
all
y
e
f
f
i
cien
t
n
eu
r
al
ar
c
h
itect
u
r
es
ca
p
ab
le
o
f
s
u
p
p
o
r
tin
g
lig
h
t
w
ei
g
h
t
i
n
f
er
e
n
ce
,
w
h
ich
i
s
p
ar
ticu
lar
l
y
i
m
p
o
r
ta
n
t
f
o
r
r
ea
l
-
ti
m
e
ap
p
licatio
n
s
a
n
d
lar
g
e
-
s
ca
le
d
ep
lo
y
m
e
n
t
en
v
i
r
o
n
m
e
n
t
s
[
1
3
]
.
P
ar
allel
d
ev
elo
p
m
e
n
ts
in
d
is
tr
ib
u
ted
ML
h
av
e
f
u
r
th
er
e
x
p
an
d
ed
th
e
ca
p
ab
ilit
ies o
f
m
o
d
er
n
i
n
telli
g
e
n
t
s
y
s
te
m
s
.
Fed
er
ated
lear
n
i
n
g
a
p
p
r
o
ac
h
es,
f
o
r
ex
a
m
p
le,
e
n
ab
le
d
ec
en
tr
alize
d
m
o
d
el
tr
ain
i
n
g
ac
r
o
s
s
d
is
tr
ib
u
ted
d
ata
s
o
u
r
ce
s
w
h
i
le
p
r
eser
v
i
n
g
p
r
iv
ac
y
an
d
i
m
p
r
o
v
in
g
s
ca
lab
i
lit
y
[
1
4
]
.
T
h
ese
ad
v
an
ce
s
ar
e
p
ar
ticu
lar
l
y
r
ele
v
an
t
f
o
r
b
io
m
ed
ical
in
f
o
r
m
a
tio
n
s
y
s
te
m
s
,
w
h
er
e
s
en
s
iti
v
e
h
ea
lt
h
-
r
elate
d
d
ata
is
o
f
ten
d
is
tr
ib
u
ted
ac
r
o
s
s
m
u
ltip
le
p
latf
o
r
m
s
.
C
o
m
p
le
m
en
tar
y
r
es
ea
r
ch
o
n
f
ea
t
u
r
e
s
elec
tio
n
e
m
p
h
asizes
t
h
e
i
m
p
o
r
ta
n
ce
o
f
id
en
ti
f
y
in
g
s
tab
le
a
n
d
r
elev
an
t
f
ea
t
u
r
es
w
h
e
n
d
ea
li
n
g
w
ith
co
m
p
le
x
an
d
h
eter
o
g
e
n
e
o
u
s
d
ataset
s
[
1
5
]
.
Su
c
h
co
n
s
id
er
atio
n
s
ar
e
cr
itical
f
o
r
b
io
m
ed
ical
m
is
i
n
f
o
r
m
atio
n
d
etec
tio
n
,
w
h
er
e
n
o
is
y
d
ata
an
d
s
u
b
tle
s
e
m
a
n
tic
v
ar
iatio
n
s
ca
n
s
i
g
n
i
f
ica
n
tl
y
in
f
lu
e
n
ce
m
o
d
el
p
er
f
o
r
m
a
n
ce
.
A
lt
h
o
u
g
h
s
u
b
s
ta
n
tial
r
esear
c
h
h
as
e
x
p
lo
r
ed
f
ak
e
n
e
w
s
d
etec
tio
n
,
b
io
m
ed
ical
te
x
t
m
in
i
n
g
,
a
n
d
DL
-
b
ased
m
i
s
i
n
f
o
r
m
atio
n
an
a
l
y
s
i
s
,
s
ev
er
al
cr
itical
g
ap
s
r
em
ai
n
—
esp
ec
ial
l
y
w
h
en
ad
d
r
ess
i
n
g
h
ea
lt
h
-
r
elate
d
n
ar
r
ativ
e
s
th
at
r
eq
u
ir
e
s
cien
ti
f
i
c
ac
c
u
r
ac
y
r
ath
er
th
a
n
s
i
m
p
le
l
in
g
u
i
s
tic
class
if
ica
tio
n
.
Mo
s
t
ex
is
t
in
g
s
t
u
d
ies
tr
ea
t
m
is
in
f
o
r
m
at
io
n
d
etec
tio
n
as
a
co
n
v
en
tio
n
al
tex
t
class
if
ic
atio
n
p
r
o
b
lem
,
r
ely
in
g
p
r
i
m
a
r
il
y
o
n
s
t
y
li
s
tic
o
r
co
n
tex
t
u
al
f
ea
tu
r
es
p
r
esen
t
in
th
e
tex
t.
E
v
e
n
ad
v
an
ce
d
ar
ch
it
ec
tu
r
es
s
u
ch
as
C
NNs,
lo
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
s
(
L
ST
Ms
)
,
an
d
tr
an
s
f
o
r
m
er
-
b
as
ed
m
o
d
els
t
y
p
ical
l
y
f
o
cu
s
o
n
c
o
n
tex
t
u
al
e
m
b
ed
d
in
g
s
w
i
th
o
u
t
ev
alu
a
tin
g
w
h
et
h
er
b
io
m
ed
ical
s
ta
te
m
e
n
t
s
ar
e
lo
g
icall
y
co
n
s
i
s
te
n
t w
it
h
v
er
if
ied
s
cien
ti
f
ic
k
n
o
w
led
g
e.
T
h
is
li
m
i
tatio
n
is
s
i
g
n
if
ica
n
t
b
ec
au
s
e
b
io
m
ed
ical
m
i
s
in
f
o
r
m
atio
n
r
ar
el
y
ap
p
ea
r
s
lin
g
u
is
t
icall
y
s
u
s
p
icio
u
s
;
i
n
s
tead
,
it
o
f
te
n
in
v
o
lv
e
s
s
u
b
tle
d
is
to
r
tio
n
s
o
f
m
ed
ical
f
ac
ts
,
i
n
co
r
r
ec
t
ca
u
s
al
r
elatio
n
s
h
ip
s
,
o
r
m
i
s
lead
in
g
as
s
o
ciatio
n
s
b
et
w
ee
n
b
io
m
ed
ical
en
titi
e
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
9
-
4864
I
n
t J
R
ec
o
n
f
i
g
u
r
ab
le
&
E
m
b
ed
d
ed
Sy
s
t
,
Vo
l.
1
5
,
No
.
2
,
J
u
l
y
202
6
:
439
-
4
4
9
442
An
o
th
er
m
aj
o
r
lim
i
tatio
n
li
e
s
in
th
e
m
i
n
i
m
al
i
n
te
g
r
atio
n
o
f
s
tr
u
ct
u
r
ed
b
io
m
ed
ical
k
n
o
w
le
d
g
e.
W
h
ile
s
o
m
e
r
esear
ch
h
a
s
ex
p
lo
r
ed
ex
p
lain
ab
le
m
o
d
eli
n
g
ap
p
r
o
ac
h
e
s
,
m
o
s
t
s
y
s
te
m
s
s
till
d
o
n
o
t
i
n
c
o
r
p
o
r
ate
b
io
m
ed
ical
o
n
to
lo
g
ies,
cu
r
ated
k
n
o
w
led
g
e
g
r
ap
h
s
,
o
r
s
tr
u
ctu
r
ed
s
cie
n
ti
f
ic
r
elatio
n
s
h
ip
n
e
t
w
o
r
k
s
.
W
it
h
o
u
t
s
u
c
h
r
elatio
n
al
k
n
o
w
led
g
e,
m
o
d
els
ca
n
n
o
t
d
eter
m
i
n
e
w
h
eth
er
c
lai
m
s
ar
e
s
cien
t
if
ica
ll
y
p
lau
s
ib
le
o
r
co
n
tr
ad
ict
estab
li
s
h
ed
b
io
m
ed
ical
e
v
id
en
ce
.
A
d
d
itio
n
all
y
,
f
e
w
ex
i
s
ti
n
g
s
tu
d
ie
s
p
er
f
o
r
m
s
e
m
an
tic
co
n
s
i
s
te
n
c
y
v
e
r
if
icatio
n
b
y
cr
o
s
s
-
ch
ec
k
i
n
g
tex
t
u
al
clai
m
s
ag
ai
n
s
t
au
th
o
r
itati
v
e
b
io
m
ed
ical
s
o
u
r
ce
s
.
I
n
s
tead
,
m
a
n
y
r
el
y
o
n
s
u
r
f
ac
e
-
le
v
el
s
i
m
ilar
it
y
m
etr
ics o
r
r
etr
iev
al
-
b
ased
m
at
ch
in
g
,
w
h
ich
ar
e
i
n
s
u
f
f
ic
ien
t
f
o
r
id
en
tify
in
g
co
m
p
le
x
m
is
i
n
f
o
r
m
at
io
n
p
atter
n
s
.
Gr
ap
h
-
b
ased
r
ea
s
o
n
i
n
g
r
e
m
ai
n
s
a
n
o
th
er
u
n
d
er
ex
p
lo
r
ed
ar
e
a.
A
lt
h
o
u
g
h
G
NN
s
h
av
e
ac
h
i
ev
ed
s
tr
o
n
g
r
esu
lt
s
in
d
o
m
ai
n
s
s
u
c
h
as
b
io
l
o
g
ical
n
e
t
w
o
r
k
a
n
al
y
s
is
,
d
r
u
g
d
is
co
v
er
y
,
a
n
d
d
is
ea
s
e
p
r
ed
icti
o
n
,
th
eir
ap
p
licatio
n
to
b
io
m
ed
ical
m
is
i
n
f
o
r
m
atio
n
d
etec
tio
n
r
e
m
ai
n
s
li
m
ited
.
B
io
m
ed
ical
k
n
o
w
led
g
e
i
s
i
n
h
er
e
n
tl
y
r
elatio
n
al,
i
n
v
o
l
v
i
n
g
i
n
ter
ac
tio
n
s
a
m
o
n
g
d
is
ea
s
es,
s
y
m
p
to
m
s
,
tr
ea
t
m
e
n
t
s
,
an
d
b
io
lo
g
ical
m
ec
h
an
i
s
m
s
.
C
o
n
s
eq
u
e
n
tl
y
,
th
e
ab
s
en
ce
o
f
g
r
ap
h
-
b
ased
r
ea
s
o
n
in
g
f
r
a
m
e
w
o
r
k
s
r
ep
r
esen
ts
a
m
is
s
ed
o
p
p
o
r
tu
n
it
y
f
o
r
i
m
p
r
o
v
in
g
m
is
i
n
f
o
r
m
atio
n
d
etec
tio
n
ca
p
ab
ilit
ies.
I
n
ter
p
r
eta
b
ilit
y
also
r
em
a
in
s
an
o
p
en
ch
allen
g
e.
Ma
n
y
ex
is
ti
n
g
s
y
s
te
m
s
clai
m
e
x
p
lai
n
ab
ilit
y
b
u
t
p
r
o
v
id
e
o
n
l
y
s
u
p
er
f
icial
i
n
ter
p
r
etatio
n
s
s
u
c
h
a
s
atte
n
tio
n
m
a
p
s
o
r
to
k
en
-
le
v
el
i
m
p
o
r
tan
ce
s
co
r
es.
I
n
h
ea
l
th
ca
r
e
co
n
tex
t
s
,
m
ea
n
i
n
g
f
u
l
ex
p
la
n
ati
o
n
s
s
h
o
u
ld
d
e
m
o
n
s
tr
ate
h
o
w
a
clai
m
co
n
tr
ad
icts
k
n
o
w
n
b
io
m
ed
ical
r
elatio
n
s
h
ip
s
o
r
v
io
lates
estab
li
s
h
ed
s
cie
n
ti
f
ic
ev
id
en
ce
.
Ho
w
e
v
er
,
s
u
c
h
d
o
m
ai
n
-
a
lig
n
ed
ex
p
la
n
atio
n
s
ar
e
r
ar
ely
p
r
o
v
id
ed
b
y
cu
r
r
en
t
m
is
i
n
f
o
r
m
atio
n
d
etec
ti
o
n
f
r
a
m
e
w
o
r
k
s
.
Fin
all
y
,
s
e
v
er
al
d
ataset
-
r
elate
d
lim
itatio
n
s
p
er
s
is
t.
Ma
n
y
w
i
d
e
ly
u
s
ed
m
i
s
i
n
f
o
r
m
atio
n
d
at
asets
f
o
c
u
s
p
r
im
ar
il
y
o
n
C
OVI
D
-
19
–
r
elate
d
co
n
ten
t,
w
h
ich
r
estrict
s
th
e
d
iv
er
s
it
y
o
f
b
io
m
ed
ical
m
i
s
i
n
f
o
r
m
atio
n
p
atter
n
s
av
ailab
le
f
o
r
m
o
d
el
tr
ain
in
g
.
A
d
d
itio
n
al
l
y
,
m
o
s
t
d
atasets
la
ck
s
tr
u
ct
u
r
ed
an
n
o
tatio
n
s
li
n
k
i
n
g
te
x
t
u
al
clai
m
s
to
b
io
m
ed
ical
e
n
titi
e
s
an
d
r
elatio
n
s
h
ip
s
,
m
a
k
in
g
it
d
if
f
ic
u
lt
t
o
lear
n
r
elatio
n
al
r
ea
s
o
n
in
g
p
atter
n
s
.
E
v
al
u
atio
n
m
et
h
o
d
o
lo
g
ies
also
ten
d
to
em
p
h
a
s
ize
class
if
icatio
n
m
etr
ic
s
s
u
c
h
as
ac
c
u
r
ac
y
o
r
F1
-
s
co
r
e
w
it
h
o
u
t
a
s
s
es
s
i
n
g
w
h
et
h
er
p
r
ed
ictio
n
s
ali
g
n
w
i
th
b
io
m
ed
i
ca
l
tr
u
t
h
.
A
s
a
r
es
u
lt,
th
er
e
r
e
m
ai
n
s
a
g
ap
b
et
w
ee
n
m
o
d
el
p
er
f
o
r
m
an
c
e
m
etr
ics a
n
d
r
ea
l
-
w
o
r
ld
ap
p
licab
ilit
y
i
n
h
ea
lt
h
ca
r
e
co
m
m
u
n
ic
atio
n
en
v
ir
o
n
m
e
n
ts
.
3.
P
RO
P
O
SE
D
M
E
T
H
O
D
T
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
o
lo
g
y
a
d
v
an
ce
s
b
io
m
ed
ical
m
is
i
n
f
o
r
m
atio
n
d
etec
tio
n
b
y
e
m
b
ed
d
in
g
te
x
t
u
a
l
clai
m
s
in
to
a
s
tr
u
ct
u
r
ed
s
cien
t
if
ic
co
n
te
x
t,
allo
w
i
n
g
th
e
s
y
s
t
e
m
to
v
er
i
f
y
th
e
f
ac
t
u
al
v
alid
i
t
y
o
f
n
e
w
s
co
n
ten
t
th
r
o
u
g
h
k
n
o
w
led
g
e
-
g
r
o
u
n
d
ed
r
ea
s
o
n
in
g
.
T
h
e
p
ip
elin
e
b
e
g
in
s
w
it
h
t
h
e
co
llectio
n
o
f
b
io
m
e
d
ical
ar
ticles
f
r
o
m
r
ep
u
tab
le
m
is
i
n
f
o
r
m
atio
n
d
ata
s
ets,
i
n
cl
u
d
in
g
C
O
VI
D
-
1
9
Hea
lth
ca
r
e
Misi
n
f
o
r
m
atio
n
Data
s
et,
2
0
2
0
(
C
o
A
I
D
)
,
R
eliab
le
C
OVI
D
-
1
9
Ne
w
s
Data
s
et,
2
0
2
1
(
R
eCOVer
y
)
,
an
d
2
0
1
8
–
2
0
2
0
b
io
m
ed
ical
h
ea
lt
h
n
e
w
s
co
r
p
u
s
(
Hea
lth
Sto
r
y
)
,
w
h
ic
h
co
llectiv
el
y
p
r
o
v
id
e
m
o
r
e
th
a
n
2
0
,
0
0
0
lab
eled
ar
ticles
w
it
h
m
etad
ata
an
d
s
o
cial
-
co
n
tex
t
f
ea
t
u
r
es.
T
h
ese
d
atasets
ar
e
s
o
u
r
ce
d
f
r
o
m
p
u
b
licl
y
a
v
ailab
l
e
r
ep
o
s
ito
r
ies
h
o
s
ted
o
n
GitH
u
b
an
d
Z
e
n
o
d
o
,
an
d
in
cl
u
d
e
lab
els
i
n
d
icati
n
g
w
h
et
h
er
n
e
w
s
ite
m
s
ar
e
f
a
k
e,
r
ea
l,
m
i
s
lead
in
g
,
o
r
p
ar
tiall
y
f
ac
t
u
al.
A
l
l
d
o
cu
m
en
t
s
u
n
d
er
g
o
a
d
etailed
p
r
ep
r
o
ce
s
s
in
g
s
tag
e
t
h
at
i
n
clu
d
e
s
s
en
te
n
ce
s
eg
m
en
ta
tio
n
,
lo
w
er
ca
s
i
n
g
,
s
to
p
w
o
r
d
r
em
o
v
al,
an
d
en
tit
y
n
o
r
m
aliza
tio
n
u
s
in
g
b
io
m
ed
ical
ter
m
i
n
o
lo
g
ie
s
s
u
ch
a
s
Me
SH
a
n
d
UM
L
S
to
en
s
u
r
e
th
at
e
x
tr
ac
ted
u
n
i
ts
ali
g
n
w
it
h
s
ta
n
d
ar
d
ized
v
o
ca
b
u
lar
ies.
B
io
m
ed
ical
n
a
m
e
d
-
en
ti
t
y
r
ec
o
g
n
itio
n
is
p
er
f
o
r
m
ed
u
s
in
g
B
io
B
E
R
T
-
NE
R
,
w
h
ich
id
e
n
ti
f
ies
k
e
y
en
ti
t
y
ca
te
g
o
r
ies
—
d
is
ea
s
e
n
a
m
e
s
,
s
y
m
p
to
m
s
,
an
ato
m
ical
ter
m
s
,
d
r
u
g
s
,
p
r
o
tein
s
,
an
d
b
io
lo
g
ical
p
r
o
ce
s
s
es
—
an
d
m
a
p
s
th
e
m
to
ca
n
o
n
ical
f
o
r
m
s
u
s
i
n
g
t
h
e
UM
L
S
s
e
m
a
n
tic
n
et
w
o
r
k
.
E
n
tit
y
ex
tr
ac
t
io
n
p
r
o
d
u
ce
s
a
s
eq
u
en
ce
o
f
en
ti
t
y
to
k
e
n
s
e
1
,
e
2
,
…
,
e
n
,
an
d
co
r
r
esp
o
n
d
i
n
g
co
n
te
x
t
u
al
e
m
b
ed
d
in
g
s
h
i
∈
R
d
,
g
en
er
ated
u
s
i
n
g
d
o
m
ai
n
-
t
u
n
ed
B
E
R
T
em
b
ed
d
in
g
s
.
On
ce
en
ti
ties
an
d
th
e
ir
co
-
o
cc
u
r
r
e
n
ce
s
ar
e
id
en
ti
f
ied
,
th
e
s
y
s
te
m
co
n
s
tr
u
ct
s
a
B
KG
b
y
in
teg
r
at
in
g
s
tr
u
ct
u
r
ed
d
atab
ases
s
u
ch
as
UM
L
S
Me
tath
esa
u
r
u
s
,
Me
SH
d
escr
ip
to
r
s
,
Dr
u
g
B
an
k
p
h
ar
m
ac
o
lo
g
ical
r
elatio
n
s
,
C
T
D
to
x
ico
g
en
o
m
ic
i
n
ter
ac
tio
n
s
,
an
d
m
a
n
u
a
ll
y
v
er
if
ied
r
elat
io
n
s
ex
tr
ac
ted
f
r
o
m
P
u
b
Me
d
ab
s
tr
ac
ts
.
T
h
e
B
K
G
is
r
ep
r
esen
ted
as
a
d
ir
ec
ted
m
u
lti
g
r
ap
h
G=
(
V,
E
)
,
w
h
er
e
n
o
d
es
v
i
∈
V
co
r
r
esp
o
n
d
to
b
i
o
m
ed
ical
en
titi
e
s
an
d
ed
g
es
(
v
i
,
v
j
,
r
ij
)
∈
E
r
ep
r
esen
t
r
elatio
n
al
t
y
p
es
s
u
c
h
as
ca
u
s
es,
tr
ea
ts
,
co
-
o
cc
u
r
s
-
w
it
h
,
i
n
ter
ac
ts
-
w
it
h
,
o
r
co
n
tr
ain
d
icate
d
-
f
o
r
.
No
d
e
attr
i
b
u
tes
ar
e
in
itial
ized
u
s
i
n
g
B
io
W
o
r
d
Vec
em
b
ed
d
in
g
s
,
w
h
ile
ed
g
es
in
cl
u
d
e
b
o
th
s
y
m
b
o
lic
r
elatio
n
lab
els
an
d
n
u
m
er
ica
l
w
e
ig
h
ts
r
e
f
lecti
n
g
r
e
latio
n
co
n
f
id
en
ce
.
T
o
ca
p
tu
r
e
s
e
m
a
n
tic
p
r
o
x
i
m
it
y
b
et
w
ee
n
t
w
o
n
o
d
es,
a
g
r
ap
h
d
i
s
tan
ce
m
etr
ic
is
ap
p
lied
as
:
(
,
)
=
min
∈
(
,
)
∑
,
(
,
)
∈
w
h
er
e
(
,
)
d
en
o
tes
all
p
o
s
s
ib
le
p
ath
s
b
et
w
ee
n
n
o
d
es
an
d
,
co
r
r
esp
o
n
d
s
to
th
e
ed
g
e
w
ei
g
h
t
.
C
lai
m
s
ex
tr
ac
ted
f
r
o
m
n
e
w
s
ar
ticle
s
a
r
e
co
n
v
er
ted
in
to
r
elatio
n
al
t
u
p
les
o
f
th
e
f
o
r
m
=
(
,
,
)
an
d
m
ap
p
ed
to
th
e
k
n
o
w
led
g
e
g
r
ap
h
to
ev
a
lu
ate
t
h
eir
p
lau
s
ib
il
it
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
R
ec
o
n
f
i
g
u
r
ab
le
&
E
m
b
ed
d
ed
Sy
s
t
I
SS
N:
2089
-
4864
Gra
p
h
n
eu
r
a
l n
etw
o
r
k
-
b
a
s
ed
b
io
med
ica
l
mis
in
fo
r
ma
tio
n
d
et
ec
tio
n
w
ith
s
ema
n
tic
…
(
S
iva
Dh
ieva
r
a
j
)
443
T
o
v
er
if
y
b
io
m
ed
ical
p
la
u
s
ib
ilit
y
,
t
h
e
m
et
h
o
d
o
lo
g
y
i
n
co
r
p
o
r
ates
a
s
e
m
an
tic
co
n
s
i
s
te
n
c
y
ch
ec
k
i
n
g
m
o
d
u
le
t
h
at
co
m
p
ar
es
ea
ch
cla
i
m
’
s
i
m
p
lied
r
elatio
n
s
h
ip
to
th
e
g
r
ap
h
’
s
f
ac
t
u
al
s
tr
u
ct
u
r
e.
T
h
e
s
e
m
a
n
tic
s
i
m
ila
r
i
t
y
b
et
w
ee
n
clai
m
e
m
b
ed
d
in
g
s
a
n
d
g
r
ap
h
-
d
er
iv
ed
e
m
b
ed
d
in
g
s
i
s
co
m
p
u
ted
u
s
i
n
g
a
co
s
i
n
e
m
e
asu
r
e
:
(
,
)
=
ℎ
.
ℎ
|
|
ℎ
|
|
|
|
ℎ
|
|
w
h
er
e
ℎ
r
ep
r
esen
ts
th
e
e
m
b
ed
d
in
g
o
f
th
e
clai
m
r
elatio
n
o
b
tain
ed
f
r
o
m
B
io
B
E
R
T
,
an
d
ℎ
is
th
e
r
elatio
n
al
e
m
b
ed
d
in
g
a
g
g
r
e
g
ated
f
r
o
m
g
r
ap
h
n
eig
h
b
o
r
h
o
o
d
s
.
A
co
n
s
is
t
en
c
y
s
co
r
e
is
co
m
p
u
ted
as
:
=
⋅
(
,
)
−
⋅
(
,
)
,
w
h
er
e
α
a
n
d
β
ar
e
lear
n
a
b
le
p
ar
a
m
eter
s
co
n
tr
o
lli
n
g
th
e
tr
ad
e
-
o
f
f
b
et
w
ee
n
s
e
m
a
n
tic
s
i
m
i
la
r
it
y
an
d
b
io
m
ed
ical
g
r
ap
h
d
is
ta
n
ce
.
H
ig
h
er
v
al
u
es
o
f
in
d
icate
s
tr
o
n
g
er
ali
g
n
m
e
n
t
b
et
w
ee
n
t
h
e
c
lai
m
a
n
d
s
cie
n
ti
f
ic
k
n
o
w
led
g
e,
w
h
ile
n
eg
at
iv
e
o
r
n
ea
r
-
ze
r
o
s
co
r
es f
lag
p
o
ten
t
ial
m
is
i
n
f
o
r
m
a
tio
n
.
Fo
r
f
in
a
l
clas
s
i
f
icatio
n
,
t
h
e
s
y
s
te
m
e
m
p
lo
y
s
a
g
r
ap
h
atten
tio
n
n
et
w
o
r
k
(
G
A
T
)
th
at
i
n
te
g
r
ates
s
tr
u
ct
u
r
al
in
f
o
r
m
atio
n
f
r
o
m
t
h
e
k
n
o
w
led
g
e
g
r
ap
h
w
i
th
co
n
tex
tu
al
e
m
b
ed
d
in
g
s
f
r
o
m
t
h
e
te
x
tu
al
c
lai
m
.
E
ac
h
n
e
w
s
i
te
m
i
s
r
ep
r
esen
ted
as
a
claim
-
i
n
d
u
c
ed
s
u
b
g
r
ap
h
⊂
,
co
n
s
is
ti
n
g
o
f
th
e
m
e
n
tio
n
ed
en
ti
ties
,
th
eir
i
m
m
ed
iate
n
eig
h
b
o
r
s
,
an
d
s
tr
u
ct
u
r
all
y
r
e
lev
an
t
b
io
m
ed
ical
r
elatio
n
s
.
T
h
e
GA
T
u
p
d
ates
n
o
d
e
r
ep
r
esen
tat
io
n
s
t
h
r
o
u
g
h
atten
tio
n
-
b
ased
p
r
o
p
ag
atio
n
u
s
in
g
:
ℎ
+
1
=
(
∑
ℎ
(
)
)
w
h
er
e
(
)
d
en
o
tes
th
e
n
e
ig
h
b
o
r
s
o
f
n
o
d
e
i
,
is
a
tr
ain
ab
le
w
ei
g
h
t
m
atr
ix
,
σ
i
s
an
ac
ti
v
atio
n
f
u
n
ctio
n
,
a
n
d
th
e
atte
n
tio
n
co
ef
f
ic
ien
t i
s
co
m
p
u
ted
as
:
=
(
(
[
ℎ
∥
ℎ
]
)
)
∑
(
(
[
ℎ
∥
ℎ
]
)
)
(
)
T
h
is
m
ec
h
a
n
is
m
en
s
u
r
e
s
th
at
t
h
e
m
o
d
el
ass
i
g
n
s
g
r
ea
ter
w
ei
g
h
t
to
m
ed
ical
l
y
i
m
p
o
r
ta
n
t
r
elatio
n
s
—
s
u
c
h
as
s
y
m
p
to
m
–
d
is
ea
s
e
o
r
d
r
u
g
–
ef
f
ec
t
p
air
s
—
w
h
i
le
d
o
w
n
-
w
ei
g
h
t
in
g
ir
r
elev
a
n
t
o
r
w
ea
k
l
y
co
n
n
ec
ted
n
o
d
es.
T
h
e
o
u
tp
u
t
r
ep
r
esen
tatio
n
o
f
th
e
s
u
b
g
r
ap
h
is
p
ass
ed
th
r
o
u
g
h
a
f
u
ll
y
co
n
n
ec
ted
la
y
er
to
p
r
o
d
u
ce
a
f
in
al
p
r
o
b
ab
ilit
y
s
co
r
e
r
ef
lectin
g
w
h
eth
er
t
h
e
cl
ai
m
is
f
ac
t
u
al,
m
is
lead
i
n
g
,
o
r
s
cien
ti
f
icall
y
i
n
co
r
r
ec
t.
T
h
e
tr
ain
in
g
p
h
ase
e
m
p
lo
y
s
a
co
m
p
o
s
ite
lo
s
s
f
u
n
ctio
n
t
h
a
t
in
te
g
r
ates
clas
s
i
f
icatio
n
ac
c
u
r
ac
y
w
it
h
s
cien
t
if
ic
co
n
s
i
s
ten
c
y
.
T
h
e
to
tal
lo
s
s
is
r
ep
r
esen
ted
as
:
=
+
,
w
h
er
e
is
th
e
cr
o
s
s
-
en
tr
o
p
y
lo
s
s
f
o
r
class
if
icat
io
n
,
=
∥
−
ℎ
∥
2
p
en
alize
s
clai
m
s
th
at
d
ev
iate
f
r
o
m
b
io
m
ed
ical
k
n
o
w
led
g
e
p
atter
n
s
,
an
d
λ
r
eg
u
late
s
th
e
s
tr
en
g
t
h
o
f
t
h
e
s
e
m
a
n
tic
co
n
s
tr
ain
t.
H
y
p
er
p
ar
a
m
eter
s
s
u
c
h
as
lear
n
i
n
g
r
ate
(
0
.
0
0
1
–
0
.
0
0
0
5
)
,
d
r
o
p
o
u
t
(
0
.
3
–
0
.
5
)
,
at
ten
tio
n
h
ea
d
s
(
6
–
1
2
)
,
an
d
em
b
ed
d
in
g
d
i
m
e
n
s
io
n
s
(
2
5
6
–
7
6
8
)
a
r
e
tu
n
ed
th
r
o
u
g
h
g
r
id
s
ea
r
ch
f
o
r
o
p
ti
m
al
p
er
f
o
r
m
a
n
ce
.
T
h
e
m
o
d
el
is
t
r
ain
ed
u
s
in
g
th
e
A
d
a
m
o
p
tim
izer
w
ith
w
e
ig
h
t
d
ec
a
y
t
o
en
s
u
r
e
s
tab
le
co
n
v
er
g
e
n
ce
.
B
y
g
r
o
u
n
d
i
n
g
clai
m
s
w
it
h
i
n
a
r
ich
b
io
m
ed
ical
g
r
ap
h
,
en
f
o
r
cin
g
s
cie
n
ti
f
ic
co
n
s
i
s
te
n
c
y
,
a
n
d
ap
p
ly
i
n
g
g
r
ap
h
n
eu
r
al
r
ea
s
o
n
i
n
g
,
th
i
s
m
eth
o
d
o
lo
g
y
tr
a
n
s
f
o
r
m
s
b
io
m
e
d
ical
m
i
s
in
f
o
r
m
a
tio
n
d
etec
tio
n
f
r
o
m
a
s
i
m
p
le
te
x
t
class
i
f
icatio
n
p
r
o
b
lem
in
to
a
s
t
r
u
ctu
r
ed
in
f
er
e
n
ce
tas
k
.
T
h
is
a
p
p
r
o
ac
h
en
ab
les
th
e
m
o
d
el
to
id
en
tify
clai
m
s
th
a
t
co
n
tr
ad
ict
b
io
m
ed
ical
k
n
o
w
l
ed
g
e
ev
en
w
h
en
t
h
e
y
ap
p
ea
r
lin
g
u
i
s
ticall
y
cr
ed
ib
le,
p
r
o
v
id
in
g
a
r
o
b
u
s
t
a
n
d
in
ter
p
r
etab
le
f
r
a
m
e
w
o
r
k
f
o
r
id
en
ti
f
y
in
g
f
r
a
u
d
u
le
n
t o
r
m
i
s
lea
d
in
g
b
io
m
ed
ical
n
e
w
s
.
4.
SYST
E
M
ARCH
I
T
E
CT
U
R
E
DE
SCRI
P
T
I
O
N
T
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
ar
c
h
ite
ctu
r
e
is
d
esig
n
ed
as
a
m
u
l
ti
la
y
er
an
a
l
y
t
ical
p
ip
elin
e
f
o
r
d
etec
tin
g
b
io
m
ed
ical
m
i
s
i
n
f
o
r
m
atio
n
b
y
i
n
teg
r
ati
n
g
ad
v
a
n
ce
d
DL
,
NL
P
,
an
d
k
n
o
w
led
g
e
-
d
r
iv
e
n
r
ea
s
o
n
in
g
.
U
n
li
k
e
c
o
n
v
e
n
tio
n
al
s
in
g
le
-
s
ta
g
e
cla
s
s
i
f
ier
s
,
th
i
s
ar
ch
itect
u
r
e
m
o
d
els
th
e
ev
o
lu
tio
n
o
f
in
f
o
r
m
atio
n
ac
r
o
s
s
d
ig
ital
p
latf
o
r
m
s
,
tr
a
n
s
f
o
r
m
i
n
g
r
a
w
te
x
tu
a
l
co
n
te
n
t
i
n
to
s
tr
u
c
tu
r
ed
,
s
e
m
an
ticall
y
en
r
ic
h
ed
r
ep
r
esen
tatio
n
s
s
u
i
tab
le
f
o
r
r
ea
s
o
n
in
g
.
T
h
e
ar
ch
itect
u
r
e
is
co
n
ce
p
tu
all
y
d
iv
id
ed
in
to
t
h
r
ee
in
ter
d
ep
en
d
en
t la
y
er
s
:
−
Data
ac
q
u
is
it
io
n
an
d
p
r
ep
r
o
ce
s
s
i
n
g
la
y
er
−
Featu
r
e
r
ep
r
esen
tatio
n
an
d
k
n
o
w
led
g
e
i
n
te
g
r
atio
n
la
y
er
−
GNN
–
b
ased
class
i
f
icatio
n
an
d
ex
p
lain
ab
ilit
y
la
y
er
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
9
-
4864
I
n
t J
R
ec
o
n
f
i
g
u
r
ab
le
&
E
m
b
ed
d
ed
Sy
s
t
,
Vo
l.
1
5
,
No
.
2
,
J
u
l
y
202
6
:
439
-
4
4
9
444
E
ac
h
la
y
er
b
u
ild
s
u
p
o
n
t
h
e
p
r
ev
io
u
s
o
n
e,
en
s
u
r
in
g
th
at
u
n
s
t
r
u
ctu
r
ed
,
n
o
is
y
te
x
t
is
tr
an
s
f
o
r
m
ed
in
to
m
ea
n
in
g
f
u
l
e
m
b
ed
d
in
g
s
w
h
il
e
p
r
eser
v
in
g
co
n
tex
tu
al
an
d
r
elatio
n
al
c
h
ar
ac
ter
is
tic
s
es
s
en
tial
f
o
r
ac
cu
r
ate
m
is
in
f
o
r
m
at
io
n
d
etec
tio
n
.
Fi
g
u
r
e
1
s
h
o
w
s
t
h
e
ar
ch
itect
u
r
e
d
i
ag
r
a
m
.
Fig
u
r
e
1
.
S
y
s
te
m
ar
ch
itect
u
r
e
4.
1
.
Da
t
a
a
cquis
it
io
n a
nd
pr
epro
ce
s
s
ing
la
y
er
T
h
is
f
o
u
n
d
atio
n
a
l
la
y
er
h
a
n
d
l
es
d
ata
co
llectio
n
,
c
lean
s
in
g
,
an
d
in
itial
r
ep
r
ese
n
tatio
n
.
T
ex
tu
al
d
ata
i
s
g
ath
er
ed
f
r
o
m
s
o
cial
m
ed
ia
p
latf
o
r
m
s
,
p
ee
r
-
r
ev
ie
w
ed
b
io
m
ed
ical
p
u
b
licatio
n
s
,
a
n
d
cu
r
at
ed
m
is
i
n
f
o
r
m
atio
n
r
ep
o
s
ito
r
i
es,
in
clu
d
in
g
R
eC
O
Ver
y
(
~2
,
0
0
0
C
OVI
D
-
1
9
ar
ti
cles)
,
C
o
A
I
D
(
>4
,
0
0
0
ar
ticles
+
2
9
6
,
0
0
0
r
elate
d
s
o
cial
p
o
s
ts
)
,
an
d
Hea
lth
Sto
r
y
(
~5
,
0
0
0
item
s
)
.
T
h
ese
d
atasets
p
r
o
v
id
e
lab
eled
in
s
tan
ce
s
o
f
tr
u
e,
m
is
lead
in
g
,
an
d
f
ak
e
b
io
m
ed
ical
co
n
ten
t.
R
a
w
co
n
te
n
t
o
f
te
n
co
n
tain
s
n
o
is
e
s
u
c
h
as
e
m
o
j
is
,
h
y
p
er
li
n
k
s
,
in
f
o
r
m
a
l
ab
b
r
ev
iatio
n
s
,
m
u
l
tili
n
g
u
a
l
ex
p
r
ess
io
n
s
,
an
d
b
o
t
-
g
en
er
at
ed
p
o
s
ts
.
P
r
e
p
r
o
ce
s
s
in
g
tr
an
s
f
o
r
m
s
t
h
i
s
n
o
is
y
i
n
p
u
t
i
n
to
clea
n
,
s
ta
n
d
ar
d
ized
s
eq
u
en
ce
s
,
w
h
ile
p
r
eser
v
i
n
g
s
y
n
tac
tic
an
d
s
e
m
a
n
tic
c
u
es.
T
h
e
p
r
ep
r
o
ce
s
s
in
g
f
u
n
c
tio
n
m
a
p
s
ea
ch
r
a
w
tex
t
in
s
ta
n
ce
=
{
1
,
2
,
…
,
}
in
to
a
clea
n
ed
to
k
en
s
eq
u
en
ce
.
̈
=
(
)
=
{
⃛
1
,
⃛
2
…
.
.
⃛
}
,
≤
h
er
e,
⃛
d
en
o
tes
n
o
r
m
a
lized
an
d
m
ea
n
in
g
f
u
l
to
k
en
s
r
etai
n
ed
af
ter
n
o
i
s
e
f
il
ter
in
g
.
I
n
ad
d
itio
n
,
m
etad
ata
s
u
c
h
as p
o
s
tin
g
ti
m
esta
m
p
,
s
o
u
r
ce
cr
ed
ib
ilit
y
,
an
d
s
h
ar
i
n
g
f
r
eq
u
e
n
c
y
ar
e
en
co
d
ed
as v
ec
to
r
s
to
c
ap
tu
r
e
te
m
p
o
r
al
an
d
r
elatio
n
al
asp
ec
ts
o
f
m
i
s
i
n
f
o
r
m
atio
n
p
r
o
p
ag
atio
n
.
4.
2
.
F
e
a
t
ure
re
presenta
t
io
n
a
nd
k
no
w
ledg
e
in
t
eg
ra
t
io
n l
a
y
er
On
ce
p
r
ep
r
o
ce
s
s
in
g
is
co
m
p
l
ete,
th
e
s
y
s
te
m
tr
an
s
f
o
r
m
s
t
o
k
en
s
in
to
co
n
tex
t
u
al
e
m
b
ed
d
in
g
s
th
a
t
ca
p
tu
r
e
s
e
m
a
n
tic,
s
y
n
tact
ic,
an
d
d
o
m
ain
-
s
p
ec
i
f
ic
n
u
a
n
ce
s
.
U
n
li
k
e
s
h
al
lo
w
m
o
d
els
(
T
F
-
I
DF
an
d
b
ag
-
of
-
w
o
r
d
s
)
,
tr
an
s
f
o
r
m
er
-
b
a
s
ed
em
b
ed
d
in
g
s
s
u
c
h
as
R
o
B
E
R
T
a
ar
e
u
s
ed
,
m
ap
p
in
g
ea
ch
to
k
en
⃛
in
to
a
h
ig
h
-
d
i
m
e
n
s
io
n
a
l
v
ec
to
r
ℎ
∈
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
R
ec
o
n
f
i
g
u
r
ab
le
&
E
m
b
ed
d
ed
Sy
s
t
I
SS
N:
2089
-
4864
Gra
p
h
n
eu
r
a
l n
etw
o
r
k
-
b
a
s
ed
b
io
med
ica
l
mis
in
fo
r
ma
tio
n
d
et
ec
tio
n
w
ith
s
ema
n
tic
…
(
S
iva
Dh
ieva
r
a
j
)
445
ℎ
=
(
⃛
)
,
=
768
to
i
m
p
r
o
v
e
s
tab
ilit
y
a
n
d
r
ed
u
c
e
r
ed
u
n
d
an
c
y
,
e
m
b
ed
d
in
g
s
ar
e
r
ef
in
ed
v
ia
a
f
ea
tu
r
e
s
elec
tio
n
o
p
er
ato
r
Φ
in
s
p
ir
ed
b
y
s
tab
ilit
y
-
o
r
ie
n
ted
s
elec
tio
n
m
et
h
o
d
s:
̂
=
(
{
ℎ
1
,
ℎ
2
,
…
,
ℎ
}
)
B
ey
o
n
d
e
m
b
ed
d
in
g
s
,
t
h
is
la
y
er
in
teg
r
ate
s
s
t
r
u
ct
u
r
ed
b
io
m
e
d
ical
k
n
o
w
led
g
e
b
y
m
ap
p
in
g
en
titi
e
s
—
d
is
ea
s
es,
d
r
u
g
s
,
p
r
o
tein
s
,
s
y
m
p
to
m
s
—
o
n
to
a
k
n
o
w
led
g
e
g
r
ap
h
G=
(
V,
E
)
,
w
h
er
e
n
o
d
es
∈
r
ep
r
esen
t
en
t
ities
,
an
d
ed
g
es
=
(
,
,
)
∈
r
ep
r
esen
t
ca
u
s
al,
th
er
ap
eu
tic,
o
r
m
o
lec
u
lar
r
elatio
n
s
h
ip
s
.
E
ac
h
n
o
d
e
is
ass
o
ciate
d
w
it
h
a
n
e
m
b
ed
d
in
g
ℎ
,
p
r
o
v
id
in
g
co
n
te
x
tu
a
l a
n
d
r
elatio
n
al
g
r
o
u
n
d
i
n
g
f
o
r
d
o
w
n
s
tr
ea
m
clas
s
if
icatio
n
.
4.
3
.
G
ra
ph
neura
l net
w
o
rk
–
ba
s
ed
c
la
s
s
if
ica
t
io
n a
nd
e
x
pl
a
ina
bil
it
y
la
y
e
r
T
h
is
la
y
er
f
o
r
m
s
th
e
a
n
al
y
tic
al
en
g
i
n
e,
co
m
b
i
n
i
n
g
r
elatio
n
al
r
ea
s
o
n
in
g
w
it
h
a
d
u
al
-
p
h
a
s
e
lear
n
in
g
m
ec
h
a
n
i
s
m
:
P
h
ase
1
–
GA
T
r
ea
s
o
n
in
g
:
T
h
e
em
b
ed
d
in
g
s
̂
an
d
s
u
b
g
r
ap
h
s
⊂
ar
e
p
r
o
ce
s
s
ed
th
r
o
u
g
h
a
GA
T
,
w
h
er
e
n
o
d
e
r
ep
r
esen
tatio
n
s
p
r
o
p
ag
ate
ac
co
r
d
in
g
to
t
h
e
r
el
ev
an
ce
o
f
n
ei
g
h
b
o
r
in
g
n
o
d
es:
ℎ
+
1
=
(
∑
ℎ
∈
(
)
)
A
tte
n
tio
n
co
ef
f
icie
n
t
ar
e
co
m
p
u
ted
as
:
=
(
(
[
ℎ
∥
ℎ
]
)
)
∑
(
(
[
ℎ
∥
ℎ
]
)
)
(
)
T
h
is
m
e
c
h
an
i
s
m
p
r
io
r
itizes
m
ed
icall
y
s
i
g
n
i
f
ica
n
t
ed
g
e
s
w
h
ile
a
tten
u
ati
n
g
les
s
i
n
f
o
r
m
ati
v
e
co
n
n
ec
tio
n
s
,
allo
w
i
n
g
t
h
e
m
o
d
el
to
d
etec
t su
b
tle
r
elatio
n
al
i
n
co
n
s
is
te
n
cie
s
.
P
h
ase
2
–
r
ef
in
e
m
en
t
w
it
h
C
N
N:
Am
b
i
g
u
o
u
s
p
r
ed
ictio
n
s
f
r
o
m
P
h
ase
1
ar
e
r
ea
s
s
ess
ed
v
ia
a
C
NN
m
o
d
u
le
ca
p
tu
r
i
n
g
l
o
n
g
-
r
an
g
e
d
ep
en
d
en
cies
i
n
th
e
te
x
t.
T
h
is
en
h
a
n
ce
s
r
o
b
u
s
tn
e
s
s
,
p
ar
ticu
la
r
l
y
f
o
r
clai
m
s
t
h
at
ar
e
lin
g
u
is
t
i
ca
ll
y
p
lau
s
ib
le
b
u
t
s
e
m
a
n
tical
l
y
i
n
co
n
s
is
te
n
t
w
it
h
th
e
k
n
o
w
led
g
e
g
r
ap
h
.
Fin
al
p
r
ed
ictio
n
s
co
m
b
in
e
b
o
t
h
p
h
ase
s
u
s
in
g
ca
lib
r
ated
ag
g
r
eg
a
tio
n
:
⃛
=
(
⋅
ℎ
+
(
1
−
)
⋅
ℎ
)
,
∈
[
0
,
1
]
E
x
p
lain
ab
ilit
y
m
o
d
u
le
T
o
en
s
u
r
e
in
ter
p
r
etab
ilit
y
,
a
n
e
x
p
lain
ab
ili
t
y
m
o
d
u
le
is
e
m
b
ed
d
ed
d
ir
ec
tly
in
to
th
e
w
o
r
k
f
lo
w
,
g
en
er
ati
n
g
to
k
e
n
-
a
n
d
n
o
d
e
-
lev
el
co
n
tr
ib
u
tio
n
s
co
r
es
(
)
an
d
(
)
:
(
)
=
∑
∑
ℎ
∈
(
)
=
1
th
is
allo
w
s
u
s
er
s
to
tr
ac
e
p
r
ed
ictio
n
s
to
m
ed
icall
y
s
i
g
n
i
f
ican
t
en
titi
e
s
an
d
r
elatio
n
s
h
ip
s
,
p
r
o
v
id
in
g
tr
an
s
p
ar
en
t
an
d
ac
tio
n
ab
le
in
s
i
g
h
t
s
in
to
w
h
y
a
clai
m
i
s
f
la
g
g
ed
as
m
i
s
lea
d
in
g
.
5.
E
XP
E
R
I
M
E
NT
A
L
SE
T
UP
T
o
ev
alu
ate
th
e
p
r
o
p
o
s
ed
GNN
–
b
ased
b
io
m
ed
ical
m
is
in
f
o
r
m
atio
n
d
etec
to
r
,
w
e
d
e
s
ig
n
ed
an
ex
p
er
i
m
e
n
tal
p
ip
elin
e
t
h
at
m
e
asu
r
es
b
o
th
p
r
ed
ictiv
e
p
er
f
o
r
m
an
ce
a
n
d
th
e
m
o
d
el
’
s
ab
ilit
y
to
r
ea
s
o
n
ag
ai
n
s
t
s
tr
u
ct
u
r
ed
b
io
m
ed
ical
k
n
o
w
l
ed
g
e.
A
ll
e
x
p
er
i
m
e
n
ts
u
s
e
th
r
ee
w
id
el
y
cited
b
io
m
ed
ic
al
m
is
i
n
f
o
r
m
atio
n
co
r
p
o
r
a
—
R
eCOVer
y
,
C
o
A
I
D
,
an
d
Hea
lth
Sto
r
y
—
a
u
g
m
en
t
ed
w
it
h
a
s
m
al
l
c
u
r
ated
s
et
o
f
p
ee
r
-
r
ev
ie
w
e
d
b
io
m
ed
ical
s
tate
m
e
n
ts
to
in
cr
ea
s
e
co
v
er
ag
e
o
f
d
o
m
ai
n
-
s
p
ec
if
ic
r
elatio
n
s
.
T
h
e
th
r
ee
p
u
b
lic
d
atasets
p
r
o
v
id
e
d
iv
er
s
e
e
x
a
m
p
le
s
o
f
h
ea
lt
h
-
r
el
ated
m
is
in
f
o
r
m
at
io
n
a
n
d
tr
u
s
t
w
o
r
th
y
r
ep
o
r
tin
g
,
co
v
er
i
n
g
C
OVI
D
-
er
a
clai
m
s
an
d
b
r
o
ad
e
r
b
i
o
m
ed
ical
n
ar
r
ati
v
es.
Fo
r
ea
ch
d
ataset
w
e
p
er
f
o
r
m
e
d
s
tr
atif
ied
s
p
lits
in
to
tr
ain
i
n
g
,
v
alid
atio
n
,
an
d
test
s
ets
u
s
i
n
g
a
7
0
/1
5
/1
5
r
atio
,
en
s
u
r
i
n
g
t
h
at
cla
s
s
p
r
o
p
o
r
tio
n
s
(
tr
u
e
v
s
.
f
al
s
e)
r
e
m
ai
n
co
n
s
is
te
n
t
ac
r
o
s
s
s
p
lit
s
.
I
n
ad
d
itio
n
to
th
e
s
in
g
le
-
s
p
lit
e
v
al
u
atio
n
,
w
e
r
an
5
-
f
o
ld
cr
o
s
s
-
v
ali
d
atio
n
o
n
th
e
tr
ain
i
n
g
+v
a
lid
ati
o
n
p
o
r
tio
n
to
ch
ec
k
s
tab
ilit
y
o
f
r
esu
lts
a
n
d
to
tu
n
e
h
y
p
er
p
ar
a
m
eter
s
.
R
a
w
tex
t
i
s
p
r
ep
r
o
ce
s
s
ed
w
i
th
a
lig
h
t
w
eig
h
t
p
ip
elin
e
t
h
at
b
ala
n
ce
s
clea
n
in
g
an
d
s
e
m
a
n
tic
p
r
eser
v
atio
n
.
T
h
e
p
ip
elin
e
r
e
m
o
v
es
n
o
n
-
i
n
f
o
r
m
ati
v
e
to
k
e
n
s
s
u
ch
a
s
ex
t
r
an
eo
u
s
HT
ML
a
n
d
n
o
r
m
a
liz
es
w
h
i
tesp
ac
e,
b
u
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
9
-
4864
I
n
t J
R
ec
o
n
f
i
g
u
r
ab
le
&
E
m
b
ed
d
ed
Sy
s
t
,
Vo
l.
1
5
,
No
.
2
,
J
u
l
y
202
6
:
439
-
4
4
9
446
in
te
n
tio
n
al
l
y
p
r
eser
v
es
p
u
n
ct
u
atio
n
a
n
d
n
u
m
er
ic
e
x
p
r
ess
io
n
s
b
ec
a
u
s
e
t
h
ese
o
f
te
n
ca
r
r
y
m
ed
icall
y
r
elev
a
n
t
in
f
o
r
m
atio
n
(
e.
g
.
,
d
o
s
ag
e
s
an
d
ti
m
e
in
ter
v
als).
T
o
k
en
izatio
n
an
d
s
u
b
w
o
r
d
en
co
d
in
g
ar
e
p
er
f
o
r
m
ed
u
s
i
n
g
t
h
e
to
k
en
izer
th
at
ac
co
m
p
a
n
ies
t
h
e
ch
o
s
en
tr
an
s
f
o
r
m
er
m
o
d
el
(
B
io
B
E
R
T
f
o
r
b
io
m
ed
ical
ex
p
er
i
m
en
ts
an
d
B
E
R
T
-
b
ase
f
o
r
m
o
r
e
g
e
n
er
al
b
aselin
e
s
)
.
W
e
ap
p
ly
s
e
n
ten
ce
s
eg
m
e
n
t
atio
n
an
d
t
h
en
u
s
e
a
d
o
m
ai
n
-
s
p
ec
if
ic
n
a
m
ed
e
n
tit
y
r
ec
o
g
n
itio
n
(
NE
R
)
m
o
d
el
f
i
n
e
-
t
u
n
ed
o
n
b
io
m
ed
ical
co
r
p
o
r
a
to
ex
tr
ac
t
en
t
ities
s
u
c
h
a
s
d
is
e
ases
,
d
r
u
g
s
,
g
e
n
e
s
,
s
y
m
p
to
m
s
,
an
d
clin
ical
o
u
t
co
m
e
s
.
E
x
tr
ac
ted
en
tit
ies
a
r
e
n
o
r
m
al
ized
u
s
i
n
g
co
n
tr
o
lled
v
o
ca
b
u
lar
ies
(
UM
L
S/
Me
SH id
en
ti
f
ier
s
w
h
e
r
e
p
o
s
s
ib
le)
to
r
ed
u
ce
s
y
n
o
n
y
m
y
an
d
en
ab
le
r
eliab
le
lin
k
i
n
g
in
to
th
e
k
n
o
w
led
g
e
g
r
ap
h
.
T
h
e
B
KG
is
ass
e
m
b
led
b
y
m
er
g
in
g
c
u
r
ated
s
o
u
r
ce
s
(
UM
L
S,
Me
SH,
Dr
u
g
B
an
k
,
an
d
th
e
C
o
m
p
ar
at
iv
e
T
o
x
ico
g
en
o
m
ics
Data
b
ase)
w
i
th
r
elatio
n
tr
ip
le
s
m
i
n
ed
f
r
o
m
a
co
llectio
n
o
f
v
er
if
ied
b
io
m
e
d
ical
p
u
b
licatio
n
s
.
No
d
es
r
ep
r
esen
t
n
o
r
m
alize
d
b
io
m
ed
ical
co
n
ce
p
ts
,
an
d
ed
g
e
s
en
co
d
e
o
n
to
lo
g
ical
r
elatio
n
s
(
is
-
a,
p
ar
t
-
o
f
)
a
n
d
f
ac
t
u
al
ass
o
ciatio
n
s
(
tr
ea
ts
,
ca
u
s
e
s
,
co
n
t
r
ain
d
icate
d
_
w
ith
,
i
n
ter
ac
ts
_
w
i
th
)
.
Fo
r
ea
ch
clai
m
in
th
e
d
atasets
,
w
e
ex
tr
ac
t
a
clai
m
-
ce
n
ter
ed
s
u
b
g
r
ap
h
b
y
m
ap
p
in
g
e
n
titi
e
s
i
n
t
h
e
clai
m
to
n
o
d
es
i
n
t
h
e
B
KG
a
n
d
r
etr
iev
i
n
g
d
ir
ec
t
n
eig
h
b
o
r
s
an
d
m
u
lti
-
h
o
p
p
ath
s
u
p
to
len
g
th
t
h
r
ee
.
T
h
is
s
u
b
g
r
ap
h
,
a
u
g
m
e
n
ted
w
i
t
h
cla
i
m
-
le
v
el
co
n
te
x
t
u
al
e
m
b
ed
d
in
g
s
,
f
o
r
m
s
t
h
e
in
p
u
t t
o
th
e
GNN.
W
e
co
m
p
ar
e
th
e
p
r
o
p
o
s
ed
G
A
T
-
b
ased
ar
ch
itect
u
r
e
w
it
h
a
s
u
ite
o
f
b
aseli
n
es
th
at
r
e
f
lec
t
co
m
m
o
n
ap
p
r
o
ac
h
es
in
t
h
e
liter
at
u
r
e.
B
aselin
e
s
y
s
te
m
s
i
n
cl
u
d
e:
i
)
t
h
e
L
ST
M
-
SGD
m
o
d
el
f
r
o
m
t
h
e
u
s
er
’
s
p
r
i
o
r
w
o
r
k
(re
-
i
m
p
le
m
e
n
ted
an
d
tu
n
ed
h
er
e
f
o
r
co
n
s
is
te
n
c
y
)
;
ii
)
a
C
NN
-
b
ased
clas
s
i
f
ier
;
iii
)
a
f
i
n
e
-
t
u
n
ed
tr
an
s
f
o
r
m
er
b
aselin
e
(
B
io
B
E
R
T
)
th
at
p
er
f
o
r
m
s
te
x
t
-
o
n
l
y
clas
s
i
f
icatio
n
;
iv
)
a
h
y
b
r
id
tex
t+
f
ea
t
u
r
e
m
o
d
el
th
at
co
n
ca
ten
a
tes
tr
an
s
f
o
r
m
er
e
m
b
ed
d
in
g
s
w
i
th
m
a
n
u
a
ll
y
en
g
i
n
ee
r
ed
f
ea
tu
r
e
s
(
r
ea
d
ab
ilit
y
,
s
e
n
ti
m
e
n
t,
m
et
ad
ata)
;
an
d
v
)
t
w
o
alter
n
ati
v
e
GNN
v
ar
ia
n
ts
—
a
v
an
i
lla
g
r
a
p
h
co
n
v
o
lu
t
io
n
al
n
e
t
w
o
r
k
(
G
C
N)
an
d
a
GA
T
ab
lat
ed
o
f
th
e
s
e
m
a
n
ti
c
co
n
s
is
ten
c
y
m
o
d
u
le.
T
h
is
s
et
allo
w
s
d
ir
ec
t
co
m
p
ar
is
o
n
ac
r
o
s
s
te
x
t
-
o
n
l
y
,
h
y
b
r
id
,
an
d
g
r
ap
h
-
b
a
s
ed
p
ar
ad
ig
m
s
,
an
d
is
o
lates t
h
e
co
n
tr
ib
u
tio
n
o
f
th
e
k
n
o
w
led
g
e
-
g
r
ap
h
a
n
d
s
e
m
an
tic
-
ch
ec
k
i
n
g
co
m
p
o
n
en
t
s.
T
r
ain
in
g
an
d
h
y
p
er
p
ar
a
m
eter
s
w
er
e
ch
o
s
e
n
af
ter
g
r
id
s
ea
r
c
h
es
o
n
t
h
e
v
alid
atio
n
f
o
ld
s
an
d
th
r
o
u
g
h
in
s
p
ec
tio
n
o
f
lear
n
i
n
g
c
u
r
v
e
s
.
T
h
e
tr
an
s
f
o
r
m
er
e
m
b
ed
d
in
g
s
(
B
io
B
E
R
T
)
ar
e
f
in
e
-
tu
n
ed
w
it
h
a
lear
n
in
g
r
ate
o
f
2
×1
0
⁻⁵,
b
atch
s
ize
1
6
,
an
d
u
p
to
5
ep
o
ch
s
w
it
h
ea
r
l
y
s
to
p
p
in
g
(
p
atien
ce
=3
)
.
T
h
e
GNN
co
m
p
o
n
en
t
u
s
e
s
P
y
T
o
r
ch
Geo
m
etr
ic
an
d
i
s
tr
ain
ed
w
i
th
A
d
a
m
o
p
ti
m
izer
(
lear
n
in
g
r
a
te
1
×1
0
⁻³,
w
e
ig
h
t
d
ec
a
y
1
×1
0
⁻⁵)
.
T
h
e
GA
T
u
s
es
th
r
ee
atte
n
tio
n
la
y
er
s
w
i
th
4
at
ten
tio
n
h
ea
d
s
p
er
la
y
er
an
d
a
h
id
d
en
d
i
m
e
n
s
io
n
o
f
1
2
8
.
Dr
o
p
o
u
t
o
f
0
.
3
is
ap
p
lied
b
et
w
ee
n
la
y
er
s
,
an
d
R
e
L
U
ac
t
iv
atio
n
s
ar
e
u
s
ed
f
o
r
n
o
n
-
li
n
e
ar
it
y
.
W
e
tr
ain
f
o
r
u
p
to
5
0
ep
o
ch
s
f
o
r
th
e
GNN,
ap
p
ly
i
n
g
ea
r
l
y
s
to
p
p
in
g
w
i
th
p
atien
ce
s
et
to
7
b
ased
o
n
v
alid
atio
n
F1
.
Mi
n
i
-
b
atch
g
r
ap
h
s
a
m
p
lin
g
is
ap
p
lied
to
h
an
d
le
lar
g
e
g
r
ap
h
s
;
ea
c
h
b
atc
h
co
n
tai
n
s
u
p
to
3
2
clai
m
-
ce
n
t
er
ed
s
u
b
g
r
ap
h
s
.
All
e
x
p
er
i
m
e
n
ts
ar
e
r
ep
ea
ted
w
it
h
f
i
v
e
d
if
f
er
en
t r
a
n
d
o
m
s
ee
d
s
(
0
,
4
2
,
1
2
3
,
2
0
2
2
,
9
9
9
)
a
n
d
w
e
r
ep
o
r
t m
ea
n
an
d
s
ta
n
d
ar
d
d
ev
iatio
n
f
o
r
all
p
r
i
m
ar
y
m
etr
ics.
6.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
ex
p
er
im
e
n
tal
e
v
alu
at
io
n
d
em
o
n
s
tr
ate
s
th
at
co
m
b
i
n
i
n
g
s
tr
u
ctu
r
ed
b
io
m
ed
ical
k
n
o
w
l
ed
g
e
w
it
h
g
r
ap
h
-
b
ased
r
ea
s
o
n
i
n
g
s
i
g
n
if
ic
an
tl
y
i
m
p
r
o
v
es m
i
s
in
f
o
r
m
at
io
n
d
etec
tio
n
co
m
p
ar
ed
to
co
n
v
e
n
tio
n
al
tex
t
-
ce
n
tr
ic
ap
p
r
o
ac
h
es.
T
h
e
p
r
o
p
o
s
ed
GNN
-
b
ased
ar
c
h
itect
u
r
e
w
a
s
ev
al
u
at
ed
o
n
t
h
r
ee
b
en
c
h
m
ar
k
b
io
m
ed
ical
m
is
in
f
o
r
m
at
io
n
d
ataset
s
—
R
e
C
OVe
r
y
,
C
o
A
I
D,
an
d
Hea
l
th
Sto
r
y
—
u
s
i
n
g
s
ta
n
d
ar
d
class
i
f
icatio
n
m
etr
ic
s
,
in
cl
u
d
in
g
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
m
ac
r
o
F1
-
s
co
r
e.
T
ab
le
1
p
r
o
v
id
es
a
co
m
p
ar
is
o
n
o
f
ex
is
t
in
g
ap
p
r
o
ac
h
es
in
ter
m
s
o
f
th
e
ir
u
n
d
er
l
y
in
g
ar
ch
i
tectu
r
e,
u
s
e
o
f
d
o
m
ai
n
k
n
o
w
led
g
e,
an
d
ap
p
licatio
n
d
o
m
ai
n
.
Mo
s
t
ex
i
s
ti
n
g
m
et
h
o
d
s
p
r
im
ar
il
y
r
el
y
o
n
tex
t
u
al
f
ea
t
u
r
es
an
d
DL
m
o
d
el
s
,
in
cl
u
d
i
n
g
CN
N
,
N
L
P
tech
n
iq
u
e
s
,
an
d
d
is
tr
ib
u
te
d
ML
f
r
a
m
e
w
o
r
k
s
.
Ho
w
e
v
er
,
th
ese
ap
p
r
o
ac
h
es
g
en
er
all
y
d
o
n
o
t
in
co
r
p
o
r
ate
s
tr
u
ctu
r
ed
b
io
m
ed
ical
k
n
o
w
led
g
e
to
s
u
p
p
o
r
t
s
e
m
a
n
tic
r
ea
s
o
n
i
n
g
.
I
n
co
n
tr
a
s
t,
t
h
e
p
r
o
p
o
s
ed
f
r
a
m
e
w
o
r
k
i
n
te
g
r
ate
s
G
A
T
w
it
h
a
B
KG
,
e
n
ab
lin
g
t
h
e
m
o
d
el
to
ca
p
tu
r
e
co
m
p
lex
r
elatio
n
s
h
ip
s
a
m
o
n
g
b
io
m
ed
ical
e
n
titi
e
s
an
d
i
m
p
r
o
v
e
m
i
s
i
n
f
o
r
m
atio
n
d
etec
tio
n
p
er
f
o
r
m
an
ce
.
T
h
e
p
r
o
p
o
s
ed
GNN
m
o
d
el
w
a
s
ev
alu
ated
o
n
th
e
R
e
C
OVe
r
y
,
C
o
A
I
D,
an
d
Hea
lth
Sto
r
y
d
at
asets
.
T
h
e
r
esu
lt
s
,
p
r
esen
ted
in
T
a
b
le
2
,
i
n
clu
d
e
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
e
ca
ll,
an
d
m
ac
r
o
F1
-
s
co
r
es
an
d
co
m
p
ar
e
th
e
GNN
ag
ain
s
t tr
a
n
s
f
o
r
m
er
an
d
C
N
N
b
aselin
es.
T
ab
le
1
.
C
o
m
p
ar
is
o
n
o
f
p
r
o
p
o
s
ed
m
o
d
el
w
i
th
p
r
io
r
w
o
r
k
i
n
m
is
in
f
o
r
m
at
io
n
d
etec
tio
n
S
t
u
d
y
M
e
t
h
o
d
o
l
o
g
y
K
n
o
w
l
e
d
g
e
i
n
t
e
g
r
a
t
i
o
n
A
p
p
l
i
c
a
t
i
o
n
d
o
ma
i
n
L
e
e
e
t
a
l
.
[
1
0
]
(
2
0
1
9
)
D
L
f
a
k
e
n
e
w
s d
e
t
e
c
t
i
o
n
No
F
a
k
e
n
e
w
s d
e
t
e
c
t
i
o
n
V
y
s
o
t
sk
a
e
t
a
l
.
[
6
]
(
2
0
2
4
)
N
L
P
+
DL
No
F
a
k
e
n
e
w
s d
e
t
e
c
t
i
o
n
A
l
t
h
e
n
e
y
a
n
a
n
d
A
l
h
a
d
l
a
q
[
7
]
(
2
0
2
3
)
D
i
st
r
i
b
u
t
e
d
M
L
No
F
a
k
e
n
e
w
s d
e
t
e
c
t
i
o
n
P
r
o
p
o
se
d
mo
d
e
l
GAT
+
C
N
N
F
u
l
l
b
i
o
me
d
i
c
a
l
K
G
B
i
o
me
d
i
c
a
l
m
i
si
n
f
o
r
mat
i
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
R
ec
o
n
f
i
g
u
r
ab
le
&
E
m
b
ed
d
ed
Sy
s
t
I
SS
N:
2089
-
4864
Gra
p
h
n
eu
r
a
l n
etw
o
r
k
-
b
a
s
ed
b
io
med
ica
l
mis
in
fo
r
ma
tio
n
d
et
ec
tio
n
w
ith
s
ema
n
tic
…
(
S
iva
Dh
ieva
r
a
j
)
447
T
ab
le
2
.
P
er
f
o
r
m
a
n
ce
co
m
p
ar
i
s
o
n
o
n
b
en
c
h
m
ar
k
d
ataset
s
D
a
t
a
se
t
M
o
d
e
l
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
si
o
n
(
%)
R
e
c
a
l
l
(
%)
M
a
c
r
o
-
F
1
(
%)
R
e
C
O
V
e
r
y
B
i
o
B
E
R
T
8
6
.
4
8
4
.
7
8
2
.
1
8
3
.
4
C
N
N
8
2
.
1
8
0
.
5
7
8
.
0
7
9
.
2
P
r
o
p
o
se
d
G
N
N
9
1
.
2
9
0
.
5
8
9
.
1
8
9
.
8
C
o
A
I
D
B
i
o
B
E
R
T
8
4
.
0
8
3
.
2
8
0
.
4
8
1
.
8
C
N
N
8
0
.
2
7
8
.
9
7
6
.
5
7
7
.
7
P
r
o
p
o
se
d
G
N
N
8
8
.
5
8
7
.
9
8
6
.
2
8
7
.
0
H
e
a
l
t
h
S
t
o
r
y
B
i
o
B
E
R
T
8
1
.
7
8
0
.
1
7
8
.
3
7
9
.
2
C
N
N
7
8
.
4
7
7
.
0
7
5
.
6
7
6
.
3
P
r
o
p
o
se
d
G
N
N
8
5
.
6
8
4
.
8
8
3
.
5
8
4
.
1
T
ab
le
2
h
ig
h
li
g
h
ts
t
h
at
th
e
p
r
o
p
o
s
ed
GNN
co
n
s
is
te
n
tl
y
o
u
tp
er
f
o
r
m
s
b
aseli
n
es
ac
r
o
s
s
d
atasets
,
p
ar
tic
u
lar
l
y
i
n
m
ac
r
o
-
F1
,
in
d
ic
atin
g
b
alan
ce
d
p
er
f
o
r
m
a
n
ce
o
n
b
o
th
clas
s
es o
f
m
is
in
f
o
r
m
at
i
o
n
.
6
.
1
.
Abla
t
io
n
s
t
ud
y
re
s
ults
T
o
in
v
est
ig
ate
t
h
e
co
n
tr
ib
u
tio
n
o
f
ea
ch
co
m
p
o
n
e
n
t
in
t
h
e
p
r
o
p
o
s
ed
ar
ch
itectu
r
e,
an
ab
latio
n
s
tu
d
y
w
a
s
co
n
d
u
cted
.
Ke
y
m
o
d
u
les w
er
e
r
e
m
o
v
ed
o
r
r
ep
lace
d
,
an
d
t
h
e
r
esu
lti
n
g
m
ac
r
o
F1
-
s
co
r
es
ar
e
r
ep
o
r
ted
in
T
ab
le
3
.
T
ab
le
3
.
A
b
latio
n
s
t
u
d
y
o
n
m
o
d
el
co
m
p
o
n
e
n
ts
(
R
eCO
Ver
y
d
ataset)
R
e
mo
v
e
d
mo
d
u
l
e
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
si
o
n
(
%)
R
e
c
a
l
l
(
%)
M
a
c
r
o
-
F
1
(
%)
N
o
n
e
(
f
u
l
l
mo
d
e
l
)
9
1
.
2
9
0
.
5
8
9
.
1
8
9
.
8
K
n
o
w
l
e
d
g
e
g
r
a
p
h
8
7
.
0
8
6
.
2
8
4
.
0
8
5
.
1
T
r
a
n
sf
o
r
me
r
e
mb
e
d
d
i
n
g
s
8
6
.
5
8
5
.
3
8
3
.
8
8
4
.
5
S
e
man
t
i
c
c
o
n
si
st
e
n
c
y
mo
d
u
l
e
8
5
.
2
8
4
.
0
8
2
.
1
8
3
.
0
T
e
mp
o
r
a
l
g
r
a
p
h
l
a
y
e
r
8
9
.
0
8
8
.
2
8
6
.
1
8
7
.
1
T
h
e
tab
le
co
n
f
ir
m
s
t
h
at
th
e
k
n
o
w
led
g
e
g
r
ap
h
an
d
s
e
m
a
n
tic
co
n
s
i
s
te
n
c
y
m
o
d
u
le
ar
e
th
e
m
o
s
t
s
ig
n
i
f
i
ca
n
t
co
n
tr
ib
u
to
r
s
to
m
o
d
el
p
er
f
o
r
m
an
ce
.
6
.
2
.
I
nte
rpre
t
a
bil
it
y
a
na
ly
s
i
s
I
n
ter
p
r
etab
ilit
y
m
etr
ics
w
er
e
q
u
an
t
if
ied
b
y
e
x
p
er
t
ev
al
u
atio
n
o
f
atten
t
io
n
-
b
ased
ex
p
la
n
atio
n
s
.
Do
m
ai
n
ex
p
er
ts
r
ated
th
e
u
s
e
f
u
l
n
es
s
o
f
ex
p
lan
atio
n
s
o
n
a
s
ca
le
o
f
1
–
5
.
T
ab
le
4
p
r
esen
ts
th
e
p
er
ce
n
t
ag
e
o
f
e
x
p
lan
atio
n
s
d
ee
m
ed
“u
s
e
f
u
l”
o
r
“
h
ig
h
l
y
u
s
ef
u
l.”
R
e
s
u
l
ts
i
n
d
icate
t
h
at
o
v
er
7
0
%
o
f
th
e
e
x
p
la
n
atio
n
s
p
r
o
v
id
ed
b
y
th
e
GNN
ar
e
co
n
s
id
er
ed
u
s
ef
u
l,
d
e
m
o
n
s
tr
atin
g
th
e
e
f
f
ec
tiv
e
n
e
s
s
o
f
t
h
e
in
teg
r
ated
e
x
p
lain
ab
ilit
y
m
o
d
u
le.
T
ab
le
4
.
I
n
te
r
p
r
eta
b
ilit
y
ev
al
u
atio
n
ac
r
o
s
s
d
atasets
D
a
t
a
se
t
U
se
f
u
l
(
%)
H
i
g
h
l
y
u
se
f
u
l
(
%)
C
o
mb
i
n
e
d
(
%)
R
e
C
O
V
e
r
y
32
45
77
C
o
A
I
D
28
48
76
H
e
a
l
t
h
S
t
o
r
y
30
43
73
6
.
3
.
Ro
bu
s
t
nes
s
un
der
dis
t
ri
bu
t
io
na
l sh
if
t
s
T
h
e
m
o
d
el’
s
s
tab
ilit
y
u
n
d
er
tem
p
o
r
al
an
d
to
p
ical
s
h
i
f
ts
w
a
s
ev
alu
a
te
d
b
y
test
in
g
o
n
o
u
t
-
of
-
d
is
tr
ib
u
tio
n
clai
m
s
.
T
ab
le
5
s
h
o
w
s
m
ac
r
o
F1
-
s
co
r
es
f
o
r
clai
m
s
e
m
er
g
in
g
af
ter
th
e
tr
ai
n
i
n
g
p
er
io
d
.
T
ab
le
5
.
R
o
b
u
s
tn
es
s
to
d
is
tr
ib
u
tio
n
al
s
h
if
t
s
D
a
t
a
se
t
M
o
d
e
l
In
-
d
i
st
r
i
b
u
t
i
o
n
m
a
c
r
o
-
F
1
(
%)
O
u
t
-
of
-
d
i
s
t
r
i
b
u
t
i
o
n
ma
c
r
o
-
F
1
(
%)
R
e
C
O
V
e
r
y
Bi
o
B
E
R
T
8
3
.
4
7
9
.
1
C
N
N
7
9
.
2
7
4
.
3
P
r
o
p
o
se
d
G
N
N
8
9
.
8
8
7
.
0
C
o
A
I
D
B
i
o
B
E
R
T
8
1
.
8
7
7
.
5
C
N
N
7
7
.
7
7
3
.
2
P
r
o
p
o
se
d
G
N
N
8
7
.
0
8
4
.
1
T
h
e
GNN
m
o
d
el
s
h
o
w
s
m
in
i
m
al
d
eg
r
ad
atio
n
u
n
d
er
d
is
tr
ib
u
tio
n
al
s
h
i
f
t
s
,
co
n
f
ir
m
i
n
g
its
r
elian
ce
o
n
s
tab
le
b
io
m
ed
ical
r
elatio
n
s
r
at
h
er
th
a
n
s
u
r
f
ac
e
li
n
g
u
i
s
tic
f
ea
t
u
r
es
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
9
-
4864
I
n
t J
R
ec
o
n
f
i
g
u
r
ab
le
&
E
m
b
ed
d
ed
Sy
s
t
,
Vo
l.
1
5
,
No
.
2
,
J
u
l
y
202
6
:
439
-
4
4
9
448
7.
CO
NCLU
SI
O
N
I
n
th
i
s
s
t
u
d
y
,
w
e
p
r
o
p
o
s
ed
a
m
u
lti
la
y
er
ed
,
k
n
o
w
led
g
e
-
g
r
o
u
n
d
ed
f
r
a
m
e
w
o
r
k
f
o
r
b
io
m
ed
ical
m
is
in
f
o
r
m
at
io
n
d
etec
tio
n
th
a
t
in
teg
r
ate
s
DL
,
tr
an
s
f
o
r
m
er
-
b
ased
e
m
b
ed
d
in
g
s
,
GNN
s
,
an
d
ex
p
lain
ab
i
li
t
y
m
o
d
u
les.
B
y
co
m
b
i
n
in
g
tex
t
u
al
s
e
m
an
t
ics
w
it
h
r
elatio
n
al
r
ea
s
o
n
in
g
o
v
er
a
s
tr
u
ctu
r
ed
B
KG
,
th
e
s
y
s
te
m
d
em
o
n
s
tr
ate
s
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
ac
r
o
s
s
m
u
ltip
le
b
en
c
h
m
ar
k
d
ata
s
ets,
o
u
tp
er
f
o
r
m
i
n
g
tr
ad
itio
n
al
te
x
t
-
o
n
l
y
an
d
tr
an
s
f
o
r
m
er
-
b
ased
m
o
d
els
in
ac
cu
r
ac
y
,
p
r
e
cisi
o
n
,
r
ec
all,
an
d
m
ac
r
o
F
1
-
s
co
r
es.
T
h
e
d
u
al
-
p
h
ase
ar
ch
itect
u
r
e,
co
m
p
r
is
in
g
g
r
ap
h
-
b
ased
r
ea
s
o
n
in
g
a
n
d
s
eq
u
e
n
tial
r
ef
in
e
m
e
n
t,
en
s
u
r
es
r
o
b
u
s
tn
e
s
s
i
n
d
etec
tin
g
lin
g
u
is
ticall
y
p
lau
s
ib
le
y
et
f
ac
tu
al
l
y
i
n
co
r
r
ec
t c
lai
m
s
,
w
h
ile
atten
tio
n
-
b
ase
d
in
ter
p
r
etab
ilit
y
p
r
o
v
id
es tr
an
s
p
ar
en
c
y
f
o
r
ex
p
er
t
v
alid
atio
n
.
A
b
latio
n
s
tu
d
ies
c
o
n
f
ir
m
th
e
cr
itical
r
o
le
o
f
k
n
o
w
led
g
e
i
n
te
g
r
atio
n
,
s
e
m
an
t
i
c
co
n
s
is
te
n
c
y
,
an
d
te
m
p
o
r
al
p
r
o
p
ag
atio
n
in
m
ai
n
tai
n
in
g
h
i
g
h
p
r
ed
icti
v
e
ac
c
u
r
ac
y
.
A
l
th
o
u
g
h
c
h
alle
n
g
e
s
r
e
m
ain
in
h
a
n
d
lin
g
e
m
er
g
i
n
g
b
io
m
ed
ic
al
c
lai
m
s
an
d
m
etap
h
o
r
ical
la
n
g
u
a
g
e,
t
h
e
f
r
a
m
e
w
o
r
k
estab
li
s
h
e
s
a
r
o
b
u
s
t
f
o
u
n
d
atio
n
f
o
r
r
eliab
le
an
d
ex
p
lai
n
ab
le
m
is
i
n
f
o
r
m
atio
n
d
etec
t
io
n
i
n
t
h
e
b
io
m
ed
ical
d
o
m
ai
n
.
F
u
t
u
r
e
r
es
ea
r
ch
w
ill
f
o
cu
s
o
n
ex
ten
d
i
n
g
t
h
e
f
r
a
m
e
w
o
r
k
t
h
r
o
u
g
h
m
u
lti
m
o
d
al
d
ata
f
u
s
io
n
,
i
n
co
r
p
o
r
atin
g
co
m
p
le
m
e
n
tar
y
m
ed
ical
d
ata
s
o
u
r
ce
s
s
u
c
h
as
cl
in
ica
l
i
m
ag
e
s
,
b
io
m
ed
ical
liter
at
u
r
e,
an
d
ele
ctr
o
n
ic
h
ea
lt
h
r
ec
o
r
d
s
to
s
t
r
en
g
t
h
e
n
co
n
te
x
t
u
al
u
n
d
er
s
ta
n
d
in
g
.
F
u
r
th
er
i
n
v
e
s
ti
g
atio
n
o
f
ad
v
an
ce
d
tr
an
s
f
o
r
m
er
ar
ch
itectu
r
es
a
n
d
v
is
io
n
-
lan
g
u
a
g
e
m
o
d
el
s
m
a
y
i
m
p
r
o
v
e
th
e
ca
p
ab
ilit
y
o
f
ca
p
t
u
r
in
g
co
m
p
le
x
b
io
m
ed
ical
n
ar
r
ativ
es.
F
UNDIN
G
I
NF
O
RM
AT
I
O
N
Au
t
h
o
r
s
s
tate
n
o
f
u
n
d
i
n
g
i
n
v
o
l
v
ed
.
AUTHO
R
CO
NT
RIB
UT
I
O
NS ST
A
T
E
M
E
NT
T
h
is
jo
u
r
n
al
u
s
e
s
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT
)
t
o
r
ec
o
g
n
ize
in
d
i
v
id
u
al
au
t
h
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
t
h
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
lla
b
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
Siv
a
D
h
ie
v
ar
aj
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Ag
u
s
t
h
i
y
ar
R
a
m
u
✓
✓
✓
✓
✓
✓
✓
✓
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
si
s
I
:
I
n
v
e
st
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
si
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
t
h
o
r
s
s
tate
n
o
co
n
f
lic
t o
f
i
n
t
er
est.
DATA AV
AI
L
AB
I
L
I
T
Y
Data
av
ailab
ilit
y
i
s
n
o
t a
p
p
licab
le
to
th
is
p
ap
er
as n
o
n
e
w
d
ata
w
er
e
cr
ea
ted
o
r
an
al
y
ze
d
in
th
is
s
t
u
d
y
.
RE
F
E
R
E
NC
E
S
[
1
]
A
.
S
.
Jara
d
a
t
e
t
a
l.
,
“
A
u
t
o
ma
t
e
d
mo
n
k
e
y
p
o
x
sk
i
n
l
e
si
o
n
d
e
t
e
c
t
i
o
n
u
si
n
g
d
e
e
p
l
e
a
r
n
i
n
g
a
n
d
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
E
n
v
i
ro
n
m
e
n
t
a
l
Re
s
e
a
r
c
h
a
n
d
P
u
b
l
i
c
H
e
a
l
t
h
,
v
o
l
.
2
0
,
n
o
.
5
,
M
a
r
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
i
j
e
r
p
h
2
0
0
5
4
4
2
2
.
[
2
]
L
.
S
a
w
a
l
h
a
a
n
d
T
.
C
.
A
k
i
n
c
i
,
“
S
h
a
l
l
o
w
l
e
a
r
n
i
n
g
v
e
r
su
s
d
e
e
p
l
e
a
r
n
i
n
g
i
n
n
a
t
u
r
a
l
l
a
n
g
u
a
g
e
p
r
o
c
e
ssi
n
g
a
p
p
l
i
c
a
t
i
o
n
s,”
i
n
S
h
a
l
l
o
w
L
e
a
r
n
i
n
g
v
s
.
D
e
e
p
L
e
a
rn
i
n
g
,
2
0
2
4
,
p
p
.
1
7
9
–
2
0
6
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
031
-
6
9
4
9
9
-
8
_
8
.
[
3
]
G
.
H
u
a
n
g
,
Y
.
L
i
,
S
.
Jame
e
l
,
Y
.
L
o
n
g
,
a
n
d
G
.
P
a
p
a
n
a
s
t
a
s
i
o
u
,
“
F
r
o
m
e
x
p
l
a
i
n
a
bl
e
t
o
i
n
t
e
r
p
r
e
t
a
b
l
e
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
n
a
t
u
r
a
l
l
a
n
g
u
a
g
e
p
r
o
c
e
ssi
n
g
i
n
h
e
a
l
t
h
c
a
r
e
:
H
o
w
f
a
r
f
r
o
m
r
e
a
l
i
t
y
?
,
”
C
o
m
p
u
t
a
t
i
o
n
a
l
a
n
d
S
t
ru
c
t
u
ra
l
B
i
o
t
e
c
h
n
o
l
o
g
y
J
o
u
r
n
a
l
,
v
o
l
.
2
4
,
p
p
.
3
6
2
–
3
7
3
,
D
e
c
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
s
b
j
.
2
0
2
4
.
0
5
.
0
0
4
.
[
4
]
M
.
Q
.
A
l
n
a
b
h
a
n
a
n
d
P
.
B
r
a
n
c
o
,
“
F
a
k
e
n
e
w
s
d
e
t
e
c
t
i
o
n
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
:
A
sy
st
e
mat
i
c
l
i
t
e
r
a
t
u
r
e
r
e
v
i
e
w
,
”
I
EE
E
Ac
c
e
ss
,
v
o
l
.
1
2
,
p
p
.
1
1
4
4
3
5
–
1
1
4
4
5
9
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
4
.
3
4
3
5
4
9
7
.
[
5
]
Q
.
C
h
e
n
e
t
a
l
.
,
“
A
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
i
n
a
c
t
i
o
n
:
A
d
d
r
e
ssi
n
g
t
h
e
C
O
V
I
D
-
1
9
p
a
n
d
e
mi
c
w
i
t
h
n
a
t
u
ra
l
l
a
n
g
u
a
g
e
p
r
o
c
e
ssi
n
g
,
”
An
n
u
a
l
Re
v
i
e
w
o
f
Bi
o
m
e
d
i
c
a
l
D
a
t
a
S
c
i
e
n
c
e
,
v
o
l
.
4
,
n
o
.
1
,
p
p
.
3
1
3
–
3
3
9
,
Ju
l
.
2
0
2
1
,
d
o
i
:
1
0
.
1
1
4
6
/
a
n
n
u
r
e
v
-
b
i
o
d
a
t
a
sc
i
-
0
2
1
8
2
1
-
0
6
1
0
4
5
.
[
6
]
V
.
V
y
s
o
t
sk
a
,
O
.
M
a
r
k
i
v
,
D
.
S
v
y
sh
c
h
,
L
.
C
h
y
r
u
n
,
S
.
C
h
y
r
u
n
,
a
n
d
R
.
R
o
ma
n
c
h
u
k
,
“
F
a
k
e
n
e
w
s
i
d
e
n
t
i
f
i
c
a
t
io
n
b
a
se
d
o
n
N
L
P
,
b
i
g
d
a
t
a
a
n
a
l
y
si
s
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
t
e
c
h
n
o
l
o
g
y
,
”
i
n
2
0
2
4
I
E
EE
1
7
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
A
d
v
a
n
c
e
d
T
r
e
n
d
s
i
n
Ra
d
i
o
e
l
e
c
t
r
o
n
i
c
s
,
T
e
l
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
a
n
d
C
o
m
p
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
(
T
C
S
ET)
,
O
c
t
.
2
0
2
4
,
p
p
.
1
9
9
–
2
0
4
,
d
o
i
:
1
0
.
1
1
0
9
/
T
C
S
ET
6
4
7
2
0
.
2
0
2
4
.
10
7
5
5
9
0
2
.
[
7
]
A
.
A
l
t
h
e
n
e
y
a
n
a
n
d
A
.
A
l
h
a
d
l
a
q
,
“
B
i
g
d
a
t
a
M
L
-
b
a
se
d
f
a
k
e
n
e
w
s
d
e
t
e
c
t
i
o
n
u
si
n
g
d
i
st
r
i
b
u
t
e
d
l
e
a
r
n
i
n
g
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
1
1
,
p
p
.
2
9
4
4
7
–
2
9
4
6
3
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
3
.
3
2
6
0
7
6
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.