I
AE
S
I
nte
rna
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
5
,
No
.
4
,
Dec
em
b
er
2
0
1
6
,
p
p
.
143
~
148
I
SS
N:
2252
-
8938
143
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
jo
u
r
n
a
l.c
o
m/o
n
lin
e/in
d
ex
.
p
h
p
/I
J
A
I
Pars
er Ex
traction
of Triples
in
U
nstructu
red Tex
t
Sh
a
un
D’
So
uza
T
e
c
h
n
ica
l
Lea
d
,
W
ip
ro
L
im
it
e
d
,
Ba
n
g
a
lo
re
,
Ka
rn
a
tak
a
,
In
d
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Au
g
1
4
,
2
0
1
6
R
ev
i
s
ed
Oct
1
8
,
2
0
1
6
A
cc
ep
ted
No
v
2
1
,
2
0
1
6
T
h
e
w
e
b
c
o
n
tain
s
v
a
st
re
p
o
sito
ri
e
s
o
f
u
n
stru
c
tu
re
d
tex
t.
W
e
in
v
e
stig
a
te
th
e
o
p
p
o
rt
u
n
it
y
f
o
r
b
u
il
d
i
n
g
a
k
n
o
w
led
g
e
g
ra
p
h
f
ro
m
th
e
s
e
tex
t
s
o
u
rc
e
s.
W
e
g
e
n
e
ra
te
a
s
e
t
o
f
tri
p
les
w
h
ich
c
a
n
b
e
u
se
d
in
k
n
o
w
led
g
e
g
a
th
e
rin
g
a
n
d
in
teg
ra
ti
o
n
.
W
e
d
e
f
in
e
th
e
a
rc
h
it
e
c
tu
r
e
o
f
a
lan
g
u
a
g
e
c
o
m
p
il
e
r
f
o
r
p
ro
c
e
ss
in
g
su
b
jec
t
-
p
re
d
ica
te
-
o
b
jec
t
tri
p
les
u
sin
g
th
e
Op
e
n
NL
P
p
a
rse
r.
W
e
im
p
le
m
e
n
t
a
d
e
p
th
-
f
irst
se
a
rc
h
trav
e
rsa
l
o
n
th
e
P
OS
tag
g
e
d
s
y
n
tac
ti
c
tree
a
p
p
e
n
d
in
g
p
re
d
ica
te
a
n
d
o
b
jec
t
in
f
o
rm
a
ti
o
n
.
A
p
a
rs
e
r
e
n
a
b
les
h
ig
h
e
r
p
re
c
isio
n
a
n
d
h
ig
h
e
r
re
c
a
ll
e
x
trac
ti
o
n
s
o
f
sy
n
tac
ti
c
re
l
a
ti
o
n
sh
i
p
s
a
c
ro
ss
c
o
n
j
u
n
c
ti
o
n
b
o
u
n
d
a
ries
.
W
e
a
re
a
b
le
to
e
x
trac
t
2
-
2
.
5
t
im
e
s
th
e
c
o
rre
c
t
e
x
trac
ti
o
n
s
o
f
Re
V
e
rb
.
T
h
e
e
x
trac
ti
o
n
s
a
re
u
se
d
in
a
v
a
riet
y
o
f
s
e
m
a
n
ti
c
w
e
b
a
p
p
li
c
a
ti
o
n
s
a
n
d
q
u
e
stio
n
a
n
sw
e
rin
g
.
W
e
v
e
rify
e
x
trac
ti
o
n
o
f
5
0
,
0
0
0
tr
ip
l
e
s
o
n
t
h
e
Clu
e
W
e
b
d
a
tas
e
t.
K
ey
w
o
r
d
:
NL
P
Op
en
in
f
o
r
m
atio
n
ex
tr
ac
t
io
n
R
elatio
n
e
x
tr
ac
tio
n
Co
p
y
rig
h
t
©
2
0
1
6
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Sh
a
u
n
D
’
So
u
za
,
T
ec
h
n
ical
L
ea
d
(
C
T
O
Of
f
ice
)
W
ip
r
o
L
i
m
ited
,
B
an
g
alo
r
e
,
Kar
n
ata
k
a
,
I
n
d
ia
E
m
ail:
s
h
a
u
n
.
d
s
o
u
za
1
@
w
ip
r
o
.
co
m
1.
I
NT
RO
D
UCT
I
O
N
T
h
er
e
is
a
co
n
s
id
er
ab
le
a
m
o
u
n
t
o
f
r
esear
c
h
i
n
n
atu
r
al
lan
g
u
a
g
e
p
r
o
ce
s
s
i
n
g
(
N
L
P
)
.
W
ith
t
h
e
av
ailab
ilit
y
o
f
a
lar
g
er
s
et
o
f
NL
P
to
o
ls
li
k
e
Op
en
N
L
P
[
3
]
,
it
is
to
d
a
y
p
o
s
s
ib
le
to
P
OS
t
ag
a
n
d
ch
u
n
k
v
a
s
t
a
m
o
u
n
t
o
f
u
n
s
tr
u
ctu
r
ed
te
x
t
t
h
at
is
a
v
ailab
le
o
n
t
h
e
i
n
ter
n
e
t.
P
r
o
j
ec
ts
lik
e
C
l
u
eW
eb
,
Op
e
n
I
E
an
d
W
ik
ip
ed
ia
p
r
o
v
id
e
a
co
r
p
u
s
o
f
te
x
t
d
ata
w
h
ic
h
ca
n
b
e
u
s
ed
f
o
r
o
n
to
l
o
g
ical
e
n
g
in
ee
r
i
n
g
.
Op
en
N
L
P
s
u
p
p
o
r
ts
th
e
P
OS
tag
g
in
g
an
d
ch
u
n
k
i
n
g
o
f
d
ata.
I
t
o
u
tp
u
ts
a
p
ar
s
e
tr
ee
f
o
r
th
e
d
ata
w
h
ic
h
en
ca
p
s
u
late
s
th
e
s
y
n
tac
tic
co
n
te
n
t
in
a
n
-
ar
y
tr
ee
d
ata
s
tr
u
ctu
r
e.
P
OS
tag
d
ata
p
r
o
v
id
es
a
h
ig
h
er
lev
el
o
f
u
n
d
er
s
ta
n
d
in
g
as
co
m
p
ar
ed
to
a
b
ag
o
f
w
o
r
d
s
ap
p
r
o
ac
h
to
w
eb
s
ea
r
c
h
to
d
ay
.
W
e
ex
p
lo
r
e
o
p
p
o
r
tu
n
itie
s
f
o
r
lan
g
u
a
g
e
in
f
er
en
ce
an
d
u
n
d
er
s
ta
n
d
in
g
th
r
o
u
g
h
s
u
b
j
ec
t
-
p
r
ed
icate
-
o
b
j
ec
t a
n
al
y
s
is
o
f
w
eb
s
ca
le
u
n
s
tr
u
ctu
r
ed
d
ata.
Var
io
u
s
m
et
h
o
d
s
ar
e
u
s
ed
to
ex
tr
ac
t
s
u
b
j
ec
t
-
p
r
ed
icate
-
o
b
j
ec
t
tr
ip
les
i
n
u
n
s
tr
u
ct
u
r
ed
d
ata.
DB
p
ed
ia
ex
tr
ac
to
r
is
u
s
ed
to
g
e
n
er
ate
tr
ip
les u
s
in
g
a
n
n
o
tated
f
ield
i
n
f
o
r
m
at
io
n
i
n
W
ik
ip
ed
ia.
Op
en
I
E
[
1
]
u
s
ed
P
OS a
n
d
ch
u
n
k
er
d
ata
w
h
ile
C
lau
s
eI
E
[
2
]
u
s
es a
p
ar
s
er
to
o
u
tp
u
t a
s
et
o
f
w
o
r
d
tr
ip
les.
B
o
o
ts
tr
ap
p
in
g
f
u
n
ctio
n
s
u
s
e
N
-
g
r
a
m
m
o
d
els
to
g
e
n
er
ate
a
te
m
p
late
f
o
r
a
g
iv
e
n
co
m
b
i
n
atio
n
o
f
n
o
u
n
p
h
r
ases
.
T
h
ese
ar
e
u
s
ed
to
s
ea
r
ch
a
lar
g
er
co
r
p
u
s
o
f
d
ata
f
o
r
s
i
m
ilar
te
m
p
lates
an
d
g
en
er
ate
v
al
u
es.
NE
R
tag
g
er
s
ar
e
u
s
ed
to
an
n
o
tate
p
er
s
o
n
an
d
lo
ca
tio
n
i
n
f
o
r
m
atio
n
.
W
e
ass
u
m
e
a
co
n
te
x
t f
r
ee
g
r
am
m
ar
(
C
FG)
f
o
r
E
n
g
li
s
h
la
n
g
u
ag
e
[
4
]
.
G
=
(
N
,
Σ
,
R
,
S)
N
∈
{n
o
n
-
ter
m
i
n
al
s
y
m
b
o
ls
}
Σ
∈
{ter
m
i
n
al
s
y
m
b
o
ls
}
R
∈
{r
u
les} o
f
t
h
e
f
o
r
m
X
→
Y
1
Y
n
f
o
r
n
≥
0
,
X
∈
N,
Y
i
∈
(
N
⋃
Σ)
S
∈
N
s
tar
t s
y
m
b
o
l {
T
OP
}
N
=
{S,
NP
,
VP,
PP
,
D
T
,
VB
,
NN,
I
N}
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
IJ
-
AI
Vo
l.
5
,
No
.
4
,
Dec
em
b
e
r
2
0
1
6
:
1
4
3
–
1
4
8
144
S =
S
Σ
=
w
o
r
d
in
t
h
e
E
n
g
li
s
h
la
n
g
u
ag
e
R
=
S
→
NP
VP
VP
→
VB
VP
→
VB
NP
VP
→
VP
PP
NP
→
DT
NN
NP
→
NP
PP
PP
→
I
N
NP
2.
RE
S
E
ARCH
M
E
T
H
O
D
W
e
f
o
u
n
d
a
li
m
itatio
n
o
f
e
x
tr
ac
to
r
s
th
at
w
er
e
u
n
ab
le
to
ex
tr
ac
t
t
h
e
v
er
b
p
h
r
a
s
e
ac
c
u
r
atel
y
a
n
d
in
s
tead
ap
p
en
d
ed
a
lar
g
e
a
m
o
u
n
t
o
f
ad
d
itio
n
al
w
o
r
d
s
i
n
cl
u
d
in
g
th
e
tr
aili
n
g
n
o
u
n
a
n
d
p
r
ep
o
s
itio
n
co
n
te
x
t.
T
h
e
ex
tr
ac
to
r
s
w
er
e
u
n
ab
le
to
p
r
o
ce
s
s
s
e
n
te
n
ce
a
n
d
co
n
j
u
n
cti
o
n
v
a
lu
e
s
r
es
u
lti
n
g
i
n
in
co
r
r
ec
t
v
er
b
a
n
d
o
b
j
ec
t
p
h
r
ases
.
A
p
ar
s
e
tr
ee
is
ab
le
to
ca
p
tu
r
e
co
n
j
u
n
ctio
n
a
n
d
o
b
j
e
ct
p
h
r
ase
in
f
o
r
m
a
tio
n
co
r
r
ec
tly
.
A
lt
h
o
u
g
h
t
h
er
e
i
s
an
o
v
er
h
ea
d
o
n
t
h
e
p
ar
s
in
g
ti
m
e.
W
e
ev
al
u
ate
t
h
e
p
ar
s
er
tr
ee
f
o
r
s
eq
u
en
ce
s
o
f
NP
n
o
u
n
p
h
r
ases
(
s
u
b
j
ec
t,
o
b
j
ec
t
)
an
d
V
B
-
v
er
b
s
(
p
r
ed
icate
)
.
Op
en
NL
P
g
en
er
at
es
a
p
ar
s
e
tr
ee
u
s
in
g
t
h
e
C
FG
r
u
les.
W
e
i
m
p
le
m
e
n
t
a
n
i
n
-
o
r
d
er
tr
av
er
s
al
o
f
t
h
e
s
y
n
tactic
tr
ee
to
d
etec
t
SVO
p
h
r
ases
.
W
e
m
ai
n
tai
n
a
lis
t
o
f
all
NP
p
h
r
ases
in
th
e
s
e
n
ten
ce
.
W
e
th
en
tr
av
er
s
e
th
e
tr
ee
to
d
etec
t su
b
j
ec
t o
b
j
ec
t p
air
s
an
d
th
e
p
r
ed
icate
.
f
u
n
c
t
i
o
n
S
U
B
JEC
T
-
NOUN
-
P
H
R
A
S
E(
p
a
r
se
)
k
i
d
s
←
C
H
I
L
D
(
p
a
rse
)
f
o
r
i
=
1
t
o
S
I
Z
E(
k
i
d
s
)
do
if
T
Y
P
E(
k
i
d
s
[i
]
)
=
N
P
t
h
e
n
su
b
j
e
c
t
=
k
i
d
s
[
i
]
f
o
r
j
=
i
+
1
t
o
S
I
Z
E(
k
i
d
s
)
do
if
T
Y
P
E(
k
i
d
s
[j
]
)
=
V
P
|
P
P
|
S
B
A
R
t
h
e
n
e
x
p
l
o
re
d
←
a
n
e
mp
t
y
se
t
w
h
i
l
e
k
i
d
s
[j
]
n
o
t
i
n
e
x
p
l
o
r
e
d
do
e
x
t
r
a
c
t
i
o
n
←
A
P
P
EN
D
(
su
b
j
e
c
t
,
P
R
ED
I
C
A
T
E
-
V
ER
B
-
P
H
R
A
S
E(
k
i
d
s
[
j
]
))
P
R
I
N
T
(
e
x
t
ra
c
t
i
o
n
)
S
U
B
JEC
T
-
N
O
U
N
_
P
H
R
A
S
E(
k
i
d
s
[
i
]
)
f
u
n
c
t
i
o
n
P
R
ED
I
C
A
T
E
-
V
ER
B
-
P
H
R
A
S
E(
p
a
rs
e
)
r
e
t
u
r
n
s
so
l
u
t
i
o
n
,
f
a
i
l
u
r
e
k
i
d
s
←
C
H
I
L
D
(
p
a
r
se
)
i
n
i
t
i
a
l
i
z
e
p
re
d
i
c
a
t
e
st
r
i
n
g
t
o
b
e
e
mp
t
y
f
o
r
i
=
1
t
o
S
I
Z
E(
k
i
d
s
)
do
if
T
Y
P
E(
k
i
d
s
[i
]
)
=
V
P
|
S
t
h
e
n
if
k
i
d
s
[i
]
n
o
t
i
n
e
x
p
l
o
r
e
d
t
h
e
n
r
e
t
u
r
n
A
P
P
EN
D
(
p
red
i
c
a
t
e
,
P
R
ED
I
C
A
T
E
-
V
ER
B
-
P
H
R
A
S
E(
k
i
d
s
[
i
]
))
e
l
se
i
f
T
Y
P
E(
k
i
d
s
[i
]
)
=
V
B
|
JJ |
R
B
|
M
D
|
T
O
|
A
D
V
P
|
D
T
|
N
N
|
I
N
t
h
e
n
p
re
d
i
c
a
t
e
←
A
P
P
EN
D
(
p
red
i
c
a
t
e
,
k
i
d
s[i
]
);
f
o
r
j
=
i
+
1
t
o
S
I
Z
E(
k
i
d
s
)
do
if
T
Y
P
E(
k
i
d
s
[j
]
)
=
N
P
|
P
P
|
A
D
JP
|
S
|
S
B
A
R
t
h
e
n
r
e
t
u
r
n
A
P
P
EN
D
(
p
red
i
c
a
t
e
,
O
B
JE
C
T
-
N
O
U
N
_
P
H
R
A
S
E(
k
i
d
s
[
j
]
))
a
d
d
p
a
r
se
t
o
e
x
p
l
o
r
e
d
r
e
t
u
r
n
f
a
i
l
u
r
e
Fig
u
r
e
1
.
Su
b
j
ec
t
-
P
r
ed
icate
P
h
r
ase
A
l
g
o
r
ith
m
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
AI
I
SS
N:
2252
-
8938
P
a
r
s
er E
xtra
ctio
n
o
f Tr
ip
les i
n
Un
s
tr
u
ctu
r
ed
Text
(
S
h
a
u
n
D
’
S
o
u
z
a
)
145
f
u
n
c
t
i
o
n
O
B
JEC
T
-
NOUN
-
P
H
R
A
S
E(
p
a
rs
e
)
r
e
t
u
r
n
s
so
l
u
t
i
o
n
,
f
a
i
l
u
r
e
f
o
u
n
d
←
f
a
l
se
k
i
d
s
←
C
H
I
L
D
(
p
a
r
se
)
i
n
i
t
i
a
l
i
z
e
o
b
j
e
c
t
st
r
i
n
g
t
o
b
e
e
mp
t
y
f
o
r
i
=
1
t
o
S
I
Z
E(
k
i
d
s
)
do
if
T
Y
P
E(
k
i
d
s
[i
]
)
=
N
P
|
S
t
h
e
n
f
o
u
n
d
←
t
r
u
e
if
k
i
d
s
[i
]
n
o
t
i
n
e
x
p
l
o
r
e
d
t
h
e
n
r
e
t
u
r
n
A
P
P
EN
D
(
o
b
j
e
c
t
,
O
B
JE
C
T
-
NOUN
-
P
H
R
A
S
E(
k
i
d
s
[i
]
))
e
l
se
r
e
t
u
r
n
A
P
P
EN
D
(
o
b
j
e
c
t
,
G
ET
-
C
O
V
ER
ED
-
T
EX
T
(
k
i
d
s
[
i
]
))
e
l
se
i
f
T
Y
P
E(
k
i
d
s
[i
]
)
=
P
P
t
h
e
n
if
k
i
d
s
[i
]
n
o
t
i
n
e
x
p
l
o
r
e
d
t
h
e
n
r
e
t
u
r
n
A
P
P
EN
D
(
o
b
j
e
c
t
,
O
B
JE
C
T
-
P
R
EPO
S
TI
O
N
-
P
H
R
A
S
E(
k
i
d
s
[
i
]
))
e
l
se
r
e
t
u
r
n
A
P
P
EN
D
(
o
b
j
e
c
t
,
G
ET
-
C
O
V
ER
ED
-
T
EX
T
(
k
i
d
s
[
i
]
))
el
se
i
f
T
Y
P
E(
k
i
d
s
[i
]
)
=
I
N
|
T
O
t
h
e
n
o
b
j
e
c
t
←
A
P
P
EN
D
(
o
b
j
e
c
t
,
k
i
d
s
[
i
]
)
a
d
d
p
a
r
se
t
o
e
x
p
l
o
r
e
d
if
n
o
t
f
o
u
n
d
a
n
d
T
Y
P
E(
p
a
rse
)
=
N
P
t
h
e
n
r
e
t
u
r
n
A
P
P
EN
D
(
o
b
j
e
c
t
,
p
a
rse
)
r
e
t
u
r
n
f
a
i
l
u
r
e
f
u
n
c
t
i
o
n
O
B
JEC
T
-
P
R
EPO
S
I
TI
O
N
-
P
H
R
A
S
E(
p
a
rse
)
r
e
t
u
r
n
s
so
l
u
t
i
o
n
,
f
a
i
l
u
r
e
k
i
d
s
←
C
H
I
L
D
(
p
a
rse
)
i
n
i
t
i
a
l
i
z
e
p
re
p
o
si
t
i
o
n
st
r
i
n
g
t
o
b
e
e
mp
t
y
f
o
r
i
=
1
t
o
S
I
Z
E(
k
i
d
s
)
do
if
T
Y
P
E(
k
i
d
s
[i
]
)
=
N
P
a
n
d
n
o
t
i
n
e
x
p
l
o
re
d
t
h
e
n
r
e
t
u
r
n
A
P
P
EN
D
(
p
rep
o
s
i
t
i
o
n
,
O
B
JE
C
T
-
NOUN
-
P
H
R
A
S
E(
k
i
d
s
[
i
]
))
e
l
se
i
f
T
Y
P
E(
k
i
d
s
[i
]
)
=
P
P
a
n
d
n
o
t
i
n
e
x
p
l
o
re
d
t
h
e
n
r
e
t
u
r
n
A
P
P
EN
D
(
p
rep
o
s
i
t
i
o
n
,
O
B
JE
C
T
-
P
R
EPO
S
TI
O
N
-
P
H
R
A
S
E(
k
i
d
s[i
]
))
e
l
se
i
f
T
Y
P
E(
k
i
d
s
[i
]
)
=
I
N
|
T
O
|
JJ
|
A
D
V
P
t
h
e
n
p
re
p
o
si
t
i
o
n
←
A
P
P
EN
D
(
p
re
p
o
s
i
t
i
o
n
,
k
i
d
s
[
i
]
)
a
d
d
p
a
r
se
t
o
e
x
p
l
o
r
e
d
r
e
t
u
r
n
f
a
i
l
u
r
e
Fig
u
r
e
2
.
Ob
j
ec
t
P
h
r
ase
A
l
g
o
r
ith
m
W
e
im
p
le
m
e
n
t
a
d
ep
th
-
f
ir
s
t
s
e
ar
ch
o
n
th
e
n
-
ar
y
p
ar
s
e
tr
ee
.
W
e
s
ea
r
ch
th
e
p
ar
s
e
tr
ee
f
o
r
a
n
o
u
n
-
v
er
b
p
h
r
ase
in
d
icati
n
g
th
e
s
u
b
j
ec
t
-
p
r
ed
icate
-
Fig
u
r
e
1
.
T
h
e
n
o
u
n
p
h
r
ase
i
s
u
s
ed
as
t
h
e
s
u
b
j
ec
t
in
th
e
cl
au
s
e
.
W
e
lo
o
k
f
o
r
a
v
er
b
p
h
r
ase
VP
o
r
p
r
ep
o
s
itio
n
p
h
r
ase
PP
in
th
e
s
ib
lin
g
s
.
I
n
t
h
e
ca
s
e
o
f
s
u
b
s
e
q
u
en
t
co
n
j
u
n
ctio
n
s
C
C
a
n
d
W
HNP
p
h
r
ases
,
w
e
co
n
tin
u
e
to
s
ea
r
ch
th
e
s
ib
li
n
g
n
o
d
es.
Fo
r
all
f
o
u
n
d
VP
,
P
P
w
e
s
ea
r
ch
f
o
r
t
h
e
p
r
ed
icate
cl
au
s
e
in
th
e
s
e
n
te
n
ce
.
A
p
r
ed
icate
clau
s
e
co
n
s
is
t
s
o
f
a
s
eq
u
e
n
ce
o
f
v
er
b
,
ad
j
ec
tiv
es,
ad
v
er
b
an
d
m
o
d
al
id
en
ti
f
ier
s
.
T
h
ese
ar
e
ap
p
en
d
ed
to
a
s
tr
in
g
o
f
p
r
ed
icate
s
.
VP
p
h
r
as
es
ar
e
s
ea
r
ch
ed
r
ec
u
r
s
iv
el
y
ti
ll
w
e
f
i
n
d
a
ter
m
i
n
al
NP
o
b
j
ec
t
clau
s
e.
W
e
r
ep
r
ese
n
t
t
h
e
SV
O
in
t
h
e
tr
ip
les
f
o
r
m
at.
W
e
u
s
e
a
tr
ai
n
i
n
g
s
et
o
f
2
0
0
p
h
r
ases
f
r
o
m
ea
r
lier
p
u
b
licatio
n
s
o
n
i
n
f
o
r
m
atio
n
ex
tr
ac
tio
n
.
T
h
ese
g
i
v
e
u
s
a
r
an
g
e
o
f
p
ar
s
e
tr
ee
s
to
ev
al
u
ate
th
e
s
ea
r
ch
o
n
an
d
r
ef
in
e.
E
ar
lier
w
o
r
k
o
n
in
f
o
r
m
at
io
n
e
x
tr
ac
tio
n
w
a
s
li
m
ited
to
th
e
ca
p
ab
ilit
ies
o
f
th
e
P
OS
an
d
C
h
u
n
k
er
ta
g
s
.
Ver
b
p
h
r
ases
w
er
e
d
etec
ted
u
s
i
n
g
s
tati
s
tical
p
r
o
b
ab
ilit
ies
o
f
f
r
eq
u
e
n
tl
y
o
cc
u
r
r
in
g
p
atter
n
s
in
t
h
e
E
n
g
li
s
h
lan
g
u
a
g
e.
W
e
i
m
p
le
m
en
t a
r
i
g
o
r
o
u
s
p
ar
s
e
tr
ee
d
esig
n
w
h
ic
h
p
r
eser
v
es th
e
la
n
g
u
a
g
e
s
y
n
tax
o
f
th
e
te
x
t d
ata.
As
th
er
e
is
a
h
i
g
h
a
v
ailab
ilit
y
o
f
co
m
p
u
ti
n
g
to
d
a
y
in
t
h
e
clo
u
d
,
w
e
i
m
p
le
m
en
t
th
e
S
VO
p
ar
s
er
as
an
o
f
f
li
n
e
f
u
n
ct
io
n
to
p
r
o
ce
s
s
t
h
e
s
y
n
tactic
tr
ee
.
W
e
p
ar
s
e
all
th
e
s
e
n
te
n
ce
s
i
n
t
h
e
te
x
t
a
n
d
g
en
er
ate
a
p
ar
s
ed
o
u
tp
u
t.
T
h
is
is
s
u
b
s
eq
u
en
t
l
y
u
s
ed
to
g
e
n
er
ate
th
e
S
VO
tr
i
p
les.
W
ith
th
e
a
v
ailab
ilit
y
o
f
co
m
p
u
ti
n
g
w
e
ca
n
i
m
p
r
o
v
e
p
er
f
o
r
m
a
n
ce
o
f
t
h
e
p
ar
s
er
b
y
p
ar
allelizi
n
g
t
h
e
p
ar
s
i
n
g
o
f
in
p
u
t se
n
ten
ce
s
.
W
e
co
n
tr
ast
t
h
e
SV
O
tr
ip
les
w
ith
p
ast
r
esear
c
h
i
n
clu
d
i
n
g
Op
en
I
E
an
d
C
la
u
s
eI
E
.
W
e
f
in
d
t
h
at
a
p
ar
s
er
b
ased
ap
p
r
o
ac
h
i
s
ab
le
to
ex
tr
ac
t
a
lar
g
e
n
u
m
b
er
o
f
S
VO
’s
ac
cu
r
atel
y
.
Av
ailab
ilit
y
o
f
a
s
y
n
tac
tic
p
ar
s
e
tr
ee
also
en
ab
les
u
s
to
ex
tr
ac
t
tr
ip
les
w
i
th
r
ed
u
ce
d
a
m
b
i
g
u
i
t
y
.
T
h
e
o
b
tain
ed
tr
ip
les
m
ap
ex
a
ctl
y
to
s
u
b
-
tr
ee
s
i
n
th
e
s
e
n
ten
ce
p
ar
s
e
tr
ee
an
d
ca
p
tu
r
e
all
th
e
s
e
m
a
n
tic
i
n
f
o
r
m
atio
n
–
s
u
b
j
ec
t
p
r
e
d
icate
.
T
h
e
n
-
ar
y
p
ar
s
e
tr
ee
en
ca
p
s
u
lates t
h
e
s
y
n
tactic
s
tr
u
ctu
r
e
o
f
th
e
s
en
ten
ce
co
m
p
lete
l
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
IJ
-
AI
Vo
l.
5
,
No
.
4
,
Dec
em
b
e
r
2
0
1
6
:
1
4
3
–
1
4
8
146
W
e
ar
e
ab
le
to
p
r
ec
is
ely
e
x
tr
a
ct
SVO
i
n
f
o
r
m
atio
n
.
I
n
t
h
e
i
n
i
tial r
ev
i
s
io
n
o
f
th
e
co
d
e
w
e
i
m
p
le
m
en
ted
p
r
ed
icate
ex
tr
ac
tio
n
s
to
in
cl
u
d
e
th
e
t
r
ailin
g
n
o
u
n
p
h
r
ase.
T
h
is
w
a
s
u
p
d
ated
to
r
eso
lv
e
th
e
o
b
j
ec
t
clau
s
e
to
co
n
tain
th
e
n
o
u
n
p
h
r
ase
NP
an
d
a
tr
aili
n
g
p
r
ep
o
s
itio
n
p
h
r
ase
P
P
-
Fi
g
u
r
e
2
.
W
e
u
s
e
a
s
et
o
f
h
e
u
r
is
tic
s
to
m
ax
i
m
ize
th
e
n
u
m
b
er
o
f
tr
ip
le
s
g
e
n
er
ated
f
o
r
ea
ch
n
o
u
n
p
h
r
ase,
v
er
b
p
h
r
ase.
3.
RE
SU
L
T
S
A
ND
AN
AL
Y
SI
S
T
h
e
SVO
ex
tr
ac
tio
n
s
ar
e
co
h
er
en
t
as
Op
en
N
L
P
ca
p
tu
r
es
t
h
e
lan
g
u
ag
e
s
y
n
ta
x
in
t
h
e
p
ar
s
e
tr
ee
.
W
e
co
m
p
ar
e
th
e
n
u
m
b
er
o
f
ex
tr
ac
tio
n
s
w
i
t
h
th
e
R
e
Ver
b
ex
tr
ac
to
r
.
W
e
o
b
s
er
v
e
a
lar
g
er
n
u
m
b
er
o
f
tr
ip
les
as
w
e
ar
e
s
ea
r
ch
in
g
f
o
r
all
n
o
u
n
p
h
r
ases
i
n
t
h
e
o
b
j
ec
t.
T
h
e
NL
P
p
ar
s
er
is
ab
le
to
ex
tr
ac
t
a
lar
g
e
n
u
m
b
er
o
f
tr
ip
le
s
m
atc
h
in
g
R
eVe
r
b
an
d
C
la
u
s
I
E
.
E
x
a
m
p
le
s
e
n
te
n
ce
T
h
e
p
r
in
cip
al
o
p
p
o
s
itio
n
p
ar
ties
b
o
y
co
tted
th
e
p
o
lls
a
f
ter
ac
cu
s
atio
n
s
o
f
v
o
te
r
ig
g
in
g
,
an
d
th
e
o
n
l
y
o
th
er
n
a
m
e
o
n
th
e
b
allo
t
w
as a
litt
le
k
n
o
w
n
c
h
alle
n
g
er
f
r
o
m
a
m
ar
g
i
n
al
p
o
liti
ca
l p
ar
t
y
Fig
u
r
e
3
.
An
E
x
a
m
p
le
Se
n
ten
ce
P
ar
s
e
T
r
ee
("
Th
e
p
r
in
cip
a
l o
p
p
o
s
itio
n
p
a
r
ties
"
,
"
b
o
yc
o
tted
"
,
"
th
e
p
o
lls
")
("
Th
e
p
r
in
cip
a
l o
p
p
o
s
itio
n
p
a
r
ties
"
,
"
b
o
yc
o
tted
"
,
"
th
e
p
o
lls
a
fter
a
cc
u
s
a
tio
n
s
")
("
Th
e
p
r
in
cip
a
l o
p
p
o
s
itio
n
p
a
r
ties
"
,
"
b
o
yc
o
tted
"
,
"
th
e
p
o
lls
a
fter
a
cc
u
s
a
tio
n
s
o
f v
o
te
r
ig
g
in
g
")
("
Th
e
o
n
ly
o
t
h
er n
a
me
o
n
th
e
b
a
llo
t
"
,
"
w
a
s
"
,
"
a
litt
le
kn
o
w
n
ch
a
llen
g
er
")
("
Th
e
o
n
ly
o
t
h
er n
a
me
o
n
th
e
b
a
llo
t
"
,
"
w
a
s
"
,
"
a
litt
le
kn
o
w
n
ch
a
llen
g
er fr
o
m
a
ma
r
g
in
a
l p
o
liti
ca
l p
a
r
ty
")
T
h
e
ab
o
v
e
ex
tr
ac
tio
n
s
ar
e
la
b
elled
co
r
r
ec
tly
i
n
t
h
e
R
eVe
r
b
d
ataset
an
d
co
n
tai
n
s
o
m
e
r
ed
u
n
d
a
n
t
ex
tr
ac
tio
n
s
.
W
e
ev
alu
ated
th
e
p
ar
s
er
ex
tr
ac
tio
n
o
n
th
e
C
lu
e
W
eb
1
2
d
ataset
an
d
w
er
e
ab
le
to
ex
tr
ac
t
m
o
r
e
th
a
n
5
0
,
0
0
0
tr
ip
les.
W
e
f
o
u
n
d
th
a
t
th
e
p
ar
s
er
w
as
ab
le
to
p
er
f
o
r
m
o
n
p
ar
w
it
h
R
eVe
r
b
a
n
d
C
lau
s
I
E
.
T
h
is
w
as
ac
h
iev
ed
u
s
in
g
t
h
e
s
y
n
tactic
f
u
n
ct
io
n
alit
y
i
n
t
h
e
p
ar
s
e
tr
ee
-
Fi
g
u
r
e
3
.
I
t
d
e
m
o
n
s
tr
ates
t
h
e
ab
ilit
y
o
f
a
p
ar
s
er
b
ased
ap
p
r
o
ac
h
in
ex
tr
ac
ti
n
g
h
ig
h
q
u
alit
y
tr
ip
les.
W
e
v
er
if
ied
th
e
ex
tr
ac
t
io
n
s
f
o
r
a
s
a
m
p
le
s
et
o
f
s
en
te
n
ce
s
i
n
t
h
e
Op
e
n
I
E
an
d
C
la
u
s
I
E
p
u
b
licatio
n
s
.
T
h
ese
w
er
e
u
s
ed
to
en
s
u
r
e
p
r
e
cisi
o
n
i
n
th
e
p
ar
s
er
e
x
tr
ac
tio
n
s
.
W
e
ad
d
itio
n
all
y
r
an
th
e
p
ar
s
er
o
n
t
h
e
C
l
u
eW
eb
d
ata
an
d
co
m
p
ar
ed
th
e
n
u
m
b
e
r
o
f
ex
tr
ac
tio
n
s
w
it
h
t
h
e
alter
n
ati
v
e
ap
p
r
o
ac
h
es.
W
e
m
ea
s
u
r
ed
th
e
d
is
tr
ib
u
tio
n
o
f
t
h
e
n
o
u
n
an
d
v
er
b
s
u
b
-
tr
e
es
i
n
t
h
e
s
e
n
te
n
ce
te
x
t
-
T
ab
le
1
.
W
e
f
o
u
n
d
t
h
at
1
0
%
o
f
th
e
p
h
r
ases
w
er
e
p
r
ep
o
s
itio
n
al.
T
h
e
d
en
s
it
y
o
f
th
e
n
o
u
n
an
d
v
er
b
p
h
r
ases
ar
e
in
ag
r
ee
m
e
n
t
w
it
h
th
e
E
n
g
l
is
h
co
n
te
x
t
f
r
ee
g
r
a
m
m
ar
(
C
F
G)
.
T
ab
le
1
.
P
h
r
ase
D
is
tr
ib
u
tio
n
N
o
u
n
F
r
e
q
u
e
n
c
y
NP
→
NN
1
4
%
NP
→
N
P
P
P
1
2
%
NP
→
D
T
N
N
1
2
%
NP
→
N
N
N
N
6%
V
e
r
b
F
r
e
q
u
e
n
c
y
VP
→
V
B
N
P
1
6
%
VP
→
V
B
V
P
1
0
%
VP
→
T
O
V
P
9%
VP
→
V
B
P
P
8%
VP
→
VB
6%
P
r
e
p
o
si
t
i
o
n
F
r
e
q
u
e
n
c
y
PP
→
I
N
N
P
8
1
%
PP
→
T
O
N
P
9%
E
ar
lier
w
o
r
k
s
li
k
e
Op
e
n
I
E
an
d
R
eVe
r
b
h
a
v
e
lo
o
k
ed
at
t
h
e
ex
tr
ac
tio
n
o
f
s
u
b
j
ec
t
-
v
er
b
-
o
b
ject
(
SVO)
tr
ip
les.
T
h
ey
w
er
e
h
o
w
e
v
er
b
ased
p
r
i
m
ar
il
y
o
n
th
e
a
v
ailab
il
i
t
y
o
f
P
OS a
n
d
ch
u
n
k
er
d
ata.
S
tr
u
ctu
r
e
o
f
th
e
v
er
b
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
AI
I
SS
N:
2252
-
8938
P
a
r
s
er E
xtra
ctio
n
o
f Tr
ip
les i
n
Un
s
tr
u
ctu
r
ed
Text
(
S
h
a
u
n
D
’
S
o
u
z
a
)
147
an
d
n
o
u
n
p
h
r
ase
s
w
er
e
d
eter
m
i
n
ed
u
s
in
g
s
ta
tis
tica
l
d
is
tr
ib
u
tio
n
o
f
t
h
e
p
h
r
ases
in
te
x
t
d
ata.
C
lau
s
I
E
u
s
ed
a
d
ep
en
d
en
c
y
p
ar
s
er
in
r
eso
l
v
in
g
th
e
SVO
r
elatio
n
s
.
P
r
o
j
ec
ts
lik
e
DB
p
ed
ia
[
5
]
w
er
e
d
esig
n
ed
to
e
x
tr
ac
t
s
tr
u
ct
u
r
e
d
d
ata
in
th
e
in
f
o
r
m
a
tio
n
b
o
x
an
d
m
ap
it
to
an
o
n
to
lo
g
y
.
T
g
r
ep
2
[
6
]
en
ab
le
u
s
to
ex
tr
ac
t
a
n
d
p
ar
s
e
a
tr
ee
w
it
h
o
u
t
ex
p
licit
l
y
co
d
in
g
th
e
r
u
le
s
.
A
s
et
o
f
r
eg
u
lar
ex
p
r
es
s
io
n
s
ar
e
u
s
ed
t
o
ex
tr
ac
t
m
atc
h
in
g
s
u
b
-
tr
ee
s
.
Fig
u
r
e
4
.
Nu
m
b
er
o
f
C
o
r
r
ec
t N
o
n
-
r
ed
u
n
d
an
t E
x
tr
ac
tio
n
s
W
e
ev
alu
ated
a
n
u
m
b
er
o
f
ex
t
r
a
ctio
n
s
o
n
th
e
R
eVe
r
b
,
W
ik
ip
ed
ia
an
d
NY
T
d
ataset.
W
e
o
b
tain
ed
th
e
s
a
m
p
le
d
atase
t
f
r
o
m
t
h
e
C
lau
s
I
E
s
o
u
r
ce
s
.
W
e
w
er
e
ab
le
to
ex
tr
ac
t
m
o
r
e
th
a
n
2
0
0
0
SVO
in
t
h
e
d
atase
t
w
it
h
1
0
0
0
m
atch
i
n
g
th
e
C
la
u
s
I
E
ex
tr
ac
tio
n
s
.
As
all
th
e
ex
tr
ac
ted
r
esu
l
ts
ar
e
s
e
m
an
t
icall
y
ac
cu
r
ate,
th
e
p
r
ec
is
io
n
o
f
t
h
e
r
es
u
lt
s
i
s
~0
.
9
.
T
h
is
v
alu
e
is
i
n
d
ep
en
d
en
t o
f
t
h
e
d
atase
t a
n
d
is
d
er
i
v
ed
f
r
o
m
t
h
e
e
x
tr
ac
ti
o
n
g
r
a
m
m
ar
r
u
le
s
.
T
h
e
ex
tr
ac
t
io
n
s
ar
e
b
ased
o
n
a
r
u
le
b
ased
s
y
s
te
m
a
n
d
ca
p
tu
r
e
th
e
s
y
n
tax
o
f
t
h
e
E
n
g
lis
h
lan
g
u
ag
e.
So
m
e
o
f
th
e
SVO
o
u
tp
u
ts
ar
e
in
co
r
r
ec
t
d
u
e
to
th
e
a
m
b
ig
u
it
ies
i
n
t
h
e
la
n
g
u
a
g
e
p
ar
s
e
tr
ee
i
n
cl
u
d
in
g
co
n
j
u
n
ctio
n
s
i
n
n
o
u
n
p
h
r
a
s
e
s
.
W
e
v
er
i
f
ied
t
h
e
ex
tr
ac
ted
tr
ip
les
to
m
ea
s
u
r
e
th
e
r
ec
all
o
f
th
e
d
ata.
T
h
e
r
ec
all
v
alu
e
is
a
f
u
n
c
tio
n
o
f
th
e
g
r
a
m
m
ar
.
W
e
ca
n
r
ef
in
e
t
h
e
r
u
les
to
f
i
n
d
ad
d
itio
n
al
tr
ip
les
in
th
e
d
ata.
T
h
is
w
o
u
ld
in
cr
ea
s
e
th
e
r
ec
all
o
n
t
h
e
ex
tr
ac
ted
v
al
u
es.
W
e
m
ea
s
u
r
ed
an
av
er
a
g
e
r
ec
al
l v
al
u
e
o
f
6
0
% o
n
t
h
e
tr
ip
le
s
-
T
ab
le
2
.
W
e
u
s
ed
th
e
e
x
tr
ac
tio
n
s
-
all
-
lab
eled
as
a
b
aselin
e
f
o
r
o
u
r
co
m
p
u
tatio
n
.
T
h
ese
in
clu
d
e
all
t
h
e
ex
tr
ac
tio
n
s
f
r
o
m
R
e
Ver
b
,
C
lau
s
I
E
a
n
d
o
th
er
OI
E
u
tili
tie
s
.
W
e
esti
m
a
ted
a
p
r
ec
is
io
n
o
f
0
.
8
f
o
r
th
e
p
ar
s
er
ex
tr
ac
tio
n
s
.
W
e
f
o
u
n
d
t
h
at
t
h
e
p
ar
s
er
was
ab
le
to
ex
tr
ac
t
2
-
2
.
5
ti
m
es
th
e
co
r
r
ec
t
ex
tr
ac
tio
n
s
o
f
R
e
Ver
b
an
d
8
0
%
o
f
th
e
co
r
r
ec
t
n
o
n
-
r
ed
u
n
d
an
t
C
lau
s
I
E
ex
tr
ac
tio
n
s
-
E
rr
o
r!
Ref
er
enc
e
s
o
urce
no
t
f
o
un
d.
.
T
ab
le
2
.
P
r
ec
is
io
n
an
d
R
ec
all
Valu
es
f
o
r
Var
io
u
s
Data
s
ets
P
r
e
c
i
si
o
n
R
e
c
a
l
l
N
Y
T
0
.
8
0
.
6
4
W
i
k
i
p
e
d
i
a
0
.
8
0
.
7
1
R
e
V
e
r
b
0
.
8
0
.
5
3
4.
CO
NCLU
SI
O
N
W
e
p
r
esen
ted
a
m
et
h
o
d
o
lo
g
y
f
o
r
ex
tr
ac
tio
n
o
f
s
u
b
j
ec
t
-
p
r
ed
icate
-
o
b
j
ec
t
tr
ip
les
in
a
te
x
t
co
r
p
u
s
.
W
e
p
lan
to
ex
te
n
d
t
h
is
w
o
r
k
to
a
lar
g
er
o
n
to
lo
g
ical
en
g
i
n
ee
r
in
g
f
o
r
k
n
o
w
led
g
e
in
f
er
e
n
ce
.
W
e
f
o
u
n
d
th
at
a
s
y
n
tactic
p
ar
s
er
w
a
s
ab
le
to
ac
cu
r
atel
y
e
x
tr
ac
t
tr
ip
les
in
a
t
ex
t.
W
e
ex
p
lo
r
ed
o
p
p
o
r
tu
n
itie
s
to
f
u
r
t
h
er
ex
te
n
d
th
is
w
o
r
k
i
n
tr
an
s
lati
n
g
an
u
n
s
tr
u
ctu
r
ed
co
r
p
u
s
o
f
d
ata
in
to
a
s
e
m
a
n
tic
o
n
to
lo
g
y
.
A
u
s
er
is
ab
le
to
ex
p
lo
r
e
th
e
tex
t u
s
i
n
g
a
tr
ip
les
s
tr
u
ct
u
r
e.
P
r
o
v
id
e
a
s
tate
m
e
n
t
t
h
at
w
h
a
t
is
e
x
p
ec
ted
,
as
s
tated
i
n
t
h
e
"
I
n
tr
o
d
u
ctio
n
"
ch
ap
ter
ca
n
u
lti
m
atel
y
r
esu
lt
i
n
"
R
es
u
lts
a
n
d
Dis
c
u
s
s
io
n
"
ch
ap
ter
,
s
o
th
er
e
is
co
m
p
a
tib
ilit
y
.
Mo
r
eo
v
er
,
it
ca
n
also
b
e
ad
d
ed
th
e
p
r
o
s
p
ec
t
o
f
th
e
d
ev
elo
p
m
e
n
t
o
f
r
esear
ch
r
esu
l
ts
an
d
ap
p
licatio
n
p
r
o
s
p
ec
ts
o
f
f
u
r
t
h
er
s
t
u
d
ies
in
to
th
e
n
e
x
t
(
b
ased
o
n
r
esu
lt a
n
d
d
is
c
u
s
s
io
n
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
IJ
-
AI
Vo
l.
5
,
No
.
4
,
Dec
em
b
e
r
2
0
1
6
:
1
4
3
–
1
4
8
148
RE
F
E
R
E
NC
E
S
[1
]
Et
z
io
n
i
O,
F
a
d
e
r
A
,
Ch
r
isten
se
n
J,
S
o
d
e
rlan
d
S
,
M
a
u
sa
m
M.
Op
e
n
I
n
fo
rm
a
t
io
n
E
x
tra
c
ti
o
n
:
th
e
S
e
c
o
n
d
G
e
n
e
ra
ti
o
n
.
P
r
o
c
e
e
d
in
g
s o
f
th
e
T
w
e
n
t
y
-
S
e
c
o
n
d
In
ter
n
a
ti
o
n
a
l
Jo
i
n
t
C
o
n
f
e
re
n
c
e
o
n
A
rti
f
icia
l
In
telli
g
e
n
c
e
,
2
0
1
1
.
[2
]
Co
rro
L
D,
G
e
m
u
ll
a
R
,
Cl
a
u
sIE:
C
la
u
se
-
B
a
se
d
O
p
e
n
I
n
f
o
rm
a
ti
o
n
E
x
tra
c
ti
o
n
.
P
r
o
c
e
e
d
in
g
s
o
f
th
e
2
2
n
d
I
n
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
W
o
rld
W
id
e
W
e
b
,
2
0
1
3
.
[3
]
Op
e
n
NL
P
,
se
e
h
t
tp
s://
o
p
e
n
n
l
p
.
a
p
a
c
h
e
.
o
rg
.
[4
]
Ho
p
c
ro
f
t
J,
Ullm
a
n
J
,
In
tr
o
d
u
c
ti
o
n
to
a
u
t
o
m
a
ta t
h
e
o
r
y
,
lan
g
u
a
g
e
s,
a
n
d
c
o
m
p
u
tatio
n
.
A
d
d
is
o
n
-
W
e
sle
y
,
1
9
7
9
.
[5
]
A
u
e
r
S
,
Bize
r
C
,
Ko
b
il
a
ro
v
G
,
Leh
m
a
n
n
J,
C
y
g
a
n
iak
R,
Iv
e
s
Z
,
D
Bp
e
d
ia:
A
n
u
c
leu
s
f
o
r
a
w
e
b
o
f
o
p
e
n
d
a
ta.
In
T
h
e
S
e
m
a
n
ti
c
W
e
b
.
2
0
0
7
;
4
8
2
5
:
7
2
2
-
7
3
5
,
S
p
rin
g
e
r
.
[6
]
T
g
r
e
p
2
,
se
e
h
tt
p
:/
/t
e
d
lab
.
m
it
.
e
d
u
/
~
d
r/T
g
re
p
2
.
B
I
O
G
RAP
H
Y
O
F
AU
T
H
O
R
S
h
a
u
n
D’So
u
z
a
o
b
tain
e
d
a
M
.
S
.
E.
d
e
g
re
e
in
El
e
c
tri
c
a
l
En
g
in
e
e
rin
g
f
ro
m
th
e
Un
iv
e
rsit
y
o
f
M
ich
ig
a
n
,
A
n
n
A
rb
o
r
a
n
d
a
B.
S
.
d
e
g
re
e
in
Co
m
p
u
ter
S
c
ien
c
e
,
El
e
c
tri
c
a
l
a
n
d
Co
m
p
u
ter
En
g
in
e
e
rin
g
f
ro
m
Co
rn
e
ll
Un
iv
e
rsity
.
He
is
c
u
rre
n
tl
y
w
o
rk
in
g
a
s
a
T
e
c
h
n
ica
l
L
e
a
d
in
th
e
CT
O
Off
ic
e
a
t
W
ip
ro
.
His
re
se
a
rc
h
i
n
tere
sts
in
c
lu
d
e
m
a
c
h
in
e
lea
rn
in
g
,
c
o
m
p
il
e
rs,
a
lg
o
rit
h
m
s
a
n
d
s
y
ste
m
s.
Evaluation Warning : The document was created with Spire.PDF for Python.