-DOCSTART- NN O O

This NN O O
paper NN O O
presents NN O O
an NN O O
algorithm NN O I-Generic
for NN O O
computing NN O I-Task
optical NN O I-Task
flow NN O I-Task
, NN O I-Task
shape NN O I-Task
, NN O I-Task
motion NN O I-Task
, NN O I-Task
lighting NN O I-Task
, NN O I-Task
and NN O I-Task
albedo NN O I-Task
from NN O O
an NN O O
image NN O I-Material
sequence NN O I-Material
of NN O O
a NN O O
rigidly-moving NN O I-Material
Lambertian NN O I-Material
object NN O I-Material
under NN O O
distant NN O I-OtherScientificTerm
illumination NN O I-OtherScientificTerm
. NN O O

The NN O O
problem NN O I-Generic
is NN O O
formulated NN O O
in NN O O
a NN O O
manner NN O O
that NN O O
subsumes NN O O
structure NN O O
from NN O O
motion NN O I-Material
, NN O O
multi-view NN O I-Material
stereo NN O I-Material
, NN O O
and NN O O
photo-metric NN O I-Material
stereo NN O I-Material
as NN O O
special NN O O
cases NN O O
. NN O O

The NN O O
algorithm NN O I-Generic
utilizes NN O O
both NN O O
spatial NN O I-OtherScientificTerm
and NN O I-OtherScientificTerm
temporal NN O I-OtherScientificTerm
intensity NN O I-OtherScientificTerm
variation NN O I-OtherScientificTerm
as NN O O
cues NN O I-Generic
: NN O O
the NN O O
former NN O I-Generic
constrains NN O O
flow NN O I-OtherScientificTerm
and NN O O
the NN O O
latter NN O I-Generic
constrains NN O O
surface NN O I-OtherScientificTerm
orientation NN O I-OtherScientificTerm
; NN O O
combining NN O O
both NN O O
cues NN O I-Generic
enables NN O O
dense NN O I-Task
reconstruction NN O I-Task
of NN O I-Task
both NN O I-Task
textured NN O I-Task
and NN O I-Task
texture-less NN O I-Task
surfaces NN O I-Task
. NN O O

The NN O O
algorithm NN O I-Generic
works NN O O
by NN O O
iteratively NN O O
estimating NN O I-Method
affine NN O I-Method
camera NN O I-Method
parameters NN O I-Method
, NN O I-Method
illumination NN O I-Method
, NN O I-Method
shape NN O I-Method
, NN O I-Method
and NN O I-Method
albedo NN O I-Method
in NN O O
an NN O O
alternating NN O O
fashion NN O O
. NN O O

Results NN O O
are NN O O
demonstrated NN O O
on NN O O
videos NN O I-Material
of NN O I-Material
hand-held NN O I-Material
objects NN O I-Material
moving NN O O
in NN O O
front NN O O
of NN O O
a NN O O
fixed NN O O
light NN O O
and NN O O
camera NN O O
. NN O O

-DOCSTART- NN O O

Past NN O O
work NN O O
of NN O O
generating NN O O
referring NN O I-OtherScientificTerm
expressions NN O I-OtherScientificTerm
mainly NN O O
utilized NN O O
attributes NN O O
of NN O O
objects NN O O
and NN O O
binary NN O I-OtherScientificTerm
relations NN O I-OtherScientificTerm
between NN O O
objects NN O O
. NN O O

However NN O O
, NN O O
such NN O O
an NN O O
approach NN O O
does NN O O
not NN O O
work NN O O
well NN O O
when NN O O
there NN O O
is NN O O
no NN O O
distinctive NN O O
attribute NN O O
among NN O O
objects NN O O
. NN O O

To NN O O
overcome NN O O
this NN O O
limitation NN O O
, NN O O
this NN O O
paper NN O O
proposes NN O O
a NN O O
method NN O O
utilizing NN O O
the NN O O
perceptual NN O O
groups NN O O
of NN O O
objects NN O O
and NN O O
n-ary NN O I-OtherScientificTerm
relations NN O I-OtherScientificTerm
among NN O O
them NN O O
. NN O O

The NN O O
key NN O O
is NN O O
to NN O O
identify NN O O
groups NN O O
of NN O O
objects NN O O
that NN O O
are NN O O
naturally NN O O
recognized NN O O
by NN O O
humans NN O O
. NN O O

We NN O O
conducted NN O O
psychological NN O O
experiments NN O O
with NN O O
42 NN O O
subjects NN O O
to NN O O
collect NN O O
referring NN O I-OtherScientificTerm
expressions NN O I-OtherScientificTerm
in NN O O
such NN O O
situations NN O O
, NN O O
and NN O O
built NN O O
a NN O O
generation NN O I-Method
algorithm NN O I-Method
based NN O O
on NN O O
the NN O O
results NN O O
. NN O O

The NN O O
evaluation NN O O
using NN O O
another NN O O
23 NN O O
subjects NN O O
showed NN O O
that NN O O
the NN O O
proposed NN O O
method NN O I-Generic
could NN O O
effectively NN O O
generate NN O O
proper NN O O
referring NN O I-OtherScientificTerm
expressions NN O I-OtherScientificTerm
. NN O O

-DOCSTART- NN O O

An NN O O
entity-oriented NN O I-Method
approach NN O I-Method
to NN O O
restricted-domain NN O I-Task
parsing NN O I-Task
is NN O O
proposed NN O O
. NN O O

In NN O O
this NN O O
approach NN O I-Generic
, NN O O
the NN O O
definitions NN O O
of NN O O
the NN O O
structure NN O I-OtherScientificTerm
and NN O I-OtherScientificTerm
surface NN O I-OtherScientificTerm
representation NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
domain NN O I-OtherScientificTerm
entities NN O I-OtherScientificTerm
are NN O O
grouped NN O O
together NN O O
. NN O O

Like NN O O
semantic NN O I-Method
grammar NN O I-Method
, NN O O
this NN O I-Generic
allows NN O O
easy NN O O
exploitation NN O O
of NN O O
limited NN O I-OtherScientificTerm
domain NN O I-OtherScientificTerm
semantics NN O I-OtherScientificTerm
. NN O O

In NN O O
addition NN O O
, NN O O
it NN O I-Generic
facilitates NN O O
fragmentary NN O I-Task
recognition NN O I-Task
and NN O O
the NN O O
use NN O O
of NN O O
multiple NN O I-Method
parsing NN O I-Method
strategies NN O I-Method
, NN O O
and NN O O
so NN O O
is NN O O
particularly NN O O
useful NN O O
for NN O O
robust NN O O
recognition NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
extra-grammatical NN O I-OtherScientificTerm
input NN O I-OtherScientificTerm
. NN O O

Several NN O O
advantages NN O O
from NN O O
the NN O O
point NN O O
of NN O O
view NN O O
of NN O O
language NN O O
definition NN O O
are NN O O
also NN O O
noted NN O O
. NN O O

Representative NN O O
samples NN O O
from NN O O
an NN O O
entity-oriented NN O I-OtherScientificTerm
language NN O I-OtherScientificTerm
definition NN O I-OtherScientificTerm
are NN O O
presented NN O O
, NN O O
along NN O O
with NN O O
a NN O O
control NN O I-OtherScientificTerm
structure NN O I-OtherScientificTerm
for NN O O
an NN O O
entity-oriented NN O I-Method
parser NN O I-Method
, NN O O
some NN O O
parsing NN O I-Method
strategies NN O I-Method
that NN O O
use NN O O
the NN O O
control NN O I-OtherScientificTerm
structure NN O I-OtherScientificTerm
, NN O O
and NN O O
worked NN O O
examples NN O O
of NN O O
parses NN O O
. NN O O

A NN O O
parser NN O I-Method
incorporating NN O O
the NN O O
control NN O I-OtherScientificTerm
structure NN O I-OtherScientificTerm
and NN O O
the NN O O
parsing NN O I-Method
strategies NN O I-Method
is NN O O
currently NN O O
under NN O O
implementation NN O O
. NN O O

-DOCSTART- NN O O

This NN O O
paper NN O O
summarizes NN O O
the NN O O
formalism NN O I-Task
of NN O I-Task
Category NN O I-OtherScientificTerm
Cooccurrence NN O I-OtherScientificTerm
Restrictions NN O I-OtherScientificTerm
-LRB- NN O I-OtherScientificTerm
CCRs NN O I-OtherScientificTerm
-RRB- NN O I-OtherScientificTerm
and NN O O
describes NN O O
two NN O O
parsing NN O I-Method
algorithms NN O I-Method
that NN O O
interpret NN O O
it NN O I-Generic
. NN O O

CCRs NN O I-OtherScientificTerm
are NN O O
Boolean NN O I-OtherScientificTerm
conditions NN O I-OtherScientificTerm
on NN O O
the NN O O
cooccurrence NN O O
of NN O O
categories NN O O
in NN O O
local NN O I-OtherScientificTerm
trees NN O I-OtherScientificTerm
which NN O O
allow NN O O
the NN O O
statement NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
generalizations NN O I-OtherScientificTerm
which NN O O
can NN O O
not NN O O
be NN O O
captured NN O O
in NN O O
other NN O O
current NN O O
syntax NN O I-Method
formalisms NN O I-Method
. NN O O

The NN O O
use NN O O
of NN O O
CCRs NN O I-OtherScientificTerm
leads NN O O
to NN O O
syntactic NN O I-OtherScientificTerm
descriptions NN O I-OtherScientificTerm
formulated NN O O
entirely NN O O
with NN O O
restrictive NN O I-OtherScientificTerm
statements NN O I-OtherScientificTerm
. NN O O

The NN O O
paper NN O O
shows NN O O
how NN O O
conventional NN O O
algorithms NN O I-Generic
for NN O O
the NN O O
analysis NN O O
of NN O O
context NN O I-Material
free NN O I-Material
languages NN O I-Material
can NN O O
be NN O O
adapted NN O O
to NN O O
the NN O O
CCR NN O I-Task
formalism NN O I-Task
. NN O O

Special NN O O
attention NN O O
is NN O O
given NN O O
to NN O O
the NN O O
part NN O O
of NN O O
the NN O O
parser NN O I-Method
that NN O O
checks NN O O
the NN O O
fulfillment NN O O
of NN O O
logical NN O I-OtherScientificTerm
well-formedness NN O I-OtherScientificTerm
conditions NN O I-OtherScientificTerm
on NN O O
trees NN O I-OtherScientificTerm
. NN O O

-DOCSTART- NN O O

We NN O O
present NN O O
a NN O O
text NN O I-Method
mining NN O I-Method
method NN O I-Method
for NN O O
finding NN O O
synonymous NN O I-OtherScientificTerm
expressions NN O I-OtherScientificTerm
based NN O O
on NN O O
the NN O O
distributional NN O I-OtherScientificTerm
hypothesis NN O I-OtherScientificTerm
in NN O O
a NN O O
set NN O O
of NN O O
coherent NN O O
corpora NN O O
. NN O O

This NN O O
paper NN O O
proposes NN O O
a NN O O
new NN O O
methodology NN O I-Generic
to NN O O
improve NN O O
the NN O O
accuracy NN O I-Metric
of NN O O
a NN O O
term NN O I-Method
aggregation NN O I-Method
system NN O I-Method
using NN O O
each NN O O
author NN O O
's NN O O
text NN O O
as NN O O
a NN O O
coherent NN O O
corpus NN O O
. NN O O

Our NN O O
approach NN O I-Generic
is NN O O
based NN O O
on NN O O
the NN O O
idea NN O O
that NN O O
one NN O O
person NN O O
tends NN O O
to NN O O
use NN O O
one NN O O
expression NN O O
for NN O O
one NN O O
meaning NN O O
. NN O O

According NN O O
to NN O O
our NN O O
assumption NN O O
, NN O O
most NN O O
of NN O O
the NN O O
words NN O O
with NN O O
similar NN O I-OtherScientificTerm
context NN O I-OtherScientificTerm
features NN O I-OtherScientificTerm
in NN O O
each NN O O
author NN O O
's NN O O
corpus NN O O
tend NN O O
not NN O O
to NN O O
be NN O O
synonymous NN O I-OtherScientificTerm
expressions NN O I-OtherScientificTerm
. NN O O

Our NN O O
proposed NN O O
method NN O I-Generic
improves NN O O
the NN O O
accuracy NN O I-Metric
of NN O O
our NN O O
term NN O I-Method
aggregation NN O I-Method
system NN O I-Method
, NN O O
showing NN O O
that NN O O
our NN O O
approach NN O I-Generic
is NN O O
successful NN O O
. NN O O

-DOCSTART- NN O O

In NN O O
this NN O O
work NN O O
, NN O O
we NN O O
present NN O O
a NN O O
technique NN O I-Generic
for NN O O
robust NN O I-Task
estimation NN O I-Task
, NN O O
which NN O O
by NN O O
explicitly NN O O
incorporating NN O O
the NN O O
inherent NN O I-OtherScientificTerm
uncertainty NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
the NN O I-OtherScientificTerm
estimation NN O I-OtherScientificTerm
procedure NN O I-OtherScientificTerm
, NN O O
results NN O O
in NN O O
a NN O O
more NN O O
efficient NN O I-Method
robust NN O I-Method
estimation NN O I-Method
algorithm NN O I-Method
. NN O O

In NN O O
addition NN O O
, NN O O
we NN O O
build NN O O
on NN O O
recent NN O O
work NN O O
in NN O O
randomized NN O I-Task
model NN O I-Task
verification NN O I-Task
, NN O O
and NN O O
use NN O O
this NN O I-Generic
to NN O O
characterize NN O O
the NN O O
` NN O O
non-randomness NN O O
' NN O O
of NN O O
a NN O O
solution NN O O
. NN O O

The NN O O
combination NN O O
of NN O O
these NN O O
two NN O O
strategies NN O I-Generic
results NN O O
in NN O O
a NN O O
robust NN O I-Method
estimation NN O I-Method
procedure NN O I-Method
that NN O O
provides NN O O
a NN O O
significant NN O O
speed-up NN O O
over NN O O
existing NN O O
RANSAC NN O I-Method
techniques NN O I-Method
, NN O O
while NN O O
requiring NN O O
no NN O O
prior NN O I-OtherScientificTerm
information NN O I-OtherScientificTerm
to NN O O
guide NN O O
the NN O O
sampling NN O I-OtherScientificTerm
process NN O I-OtherScientificTerm
. NN O O

In NN O O
particular NN O O
, NN O O
our NN O O
algorithm NN O I-Generic
requires NN O O
, NN O O
on NN O O
average NN O O
, NN O O
3-10 NN O O
times NN O O
fewer NN O O
samples NN O O
than NN O O
standard NN O O
RANSAC NN O I-Method
, NN O O
which NN O O
is NN O O
in NN O O
close NN O O
agreement NN O O
with NN O O
theoretical NN O I-OtherScientificTerm
predictions NN O I-OtherScientificTerm
. NN O O

The NN O O
efficiency NN O O
of NN O O
the NN O O
algorithm NN O I-Generic
is NN O O
demonstrated NN O O
on NN O O
a NN O O
selection NN O O
of NN O O
geometric NN O I-Task
estimation NN O I-Task
problems NN O I-Task
. NN O O

-DOCSTART- NN O O

An NN O O
attempt NN O O
has NN O O
been NN O O
made NN O O
to NN O O
use NN O O
an NN O O
Augmented NN O I-Method
Transition NN O I-Method
Network NN O I-Method
as NN O O
a NN O O
procedural NN O O
dialog NN O I-Method
model NN O I-Method
. NN O O

The NN O O
development NN O O
of NN O O
such NN O O
a NN O O
model NN O I-Generic
appears NN O O
to NN O O
be NN O O
important NN O O
in NN O O
several NN O O
respects NN O O
: NN O O
as NN O O
a NN O O
device NN O I-Generic
to NN O O
represent NN O O
and NN O O
to NN O O
use NN O O
different NN O O
dialog NN O I-OtherScientificTerm
schemata NN O I-OtherScientificTerm
proposed NN O O
in NN O O
empirical NN O O
conversation NN O I-Method
analysis NN O I-Method
; NN O O
as NN O O
a NN O O
device NN O I-Generic
to NN O O
represent NN O O
and NN O O
to NN O O
use NN O O
models NN O I-Generic
of NN O O
verbal NN O I-OtherScientificTerm
interaction NN O I-OtherScientificTerm
; NN O O
as NN O O
a NN O O
device NN O I-Generic
combining NN O O
knowledge NN O O
about NN O O
dialog NN O I-OtherScientificTerm
schemata NN O I-OtherScientificTerm
and NN O O
about NN O O
verbal NN O I-OtherScientificTerm
interaction NN O I-OtherScientificTerm
with NN O O
knowledge NN O O
about NN O O
task-oriented NN O I-Material
and NN O I-Material
goal-directed NN O I-Material
dialogs NN O I-Material
. NN O O

A NN O O
standard NN O O
ATN NN O I-Method
should NN O O
be NN O O
further NN O O
developed NN O O
in NN O O
order NN O O
to NN O O
account NN O O
for NN O O
the NN O O
verbal NN O I-OtherScientificTerm
interactions NN O I-OtherScientificTerm
of NN O O
task-oriented NN O I-Material
dialogs NN O I-Material
. NN O O

-DOCSTART- NN O O

We NN O O
present NN O O
a NN O O
practically NN O O
unsupervised NN O I-Method
learning NN O I-Method
method NN O I-Method
to NN O O
produce NN O O
single-snippet NN O I-OtherScientificTerm
answers NN O I-OtherScientificTerm
to NN O O
definition NN O O
questions NN O O
in NN O O
question NN O I-Method
answering NN O I-Method
systems NN O I-Method
that NN O O
supplement NN O O
Web NN O I-Method
search NN O I-Method
engines NN O I-Method
. NN O O

The NN O O
method NN O I-Generic
exploits NN O O
on-line NN O I-Material
encyclopedias NN O I-Material
and NN O I-Material
dictionaries NN O I-Material
to NN O O
generate NN O O
automatically NN O O
an NN O O
arbitrarily NN O O
large NN O O
number NN O O
of NN O O
positive NN O I-Material
and NN O I-Material
negative NN O I-Material
definition NN O I-Material
examples NN O I-Material
, NN O O
which NN O O
are NN O O
then NN O O
used NN O O
to NN O O
train NN O O
an NN O O
svm NN O I-Method
to NN O O
separate NN O O
the NN O O
two NN O O
classes NN O O
. NN O O

We NN O O
show NN O O
experimentally NN O O
that NN O O
the NN O O
proposed NN O O
method NN O I-Generic
is NN O O
viable NN O O
, NN O O
that NN O O
it NN O I-Generic
outperforms NN O O
the NN O O
alternative NN O I-Generic
of NN O O
training NN O O
the NN O O
system NN O I-Generic
on NN O O
questions NN O O
and NN O O
news NN O I-Material
articles NN O I-Material
from NN O O
trec NN O I-Material
, NN O O
and NN O O
that NN O O
it NN O I-Generic
helps NN O O
the NN O O
search NN O I-Method
engine NN O I-Method
handle NN O O
definition NN O O
questions NN O O
significantly NN O O
better NN O O
. NN O O

-DOCSTART- NN O O

We NN O O
revisit NN O O
the NN O O
classical NN O I-Task
decision-theoretic NN O I-Task
problem NN O I-Task
of NN O I-Task
weighted NN O I-Task
expert NN O I-Task
voting NN O I-Task
from NN O O
a NN O O
statistical NN O I-Method
learning NN O I-Method
perspective NN O I-Method
. NN O O

In NN O O
particular NN O O
, NN O O
we NN O O
examine NN O O
the NN O O
consistency NN O O
-LRB- NN O O
both NN O O
asymptotic NN O O
and NN O O
finitary NN O O
-RRB- NN O O
of NN O O
the NN O O
optimal NN O O
Nitzan-Paroush NN O I-OtherScientificTerm
weighted NN O I-OtherScientificTerm
majority NN O I-OtherScientificTerm
and NN O O
related NN O O
rules NN O O
. NN O O

In NN O O
the NN O O
case NN O O
of NN O O
known NN O O
expert NN O I-OtherScientificTerm
competence NN O I-OtherScientificTerm
levels NN O I-OtherScientificTerm
, NN O O
we NN O O
give NN O O
sharp NN O I-Method
error NN O I-Method
estimates NN O I-Method
for NN O O
the NN O O
optimal NN O I-OtherScientificTerm
rule NN O I-OtherScientificTerm
. NN O O

When NN O O
the NN O O
competence NN O I-OtherScientificTerm
levels NN O I-OtherScientificTerm
are NN O O
unknown NN O O
, NN O O
they NN O O
must NN O O
be NN O O
empirically NN O O
estimated NN O O
. NN O O

We NN O O
provide NN O O
frequentist NN O O
and NN O O
Bayesian NN O I-Method
analyses NN O I-Method
for NN O O
this NN O O
situation NN O O
. NN O O

Some NN O O
of NN O O
our NN O O
proof NN O O
techniques NN O O
are NN O O
non-standard NN O O
and NN O O
may NN O O
be NN O O
of NN O O
independent NN O O
interest NN O O
. NN O O

The NN O O
bounds NN O O
we NN O O
derive NN O O
are NN O O
nearly NN O O
optimal NN O O
, NN O O
and NN O O
several NN O O
challenging NN O O
open NN O O
problems NN O O
are NN O O
posed NN O O
. NN O O

Experimental NN O O
results NN O O
are NN O O
provided NN O O
to NN O O
illustrate NN O O
the NN O O
theory NN O O
. NN O O

-DOCSTART- NN O O

We NN O O
analyze NN O O
a NN O O
reweighted NN O I-Method
version NN O I-Method
of NN O I-Method
the NN O I-Method
Kikuchi NN O I-Method
approximation NN O I-Method
for NN O O
estimating NN O O
the NN O O
log NN O I-Task
partition NN O I-Task
function NN O I-Task
of NN O I-Task
a NN O I-Task
product NN O I-Task
distribution NN O I-Task
defined NN O O
over NN O O
a NN O O
region NN O I-OtherScientificTerm
graph NN O I-OtherScientificTerm
. NN O O

We NN O O
establish NN O O
sufficient NN O O
conditions NN O O
for NN O O
the NN O O
concavity NN O I-OtherScientificTerm
of NN O O
our NN O O
reweighted NN O I-OtherScientificTerm
objective NN O I-OtherScientificTerm
function NN O I-OtherScientificTerm
in NN O O
terms NN O O
of NN O O
weight NN O I-OtherScientificTerm
assignments NN O I-OtherScientificTerm
in NN O O
the NN O O
Kikuchi NN O I-OtherScientificTerm
expansion NN O I-OtherScientificTerm
, NN O O
and NN O O
show NN O O
that NN O O
a NN O O
reweighted NN O I-Method
version NN O I-Method
of NN O I-Method
the NN O I-Method
sum NN O I-Method
product NN O I-Method
algorithm NN O I-Method
applied NN O O
to NN O O
the NN O O
Kikuchi NN O I-OtherScientificTerm
region NN O I-OtherScientificTerm
graph NN O I-OtherScientificTerm
will NN O O
produce NN O O
global NN O I-OtherScientificTerm
optima NN O I-OtherScientificTerm
of NN O O
the NN O O
Kikuchi NN O I-Method
approximation NN O I-Method
whenever NN O O
the NN O O
algorithm NN O I-Generic
converges NN O O
. NN O O

When NN O O
the NN O O
region NN O I-OtherScientificTerm
graph NN O I-OtherScientificTerm
has NN O O
two NN O O
layers NN O O
, NN O O
corresponding NN O O
to NN O O
a NN O O
Bethe NN O I-Method
approximation NN O I-Method
, NN O O
we NN O O
show NN O O
that NN O O
our NN O O
sufficient NN O O
conditions NN O O
for NN O O
concavity NN O I-OtherScientificTerm
are NN O O
also NN O O
necessary NN O O
. NN O O

Finally NN O O
, NN O O
we NN O O
provide NN O O
an NN O O
explicit NN O O
characterization NN O O
of NN O O
the NN O O
polytope NN O O
of NN O O
concavity NN O I-OtherScientificTerm
in NN O O
terms NN O O
of NN O O
the NN O O
cycle NN O I-OtherScientificTerm
structure NN O I-OtherScientificTerm
of NN O O
the NN O O
region NN O I-OtherScientificTerm
graph NN O I-OtherScientificTerm
. NN O O

We NN O O
conclude NN O O
with NN O O
simulations NN O O
that NN O O
demonstrate NN O O
the NN O O
advantages NN O O
of NN O O
the NN O O
reweighted NN O I-Method
Kikuchi NN O I-Method
approach NN O I-Method
. NN O O

-DOCSTART- NN O O

We NN O O
apply NN O O
a NN O O
decision NN O I-Method
tree NN O I-Method
based NN O I-Method
approach NN O I-Method
to NN O O
pronoun NN O I-Task
resolution NN O I-Task
in NN O O
spoken NN O I-Task
dialogue NN O I-Task
. NN O O

Our NN O O
system NN O I-Generic
deals NN O O
with NN O O
pronouns NN O I-OtherScientificTerm
with NN O O
NP NN O I-OtherScientificTerm
- NN O I-OtherScientificTerm
and NN O I-OtherScientificTerm
non-NP-antecedents NN O I-OtherScientificTerm
. NN O O

We NN O O
present NN O O
a NN O O
set NN O O
of NN O O
features NN O I-OtherScientificTerm
designed NN O O
for NN O O
pronoun NN O I-Task
resolution NN O I-Task
in NN O O
spoken NN O I-Task
dialogue NN O I-Task
and NN O O
determine NN O O
the NN O O
most NN O O
promising NN O O
features NN O I-OtherScientificTerm
. NN O O

We NN O O
evaluate NN O O
the NN O O
system NN O I-Generic
on NN O O
twenty NN O O
Switchboard NN O I-Material
dialogues NN O I-Material
and NN O O
show NN O O
that NN O O
it NN O I-Generic
compares NN O O
well NN O O
to NN O O
Byron NN O I-Method
's NN O I-Method
-LRB- NN O I-Method
2002 NN O I-Method
-RRB- NN O I-Method
manually NN O I-Method
tuned NN O I-Method
system NN O I-Method
. NN O O

-DOCSTART- NN O O

We NN O O
present NN O O
a NN O O
new NN O O
approach NN O I-Generic
for NN O O
building NN O O
an NN O O
efficient NN O O
and NN O O
robust NN O O
classifier NN O I-Method
for NN O O
the NN O O
two NN O O
class NN O I-Task
problem NN O I-Task
, NN O O
that NN O O
localizes NN O O
objects NN O O
that NN O O
may NN O O
appear NN O O
in NN O O
the NN O O
image NN O I-Material
under NN O O
different NN O O
orien-tations NN O I-OtherScientificTerm
. NN O O

In NN O O
contrast NN O O
to NN O O
other NN O O
works NN O O
that NN O O
address NN O O
this NN O O
problem NN O I-Generic
using NN O O
multiple NN O O
classifiers NN O I-Method
, NN O O
each NN O O
one NN O O
specialized NN O O
for NN O O
a NN O O
specific NN O O
orientation NN O O
, NN O O
we NN O O
propose NN O O
a NN O O
simple NN O O
two-step NN O O
approach NN O I-Generic
with NN O O
an NN O O
estimation NN O I-Method
stage NN O I-Method
and NN O O
a NN O O
classification NN O I-Method
stage NN O I-Method
. NN O O

The NN O O
estimator NN O I-Method
yields NN O O
an NN O O
initial NN O O
set NN O O
of NN O O
potential NN O O
object NN O I-OtherScientificTerm
poses NN O I-OtherScientificTerm
that NN O O
are NN O O
then NN O O
validated NN O O
by NN O O
the NN O O
classifier NN O I-Method
. NN O O

This NN O O
methodology NN O O
allows NN O O
reducing NN O O
the NN O O
time NN O I-Metric
complexity NN O I-Metric
of NN O O
the NN O O
algorithm NN O I-Generic
while NN O O
classification NN O I-Task
results NN O O
remain NN O O
high NN O O
. NN O O

The NN O O
classifier NN O I-Method
we NN O O
use NN O O
in NN O O
both NN O O
stages NN O O
is NN O O
based NN O O
on NN O O
a NN O O
boosted NN O I-Method
combination NN O I-Method
of NN O I-Method
Random NN O I-Method
Ferns NN O I-Method
over NN O O
local NN O I-OtherScientificTerm
histograms NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
oriented NN O I-OtherScientificTerm
gradients NN O I-OtherScientificTerm
-LRB- NN O I-OtherScientificTerm
HOGs NN O I-OtherScientificTerm
-RRB- NN O I-OtherScientificTerm
, NN O O
which NN O O
we NN O O
compute NN O O
during NN O O
a NN O O
pre-processing NN O I-Method
step NN O I-Method
. NN O O

Both NN O O
the NN O O
use NN O O
of NN O O
supervised NN O I-Method
learning NN O I-Method
and NN O O
working NN O O
on NN O O
the NN O O
gradient NN O I-OtherScientificTerm
space NN O I-OtherScientificTerm
makes NN O O
our NN O O
approach NN O I-Generic
robust NN O O
while NN O O
being NN O O
efficient NN O O
at NN O O
run-time NN O O
. NN O O

We NN O O
show NN O O
these NN O O
properties NN O O
by NN O O
thorough NN O O
testing NN O O
on NN O O
standard NN O O
databases NN O O
and NN O O
on NN O O
a NN O O
new NN O O
database NN O I-Generic
made NN O O
of NN O O
motorbikes NN O I-Material
under NN O I-Material
planar NN O I-Material
rotations NN O I-Material
, NN O O
and NN O O
with NN O O
challenging NN O O
conditions NN O I-Generic
such NN O O
as NN O O
cluttered NN O I-OtherScientificTerm
backgrounds NN O I-OtherScientificTerm
, NN O O
changing NN O I-OtherScientificTerm
illumination NN O I-OtherScientificTerm
conditions NN O I-OtherScientificTerm
and NN O O
partial NN O I-OtherScientificTerm
occlusions NN O I-OtherScientificTerm
. NN O O

-DOCSTART- NN O O

The NN O O
following NN O O
describes NN O O
recent NN O O
work NN O O
on NN O O
the NN O O
Lincoln NN O I-Method
CSR NN O I-Method
system NN O I-Method
. NN O O

Some NN O O
new NN O O
variations NN O O
in NN O O
semiphone NN O I-Method
modeling NN O I-Method
have NN O O
been NN O O
tested NN O O
. NN O O

A NN O O
very NN O O
simple NN O O
improved NN O O
duration NN O I-Method
model NN O I-Method
has NN O O
reduced NN O O
the NN O O
error NN O I-Metric
rate NN O I-Metric
by NN O O
about NN O O
10 NN O O
% NN O O
in NN O O
both NN O O
triphone NN O I-Method
and NN O I-Method
semiphone NN O I-Method
systems NN O I-Method
. NN O O

A NN O O
new NN O O
training NN O I-Method
strategy NN O I-Method
has NN O O
been NN O O
tested NN O O
which NN O O
, NN O O
by NN O O
itself NN O O
, NN O O
did NN O O
not NN O O
provide NN O O
useful NN O O
improvements NN O O
but NN O O
suggests NN O O
that NN O O
improvements NN O O
can NN O O
be NN O O
obtained NN O O
by NN O O
a NN O O
related NN O O
rapid NN O I-Method
adaptation NN O I-Method
technique NN O I-Method
. NN O O

Finally NN O O
, NN O O
the NN O O
recognizer NN O I-Method
has NN O O
been NN O O
modified NN O O
to NN O O
use NN O O
bigram NN O I-Method
back-off NN O I-Method
language NN O I-Method
models NN O I-Method
. NN O O

The NN O O
system NN O I-Generic
was NN O O
then NN O O
transferred NN O O
from NN O O
the NN O O
RM NN O I-Task
task NN O I-Task
to NN O O
the NN O O
ATIS NN O I-Task
CSR NN O I-Task
task NN O I-Task
and NN O O
a NN O O
limited NN O O
number NN O O
of NN O O
development NN O O
tests NN O O
performed NN O O
. NN O O

Evaluation NN O O
test NN O O
results NN O O
are NN O O
presented NN O O
for NN O O
both NN O O
the NN O O
RM NN O I-Task
and NN O I-Task
ATIS NN O I-Task
CSR NN O I-Task
tasks NN O I-Task
. NN O O

-DOCSTART- NN O O

A NN O O
new NN O O
approach NN O I-Generic
for NN O O
Interactive NN O I-Task
Machine NN O I-Task
Translation NN O I-Task
where NN O O
the NN O O
author NN O O
interacts NN O O
during NN O O
the NN O O
creation NN O O
or NN O O
the NN O O
modification NN O O
of NN O O
the NN O O
document NN O O
is NN O O
proposed NN O O
. NN O O

The NN O O
explanation NN O O
of NN O O
an NN O O
ambiguity NN O O
or NN O O
an NN O O
error NN O O
for NN O O
the NN O O
purposes NN O O
of NN O O
correction NN O O
does NN O O
not NN O O
use NN O O
any NN O O
concepts NN O O
of NN O O
the NN O O
underlying NN O O
linguistic NN O I-Method
theory NN O I-Method
: NN O O
it NN O O
is NN O O
a NN O O
reformulation NN O O
of NN O O
the NN O O
erroneous NN O O
or NN O O
ambiguous NN O O
sentence NN O O
. NN O O

The NN O O
interaction NN O O
is NN O O
limited NN O O
to NN O O
the NN O O
analysis NN O O
step NN O O
of NN O O
the NN O O
translation NN O I-Method
process NN O I-Method
. NN O O

This NN O O
paper NN O O
presents NN O O
a NN O O
new NN O O
interactive NN O I-Method
disambiguation NN O I-Method
scheme NN O I-Method
based NN O O
on NN O O
the NN O O
paraphrasing NN O I-Method
of NN O O
a NN O O
parser NN O O
's NN O O
multiple NN O O
output NN O O
. NN O O

Some NN O O
examples NN O O
of NN O O
paraphrasing NN O O
ambiguous NN O O
sentences NN O O
are NN O O
presented NN O O
. NN O O

-DOCSTART- NN O O

We NN O O
describe NN O O
a NN O O
novel NN O O
approach NN O I-Generic
to NN O O
statistical NN O I-Task
machine NN O I-Task
translation NN O I-Task
that NN O O
combines NN O O
syntactic NN O I-OtherScientificTerm
information NN O I-OtherScientificTerm
in NN O O
the NN O O
source NN O O
language NN O O
with NN O O
recent NN O O
advances NN O O
in NN O O
phrasal NN O I-Task
translation NN O I-Task
. NN O O

This NN O O
method NN O I-Generic
requires NN O O
a NN O O
source-language NN O I-Method
dependency NN O I-Method
parser NN O I-Method
, NN O O
target NN O I-Method
language NN O I-Method
word NN O I-Method
segmentation NN O I-Method
and NN O O
an NN O O
unsupervised NN O I-Method
word NN O I-Method
alignment NN O I-Method
component NN O I-Method
. NN O O

We NN O O
align NN O O
a NN O O
parallel NN O I-Material
corpus NN O I-Material
, NN O O
project NN O O
the NN O O
source NN O I-OtherScientificTerm
dependency NN O I-OtherScientificTerm
parse NN O I-OtherScientificTerm
onto NN O O
the NN O O
target NN O O
sentence NN O O
, NN O O
extract NN O O
dependency NN O I-OtherScientificTerm
treelet NN O I-OtherScientificTerm
translation NN O I-OtherScientificTerm
pairs NN O I-OtherScientificTerm
, NN O O
and NN O O
train NN O O
a NN O O
tree-based NN O I-Method
ordering NN O I-Method
model NN O I-Method
. NN O O

We NN O O
describe NN O O
an NN O O
efficient NN O O
decoder NN O I-Method
and NN O O
show NN O O
that NN O O
using NN O O
these NN O O
tree-based NN O I-Method
models NN O I-Method
in NN O O
combination NN O O
with NN O O
conventional NN O O
SMT NN O I-Method
models NN O I-Method
provides NN O O
a NN O O
promising NN O O
approach NN O I-Generic
that NN O O
incorporates NN O O
the NN O O
power NN O O
of NN O O
phrasal NN O I-Method
SMT NN O I-Method
with NN O O
the NN O O
linguistic NN O I-OtherScientificTerm
generality NN O I-OtherScientificTerm
available NN O O
in NN O O
a NN O O
parser NN O I-Method
. NN O O

-DOCSTART- NN O O

Video NN O I-Material
provides NN O O
not NN O O
only NN O O
rich NN O O
visual NN O I-OtherScientificTerm
cues NN O I-OtherScientificTerm
such NN O O
as NN O O
motion NN O I-OtherScientificTerm
and NN O O
appearance NN O I-OtherScientificTerm
, NN O O
but NN O O
also NN O O
much NN O O
less NN O O
explored NN O O
long-range NN O I-OtherScientificTerm
temporal NN O I-OtherScientificTerm
interactions NN O I-OtherScientificTerm
among NN O O
objects NN O O
. NN O O

We NN O O
aim NN O O
to NN O O
capture NN O O
such NN O O
interactions NN O I-Generic
and NN O O
to NN O O
construct NN O O
a NN O O
powerful NN O O
intermediate-level NN O I-Method
video NN O I-Method
representation NN O I-Method
for NN O O
subsequent NN O O
recognition NN O I-Task
. NN O O

Motivated NN O O
by NN O O
this NN O O
goal NN O O
, NN O O
we NN O O
seek NN O O
to NN O O
obtain NN O O
spatio-temporal NN O I-OtherScientificTerm
over-segmentation NN O I-OtherScientificTerm
of NN O O
a NN O O
video NN O O
into NN O O
regions NN O O
that NN O O
respect NN O O
object NN O I-OtherScientificTerm
boundaries NN O I-OtherScientificTerm
and NN O O
, NN O O
at NN O O
the NN O O
same NN O O
time NN O O
, NN O O
associate NN O O
object NN O O
pix-els NN O O
over NN O O
many NN O O
video NN O O
frames NN O O
. NN O O

The NN O O
contributions NN O O
of NN O O
this NN O O
paper NN O O
are NN O O
twofold NN O O
. NN O O

First NN O O
, NN O O
we NN O O
develop NN O O
an NN O O
efficient NN O O
spatio-temporal NN O I-Method
video NN O I-Method
segmentation NN O I-Method
algorithm NN O I-Method
, NN O O
which NN O O
naturally NN O O
incorporates NN O O
long-range NN O I-OtherScientificTerm
motion NN O I-OtherScientificTerm
cues NN O I-OtherScientificTerm
from NN O O
the NN O O
past NN O O
and NN O O
future NN O O
frames NN O O
in NN O O
the NN O O
form NN O O
of NN O O
clusters NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
point NN O I-OtherScientificTerm
tracks NN O I-OtherScientificTerm
with NN O O
coherent NN O O
motion NN O O
. NN O O

Second NN O O
, NN O O
we NN O O
devise NN O O
a NN O O
new NN O O
track NN O I-Method
clustering NN O I-Method
cost NN O I-Method
function NN O I-Method
that NN O O
includes NN O O
occlusion NN O I-OtherScientificTerm
reasoning NN O I-OtherScientificTerm
, NN O O
in NN O O
the NN O O
form NN O O
of NN O O
depth NN O I-OtherScientificTerm
ordering NN O I-OtherScientificTerm
constraints NN O I-OtherScientificTerm
, NN O O
as NN O O
well NN O O
as NN O O
motion NN O I-OtherScientificTerm
similarity NN O I-OtherScientificTerm
along NN O O
the NN O O
tracks NN O O
. NN O O

We NN O O
evaluate NN O O
the NN O O
proposed NN O O
approach NN O I-Generic
on NN O O
a NN O O
challenging NN O O
set NN O O
of NN O O
video NN O I-Material
sequences NN O I-Material
of NN O I-Material
office NN O I-Material
scenes NN O I-Material
from NN O O
feature NN O O
length NN O O
movies NN O I-Material
. NN O O

-DOCSTART- NN O O

In NN O O
this NN O O
paper NN O O
, NN O O
we NN O O
introduce NN O O
KAZE NN O I-Method
features NN O I-Method
, NN O O
a NN O O
novel NN O O
multiscale NN O I-Method
2D NN O I-Method
feature NN O I-Method
detection NN O I-Method
and NN O I-Method
description NN O I-Method
algorithm NN O I-Method
in NN O O
nonlinear NN O I-OtherScientificTerm
scale NN O I-OtherScientificTerm
spaces NN O I-OtherScientificTerm
. NN O O

Previous NN O O
approaches NN O O
detect NN O O
and NN O O
describe NN O O
features NN O O
at NN O O
different NN O O
scale NN O O
levels NN O O
by NN O O
building NN O O
or NN O O
approximating NN O O
the NN O O
Gaussian NN O I-OtherScientificTerm
scale NN O I-OtherScientificTerm
space NN O I-OtherScientificTerm
of NN O O
an NN O O
image NN O O
. NN O O

However NN O O
, NN O O
Gaussian NN O I-OtherScientificTerm
blurring NN O I-OtherScientificTerm
does NN O O
not NN O O
respect NN O O
the NN O O
natural NN O O
boundaries NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
objects NN O I-OtherScientificTerm
and NN O O
smoothes NN O O
to NN O O
the NN O O
same NN O O
degree NN O O
both NN O O
details NN O O
and NN O O
noise NN O O
, NN O O
reducing NN O O
localization NN O I-Metric
accuracy NN O I-Metric
and NN O I-Metric
distinctiveness NN O I-Metric
. NN O O

In NN O O
contrast NN O O
, NN O O
we NN O O
detect NN O O
and NN O O
describe NN O O
2D NN O I-Method
features NN O I-Method
in NN O O
a NN O O
nonlinear NN O I-OtherScientificTerm
scale NN O I-OtherScientificTerm
space NN O I-OtherScientificTerm
by NN O O
means NN O O
of NN O O
nonlinear NN O I-Method
diffusion NN O I-Method
filtering NN O I-Method
. NN O O

In NN O O
this NN O O
way NN O O
, NN O O
we NN O O
can NN O O
make NN O O
blurring NN O O
locally NN O O
adaptive NN O O
to NN O O
the NN O O
image NN O I-Material
data NN O I-Material
, NN O O
reducing NN O O
noise NN O O
but NN O O
retaining NN O O
object NN O I-OtherScientificTerm
boundaries NN O I-OtherScientificTerm
, NN O O
obtaining NN O O
superior NN O O
localization NN O I-Metric
accuracy NN O I-Metric
and NN O I-Metric
distinctiviness NN O I-Metric
. NN O O

The NN O O
nonlinear NN O I-OtherScientificTerm
scale NN O I-OtherScientificTerm
space NN O I-OtherScientificTerm
is NN O O
built NN O O
using NN O O
efficient NN O O
Additive NN O I-Method
Operator NN O I-Method
Splitting NN O I-Method
-LRB- NN O I-Method
AOS NN O I-Method
-RRB- NN O I-Method
techniques NN O I-Method
and NN O O
variable NN O I-Method
con-ductance NN O I-Method
diffusion NN O I-Method
. NN O O

We NN O O
present NN O O
an NN O O
extensive NN O O
evaluation NN O O
on NN O O
benchmark NN O I-Material
datasets NN O I-Material
and NN O O
a NN O O
practical NN O O
matching NN O I-Task
application NN O I-Task
on NN O I-Task
deformable NN O I-Task
surfaces NN O I-Task
. NN O O

Even NN O O
though NN O O
our NN O O
features NN O I-Generic
are NN O O
somewhat NN O O
more NN O O
expensive NN O O
to NN O O
compute NN O O
than NN O O
SURF NN O I-Method
due NN O O
to NN O O
the NN O O
construction NN O O
of NN O O
the NN O O
nonlinear NN O I-OtherScientificTerm
scale NN O I-OtherScientificTerm
space NN O I-OtherScientificTerm
, NN O O
but NN O O
comparable NN O O
to NN O O
SIFT NN O I-Method
, NN O O
our NN O O
results NN O I-Generic
reveal NN O O
a NN O O
step NN O O
forward NN O O
in NN O O
performance NN O O
both NN O O
in NN O O
detection NN O I-Task
and NN O O
description NN O I-Task
against NN O O
previous NN O O
state-of-the-art NN O I-Generic
methods NN O I-Generic
. NN O O

-DOCSTART- NN O O

Semantic NN O I-Material
Web NN O I-Material
documents NN O I-Material
that NN O O
encode NN O O
facts NN O O
about NN O O
entities NN O O
on NN O O
the NN O O
Web NN O O
have NN O O
been NN O O
growing NN O O
rapidly NN O O
in NN O O
size NN O O
and NN O O
evolving NN O O
over NN O O
time NN O O
. NN O O

Creating NN O I-Task
summaries NN O I-Task
on NN O O
lengthy NN O I-Material
Semantic NN O I-Material
Web NN O I-Material
documents NN O I-Material
for NN O O
quick NN O O
identification NN O I-Task
of NN O I-Task
the NN O I-Task
corresponding NN O I-Task
entity NN O I-Task
has NN O O
been NN O O
of NN O O
great NN O O
contemporary NN O O
interest NN O O
. NN O O

In NN O O
this NN O O
paper NN O O
, NN O O
we NN O O
explore NN O O
automatic NN O I-Method
summa-rization NN O I-Method
techniques NN O I-Method
that NN O O
characterize NN O O
and NN O O
enable NN O O
identification NN O O
of NN O O
an NN O O
entity NN O O
and NN O O
create NN O O
summaries NN O O
that NN O O
are NN O O
human NN O O
friendly NN O O
. NN O O

Specifically NN O O
, NN O O
we NN O O
highlight NN O O
the NN O O
importance NN O O
of NN O O
diversified NN O I-OtherScientificTerm
-LRB- NN O I-OtherScientificTerm
faceted NN O I-OtherScientificTerm
-RRB- NN O I-OtherScientificTerm
summaries NN O I-OtherScientificTerm
by NN O O
combining NN O O
three NN O O
dimensions NN O O
: NN O O
diversity NN O I-OtherScientificTerm
, NN O O
uniqueness NN O I-OtherScientificTerm
, NN O O
and NN O O
popularity NN O I-OtherScientificTerm
. NN O O

Our NN O O
novel NN O O
diversity-aware NN O I-Method
entity NN O I-Method
summarization NN O I-Method
approach NN O I-Method
mimics NN O O
human NN O I-Method
conceptual NN O I-Method
clustering NN O I-Method
techniques NN O I-Method
to NN O O
group NN O O
facts NN O O
, NN O O
and NN O O
picks NN O O
representative NN O O
facts NN O O
from NN O O
each NN O O
group NN O O
to NN O O
form NN O O
concise NN O O
-LRB- NN O O
i.e. NN O O
, NN O O
short NN O O
-RRB- NN O O
and NN O O
comprehensive NN O O
-LRB- NN O O
i.e. NN O O
, NN O O
improved NN O O
coverage NN O O
through NN O O
diversity NN O O
-RRB- NN O O
summaries NN O O
. NN O O

We NN O O
evaluate NN O O
our NN O O
approach NN O I-Generic
against NN O O
the NN O O
state-of-the-art NN O I-Generic
techniques NN O I-Generic
and NN O O
show NN O O
that NN O O
our NN O O
work NN O O
improves NN O O
both NN O O
the NN O O
quality NN O I-Metric
and NN O O
the NN O O
efficiency NN O I-Metric
of NN O O
entity NN O I-Task
summarization NN O I-Task
. NN O O

-DOCSTART- NN O O

We NN O O
present NN O O
a NN O O
framework NN O I-Generic
for NN O O
the NN O O
fast NN O I-Task
computation NN O I-Task
of NN O I-Task
lexical NN O I-Task
affinity NN O I-Task
models NN O I-Task
. NN O O

The NN O O
framework NN O I-Generic
is NN O O
composed NN O O
of NN O O
a NN O O
novel NN O O
algorithm NN O I-Generic
to NN O O
efficiently NN O O
compute NN O O
the NN O O
co-occurrence NN O I-OtherScientificTerm
distribution NN O I-OtherScientificTerm
between NN O O
pairs NN O O
of NN O O
terms NN O O
, NN O O
an NN O O
independence NN O I-Method
model NN O I-Method
, NN O O
and NN O O
a NN O O
parametric NN O I-Method
affinity NN O I-Method
model NN O I-Method
. NN O O

In NN O O
comparison NN O O
with NN O O
previous NN O O
models NN O I-Generic
, NN O O
which NN O O
either NN O O
use NN O O
arbitrary NN O O
windows NN O O
to NN O O
compute NN O O
similarity NN O I-OtherScientificTerm
between NN O O
words NN O O
or NN O O
use NN O O
lexical NN O I-OtherScientificTerm
affinity NN O I-OtherScientificTerm
to NN O O
create NN O O
sequential NN O I-Method
models NN O I-Method
, NN O O
in NN O O
this NN O O
paper NN O O
we NN O O
focus NN O O
on NN O O
models NN O I-Generic
intended NN O O
to NN O O
capture NN O O
the NN O O
co-occurrence NN O I-OtherScientificTerm
patterns NN O I-OtherScientificTerm
of NN O O
any NN O O
pair NN O O
of NN O O
words NN O O
or NN O O
phrases NN O O
at NN O O
any NN O O
distance NN O O
in NN O O
the NN O O
corpus NN O O
. NN O O

The NN O O
framework NN O I-Generic
is NN O O
flexible NN O O
, NN O O
allowing NN O O
fast NN O O
adaptation NN O O
to NN O O
applications NN O O
and NN O O
it NN O O
is NN O O
scalable NN O O
. NN O O

We NN O O
apply NN O O
it NN O I-Generic
in NN O O
combination NN O O
with NN O O
a NN O O
terabyte NN O I-Material
corpus NN O I-Material
to NN O O
answer NN O O
natural NN O I-Task
language NN O I-Task
tests NN O I-Task
, NN O O
achieving NN O O
encouraging NN O O
results NN O O
. NN O O

-DOCSTART- NN O O

This NN O O
paper NN O O
introduces NN O O
a NN O O
system NN O I-Generic
for NN O O
categorizing NN O I-Task
unknown NN O I-Task
words NN O I-Task
. NN O O

The NN O O
system NN O I-Generic
is NN O O
based NN O O
on NN O O
a NN O O
multi-component NN O I-Method
architecture NN O I-Method
where NN O O
each NN O O
component NN O I-Generic
is NN O O
responsible NN O O
for NN O O
identifying NN O O
one NN O O
class NN O O
of NN O O
unknown NN O I-OtherScientificTerm
words NN O I-OtherScientificTerm
. NN O O

The NN O O
focus NN O O
of NN O O
this NN O O
paper NN O O
is NN O O
the NN O O
components NN O I-Generic
that NN O O
identify NN O O
names NN O I-OtherScientificTerm
and NN O O
spelling NN O I-OtherScientificTerm
errors NN O I-OtherScientificTerm
. NN O O

Each NN O O
component NN O I-Generic
uses NN O O
a NN O O
decision NN O I-Method
tree NN O I-Method
architecture NN O I-Method
to NN O O
combine NN O O
multiple NN O O
types NN O O
of NN O O
evidence NN O O
about NN O O
the NN O O
unknown NN O I-OtherScientificTerm
word NN O I-OtherScientificTerm
. NN O O

The NN O O
system NN O I-Generic
is NN O O
evaluated NN O O
using NN O O
data NN O O
from NN O O
live NN O I-Material
closed NN O I-Material
captions NN O I-Material
- NN O O
a NN O O
genre NN O O
replete NN O O
with NN O O
a NN O O
wide NN O O
variety NN O O
of NN O O
unknown NN O I-OtherScientificTerm
words NN O I-OtherScientificTerm
. NN O O

-DOCSTART- NN O O

At NN O O
MIT NN O O
Lincoln NN O O
Laboratory NN O O
, NN O O
we NN O O
have NN O O
been NN O O
developing NN O O
a NN O O
Korean-to-English NN O I-Method
machine NN O I-Method
translation NN O I-Method
system NN O I-Method
CCLINC NN O I-Method
-LRB- NN O I-Method
Common NN O I-Method
Coalition NN O I-Method
Language NN O I-Method
System NN O I-Method
at NN O I-Method
Lincoln NN O I-Method
Laboratory NN O I-Method
-RRB- NN O I-Method
. NN O O

The NN O O
CCLINC NN O I-Method
Korean-to-English NN O I-Method
translation NN O I-Method
system NN O I-Method
consists NN O O
of NN O O
two NN O O
core NN O I-Generic
modules NN O I-Generic
, NN O O
language NN O I-Method
understanding NN O I-Method
and NN O I-Method
generation NN O I-Method
modules NN O I-Method
mediated NN O O
by NN O O
a NN O O
language NN O I-Method
neutral NN O I-Method
meaning NN O I-Method
representation NN O I-Method
called NN O O
a NN O O
semantic NN O I-OtherScientificTerm
frame NN O I-OtherScientificTerm
. NN O O

The NN O O
key NN O O
features NN O O
of NN O O
the NN O O
system NN O I-Generic
include NN O O
: NN O O
-LRB- NN O O
i NN O O
-RRB- NN O O
Robust NN O O
efficient NN O O
parsing NN O I-Task
of NN O I-Task
Korean NN O I-Material
-LRB- NN O O
a NN O O
verb NN O I-Material
final NN O I-Material
language NN O I-Material
with NN O O
overt NN O I-OtherScientificTerm
case NN O I-OtherScientificTerm
markers NN O I-OtherScientificTerm
, NN O O
relatively NN O O
free NN O O
word NN O O
order NN O O
, NN O O
and NN O O
frequent NN O O
omissions NN O O
of NN O O
arguments NN O O
-RRB- NN O O
. NN O O

-LRB- NN O O
ii NN O O
-RRB- NN O O
High NN O O
quality NN O O
translation NN O I-Task
via NN O O
word NN O I-Method
sense NN O I-Method
disambiguation NN O I-Method
and NN O O
accurate NN O O
word NN O I-Task
order NN O I-Task
generation NN O I-Task
of NN O O
the NN O O
target NN O O
language NN O O
. NN O O

-LRB- NN O O
iii NN O O
-RRB- NN O O
Rapid NN O O
system NN O O
development NN O O
and NN O O
porting NN O O
to NN O O
new NN O O
domains NN O O
via NN O O
knowledge-based NN O I-Method
automated NN O I-Method
acquisition NN O I-Method
of NN O I-Method
grammars NN O I-Method
. NN O O

Having NN O O
been NN O O
trained NN O O
on NN O O
Korean NN O I-Material
newspaper NN O I-Material
articles NN O I-Material
on NN O O
missiles NN O I-Material
and NN O I-Material
chemical NN O I-Material
biological NN O I-Material
warfare NN O I-Material
, NN O O
the NN O O
system NN O I-Generic
produces NN O O
the NN O O
translation NN O O
output NN O O
sufficient NN O O
for NN O O
content NN O O
understanding NN O O
of NN O O
the NN O O
original NN O O
document NN O O
. NN O O

-DOCSTART- NN O O

The NN O O
JAVELIN NN O I-Method
system NN O I-Method
integrates NN O O
a NN O O
flexible NN O O
, NN O O
planning-based NN O I-Method
architecture NN O I-Method
with NN O O
a NN O O
variety NN O O
of NN O O
language NN O I-Method
processing NN O I-Method
modules NN O I-Method
to NN O O
provide NN O O
an NN O O
open-domain NN O I-Task
question NN O I-Task
answering NN O I-Task
capability NN O I-Task
on NN O O
free NN O O
text NN O O
. NN O O

The NN O O
demonstration NN O O
will NN O O
focus NN O O
on NN O O
how NN O O
JAVELIN NN O I-Method
processes NN O O
questions NN O O
and NN O O
retrieves NN O O
the NN O O
most NN O O
likely NN O O
answer NN O O
candidates NN O O
from NN O O
the NN O O
given NN O O
text NN O O
corpus NN O O
. NN O O

The NN O O
operation NN O O
of NN O O
the NN O O
system NN O I-Generic
will NN O O
be NN O O
explained NN O O
in NN O O
depth NN O O
through NN O O
browsing NN O O
the NN O O
repository NN O O
of NN O O
data NN O O
objects NN O O
created NN O O
by NN O O
the NN O O
system NN O I-Generic
during NN O O
each NN O O
question NN O I-Task
answering NN O I-Task
session NN O O
. NN O O

-DOCSTART- NN O O

We NN O O
present NN O O
the NN O O
first NN O O
application NN O O
of NN O O
the NN O O
head-driven NN O I-Method
statistical NN O I-Method
parsing NN O I-Method
model NN O I-Method
of NN O O
Collins NN O O
-LRB- NN O O
1999 NN O O
-RRB- NN O O
as NN O O
a NN O O
simultaneous NN O I-Method
language NN O I-Method
model NN O I-Method
and NN O O
parser NN O I-Method
for NN O O
large-vocabulary NN O I-Task
speech NN O I-Task
recognition NN O I-Task
. NN O O

The NN O O
model NN O I-Generic
is NN O O
adapted NN O O
to NN O O
an NN O O
online NN O I-Method
left NN O I-Method
to NN O I-Method
right NN O I-Method
chart-parser NN O I-Method
for NN O O
word NN O I-OtherScientificTerm
lattices NN O I-OtherScientificTerm
, NN O O
integrating NN O O
acoustic NN O I-OtherScientificTerm
, NN O I-OtherScientificTerm
n-gram NN O I-OtherScientificTerm
, NN O I-OtherScientificTerm
and NN O I-OtherScientificTerm
parser NN O I-OtherScientificTerm
probabilities NN O I-OtherScientificTerm
. NN O O

The NN O O
parser NN O I-Method
uses NN O O
structural NN O I-OtherScientificTerm
and NN O I-OtherScientificTerm
lexical NN O I-OtherScientificTerm
dependencies NN O I-OtherScientificTerm
not NN O O
considered NN O O
by NN O O
n-gram NN O I-Method
models NN O I-Method
, NN O O
conditioning NN O O
recognition NN O O
on NN O O
more NN O O
linguistically-grounded NN O O
relationships NN O O
. NN O O

Experiments NN O O
on NN O O
the NN O O
Wall NN O I-Material
Street NN O I-Material
Journal NN O I-Material
treebank NN O I-Material
and NN O O
lattice NN O I-Material
corpora NN O I-Material
show NN O O
word NN O I-Metric
error NN O I-Metric
rates NN O I-Metric
competitive NN O O
with NN O O
the NN O O
standard NN O O
n-gram NN O I-Method
language NN O I-Method
model NN O I-Method
while NN O O
extracting NN O O
additional NN O O
structural NN O I-OtherScientificTerm
information NN O I-OtherScientificTerm
useful NN O O
for NN O O
speech NN O I-Task
understanding NN O I-Task
. NN O O

-DOCSTART- NN O O

Image NN O I-Task
composition NN O I-Task
-LRB- NN O I-Task
or NN O I-Task
mosaicing NN O I-Task
-RRB- NN O I-Task
has NN O O
attracted NN O O
a NN O O
growing NN O O
attention NN O O
in NN O O
recent NN O O
years NN O O
as NN O O
one NN O O
of NN O O
the NN O O
main NN O O
elements NN O O
in NN O O
video NN O I-Task
analysis NN O I-Task
and NN O I-Task
representation NN O I-Task
. NN O O

In NN O O
this NN O O
paper NN O O
we NN O O
deal NN O O
with NN O O
the NN O O
problem NN O O
of NN O O
global NN O I-Task
alignment NN O I-Task
and NN O O
super-resolution NN O I-Task
. NN O O

We NN O O
also NN O O
propose NN O O
to NN O O
evaluate NN O O
the NN O O
quality NN O O
of NN O O
the NN O O
resulting NN O O
mosaic NN O I-Task
by NN O O
measuring NN O O
the NN O O
amount NN O I-Metric
of NN O I-Metric
blurring NN O I-Metric
. NN O O

Global NN O I-Task
registration NN O I-Task
is NN O O
achieved NN O O
by NN O O
combining NN O O
a NN O O
graph-based NN O I-Method
technique NN O I-Method
-- NN O O
that NN O O
exploits NN O O
the NN O O
topological NN O I-OtherScientificTerm
structure NN O I-OtherScientificTerm
of NN O O
the NN O O
sequence NN O O
induced NN O O
by NN O O
the NN O O
spatial NN O I-OtherScientificTerm
overlap NN O I-OtherScientificTerm
-- NN O O
with NN O O
a NN O O
bundle NN O I-Method
adjustment NN O I-Method
which NN O O
uses NN O O
only NN O O
the NN O O
homographies NN O I-OtherScientificTerm
computed NN O O
in NN O O
the NN O O
previous NN O O
steps NN O O
. NN O O

Experimental NN O O
comparison NN O O
with NN O O
other NN O O
techniques NN O I-Generic
shows NN O O
the NN O O
effectiveness NN O O
of NN O O
our NN O O
approach NN O I-Generic
. NN O O

-DOCSTART- NN O O

The NN O O
project NN O O
presented NN O O
here NN O O
is NN O O
a NN O O
part NN O O
of NN O O
a NN O O
long NN O O
term NN O O
research NN O O
program NN O O
aiming NN O O
at NN O O
a NN O O
full NN O O
lexicon NN O I-Method
grammar NN O I-Method
for NN O I-Method
Polish NN O I-Method
-LRB- NN O I-Method
SyntLex NN O I-Method
-RRB- NN O I-Method
. NN O O

The NN O O
main NN O O
of NN O O
this NN O O
project NN O O
is NN O O
computer-assisted NN O I-Task
acquisition NN O I-Task
and NN O I-Task
morpho-syntactic NN O I-Task
description NN O I-Task
of NN O I-Task
verb-noun NN O I-Task
collocations NN O I-Task
in NN O O
Polish NN O I-Material
. NN O O

We NN O O
present NN O O
methodology NN O O
and NN O O
resources NN O O
obtained NN O O
in NN O O
three NN O O
main NN O O
project NN O O
phases NN O I-Generic
which NN O O
are NN O O
: NN O O
dictionary-based NN O I-Task
acquisition NN O I-Task
of NN O I-Task
collocation NN O I-Task
lexicon NN O I-Task
, NN O O
feasibility NN O I-Task
study NN O I-Task
for NN O O
corpus-based NN O I-Task
lexicon NN O I-Task
enlargement NN O I-Task
phase NN O I-Task
, NN O O
corpus-based NN O I-Task
lexicon NN O I-Task
enlargement NN O I-Task
and NN O I-Task
collocation NN O I-Task
description NN O I-Task
. NN O O

In NN O O
this NN O O
paper NN O O
we NN O O
focus NN O O
on NN O O
the NN O O
results NN O O
of NN O O
the NN O O
third NN O O
phase NN O O
. NN O O

The NN O O
presented NN O O
here NN O O
corpus-based NN O I-Method
approach NN O I-Method
permitted NN O O
us NN O O
to NN O O
triple NN O O
the NN O O
size NN O O
the NN O O
verb-noun NN O I-Material
collocation NN O I-Material
dictionary NN O I-Material
for NN O O
Polish NN O I-Material
. NN O O

In NN O O
the NN O O
paper NN O O
we NN O O
describe NN O O
the NN O O
SyntLex NN O I-Material
Dictionary NN O I-Material
of NN O I-Material
Collocations NN O I-Material
and NN O O
announce NN O O
some NN O O
future NN O O
research NN O O
intended NN O O
to NN O O
be NN O O
a NN O O
separate NN O O
project NN O O
continuation NN O O
. NN O O

-DOCSTART- NN O O

Along NN O O
with NN O O
the NN O O
increasing NN O O
requirements NN O O
, NN O O
the NN O O
hash-tag NN O I-Task
recommendation NN O I-Task
task NN O I-Task
for NN O O
microblogs NN O I-Material
has NN O O
been NN O O
receiving NN O O
considerable NN O O
attention NN O O
in NN O O
recent NN O O
years NN O O
. NN O O

Various NN O O
researchers NN O O
have NN O O
studied NN O O
the NN O O
problem NN O O
from NN O O
different NN O O
aspects NN O O
. NN O O

However NN O O
, NN O O
most NN O O
of NN O O
these NN O O
methods NN O O
usually NN O O
need NN O O
handcrafted NN O I-OtherScientificTerm
features NN O I-OtherScientificTerm
. NN O O

Motivated NN O O
by NN O O
the NN O O
successful NN O O
use NN O O
of NN O O
convolutional NN O I-Method
neural NN O I-Method
networks NN O I-Method
-LRB- NN O I-Method
CNNs NN O I-Method
-RRB- NN O I-Method
for NN O O
many NN O O
natural NN O I-Task
language NN O I-Task
processing NN O I-Task
tasks NN O I-Task
, NN O O
in NN O O
this NN O O
paper NN O O
, NN O O
we NN O O
adopt NN O O
CNNs NN O I-Method
to NN O O
perform NN O O
the NN O O
hashtag NN O I-Task
recommendation NN O I-Task
problem NN O I-Task
. NN O O

To NN O O
incorporate NN O O
the NN O O
trigger NN O I-OtherScientificTerm
words NN O I-OtherScientificTerm
whose NN O O
effectiveness NN O O
have NN O O
been NN O O
experimentally NN O O
evaluated NN O O
in NN O O
several NN O O
previous NN O O
works NN O O
, NN O O
we NN O O
propose NN O O
a NN O O
novel NN O O
architecture NN O I-Generic
with NN O O
an NN O O
attention NN O I-OtherScientificTerm
mechanism NN O I-OtherScientificTerm
. NN O O

The NN O O
results NN O O
of NN O O
experiments NN O O
on NN O O
the NN O O
data NN O I-Generic
collected NN O O
from NN O O
a NN O O
real NN O O
world NN O O
microblogging NN O O
service NN O O
demonstrated NN O O
that NN O O
the NN O O
proposed NN O O
model NN O I-Generic
outperforms NN O O
state-of-the-art NN O I-Generic
methods NN O I-Generic
. NN O O

By NN O O
incorporating NN O O
trigger NN O I-OtherScientificTerm
words NN O I-OtherScientificTerm
into NN O O
the NN O O
consideration NN O O
, NN O O
the NN O O
relative NN O O
improvement NN O O
of NN O O
the NN O O
proposed NN O O
method NN O I-Generic
over NN O O
the NN O O
state-of-the-art NN O I-Generic
method NN O I-Generic
is NN O O
around NN O O
9.4 NN O O
% NN O O
in NN O O
the NN O O
F1-score NN O I-Metric
. NN O O

-DOCSTART- NN O O

In NN O O
this NN O O
paper NN O O
, NN O O
we NN O O
improve NN O O
an NN O O
unsupervised NN O I-Method
learning NN O I-Method
method NN O I-Method
using NN O O
the NN O O
Expectation-Maximization NN O I-Method
-LRB- NN O I-Method
EM NN O I-Method
-RRB- NN O I-Method
algorithm NN O I-Method
proposed NN O O
by NN O O
Nigam NN O O
et NN O O
al. NN O O
for NN O O
text NN O I-Task
classification NN O I-Task
problems NN O I-Task
in NN O O
order NN O O
to NN O O
apply NN O O
it NN O I-Generic
to NN O O
word NN O I-Task
sense NN O I-Task
disambiguation NN O I-Task
-LRB- NN O I-Task
WSD NN O I-Task
-RRB- NN O I-Task
problems NN O I-Task
. NN O O

The NN O O
improved NN O O
method NN O I-Generic
stops NN O O
the NN O O
EM NN O I-Method
algorithm NN O I-Method
at NN O O
the NN O O
optimum NN O I-OtherScientificTerm
iteration NN O I-OtherScientificTerm
number NN O I-OtherScientificTerm
. NN O O

To NN O O
estimate NN O O
that NN O O
number NN O I-Generic
, NN O O
we NN O O
propose NN O O
two NN O O
methods NN O O
. NN O O

In NN O O
experiments NN O O
, NN O O
we NN O O
solved NN O O
50 NN O O
noun NN O I-Task
WSD NN O I-Task
problems NN O I-Task
in NN O O
the NN O O
Japanese NN O I-Task
Dictionary NN O I-Task
Task NN O I-Task
in NN O O
SENSEVAL2 NN O I-Material
. NN O O

The NN O O
score NN O O
of NN O O
our NN O O
method NN O I-Generic
is NN O O
a NN O O
match NN O O
for NN O O
the NN O O
best NN O O
public NN O O
score NN O O
of NN O O
this NN O O
task NN O I-Generic
. NN O O

Furthermore NN O O
, NN O O
our NN O O
methods NN O I-Generic
were NN O O
confirmed NN O O
to NN O O
be NN O O
effective NN O O
also NN O O
for NN O O
verb NN O I-Task
WSD NN O I-Task
problems NN O I-Task
. NN O O

-DOCSTART- NN O O

Dividing NN O I-Task
sentences NN O I-Task
in NN O I-Task
chunks NN O I-Task
of NN O I-Task
words NN O I-Task
is NN O O
a NN O O
useful NN O O
preprocessing NN O O
step NN O O
for NN O O
parsing NN O I-Task
, NN O O
information NN O I-Task
extraction NN O I-Task
and NN O O
information NN O I-Task
retrieval NN O I-Task
. NN O O

-LRB- NN O O
Ramshaw NN O O
and NN O O
Marcus NN O O
, NN O O
1995 NN O O
-RRB- NN O O
have NN O O
introduced NN O O
a NN O O
`` NN O O
convenient NN O O
'' NN O O
data NN O I-Method
representation NN O I-Method
for NN O O
chunking NN O I-Task
by NN O O
converting NN O O
it NN O I-Generic
to NN O O
a NN O O
tagging NN O I-Task
task NN O I-Task
. NN O O

In NN O O
this NN O O
paper NN O O
we NN O O
will NN O O
examine NN O O
seven NN O O
different NN O O
data NN O I-Method
representations NN O I-Method
for NN O O
the NN O O
problem NN O O
of NN O O
recognizing NN O I-Task
noun NN O I-Task
phrase NN O I-Task
chunks NN O I-Task
. NN O O

We NN O O
will NN O O
show NN O O
that NN O O
the NN O O
data NN O I-Method
representation NN O I-Method
choice NN O O
has NN O O
a NN O O
minor NN O O
influence NN O O
on NN O O
chunking NN O I-Task
performance NN O O
. NN O O

However NN O O
, NN O O
equipped NN O O
with NN O O
the NN O O
most NN O O
suitable NN O O
data NN O I-Method
representation NN O I-Method
, NN O O
our NN O O
memory-based NN O I-Method
learning NN O I-Method
chunker NN O I-Method
was NN O O
able NN O O
to NN O O
improve NN O O
the NN O O
best NN O O
published NN O O
chunking NN O I-Task
results NN O O
for NN O O
a NN O O
standard NN O O
data NN O I-Material
set NN O I-Material
. NN O O

-DOCSTART- NN O O

In NN O O
this NN O O
paper NN O O
we NN O O
describe NN O O
and NN O O
evaluate NN O O
a NN O O
Question NN O I-Method
Answering NN O I-Method
system NN O I-Method
that NN O O
goes NN O O
beyond NN O O
answering NN O O
factoid NN O O
questions NN O O
. NN O O

We NN O O
focus NN O O
on NN O O
FAQ-like NN O I-Material
questions NN O I-Material
and NN O I-Material
answers NN O I-Material
, NN O O
and NN O O
build NN O O
our NN O O
system NN O I-Generic
around NN O O
a NN O O
noisy-channel NN O I-Method
architecture NN O I-Method
which NN O O
exploits NN O O
both NN O O
a NN O O
language NN O I-Method
model NN O I-Method
for NN O O
answers NN O O
and NN O O
a NN O O
transformation NN O I-Method
model NN O I-Method
for NN O O
answer/question NN O O
terms NN O O
, NN O O
trained NN O O
on NN O O
a NN O O
corpus NN O O
of NN O O
1 NN O O
million NN O O
question/answer NN O O
pairs NN O O
collected NN O O
from NN O O
the NN O O
Web NN O I-Material
. NN O O

-DOCSTART- NN O O

In NN O O
this NN O O
paper NN O O
we NN O O
evaluate NN O O
four NN O O
objective NN O O
measures NN O I-Metric
of NN O I-Metric
speech NN O I-Metric
with NN O O
regards NN O O
to NN O O
intelligibility NN O I-Task
prediction NN O I-Task
of NN O O
synthesized NN O I-Material
speech NN O I-Material
in NN O O
diverse NN O I-OtherScientificTerm
noisy NN O I-OtherScientificTerm
situations NN O I-OtherScientificTerm
. NN O O

We NN O O
evaluated NN O O
three NN O O
intel-ligibility NN O I-Metric
measures NN O I-Metric
, NN O O
the NN O O
Dau NN O I-Metric
measure NN O I-Metric
, NN O O
the NN O O
glimpse NN O I-Metric
proportion NN O I-Metric
and NN O O
the NN O O
Speech NN O I-Metric
Intelligibility NN O I-Metric
Index NN O I-Metric
-LRB- NN O I-Metric
SII NN O I-Metric
-RRB- NN O I-Metric
and NN O O
a NN O O
quality NN O I-Metric
measure NN O I-Metric
, NN O O
the NN O O
Perceptual NN O I-Metric
Evaluation NN O I-Metric
of NN O I-Metric
Speech NN O I-Metric
Quality NN O I-Metric
-LRB- NN O I-Metric
PESQ NN O I-Metric
-RRB- NN O I-Metric
. NN O O

For NN O O
the NN O O
generation NN O I-Task
of NN O I-Task
synthesized NN O I-Task
speech NN O I-Task
we NN O O
used NN O O
a NN O O
state NN O O
of NN O O
the NN O O
art NN O O
HMM-based NN O I-Method
speech NN O I-Method
synthesis NN O I-Method
system NN O I-Method
. NN O O

The NN O O
noisy NN O I-OtherScientificTerm
conditions NN O I-OtherScientificTerm
comprised NN O O
four NN O O
additive NN O I-OtherScientificTerm
noises NN O I-OtherScientificTerm
. NN O O

The NN O O
measures NN O I-Generic
were NN O O
compared NN O O
with NN O O
subjective NN O I-Metric
intelligibility NN O I-Metric
scores NN O I-Metric
obtained NN O O
in NN O O
listening NN O O
tests NN O O
. NN O O

The NN O O
results NN O O
show NN O O
the NN O O
Dau NN O I-Metric
and NN O O
the NN O O
glimpse NN O I-Metric
measures NN O I-Metric
to NN O O
be NN O O
the NN O O
best NN O O
predictors NN O I-Method
of NN O I-Method
intelligibility NN O I-Method
, NN O O
with NN O O
correlations NN O I-Metric
of NN O O
around NN O O
0.83 NN O O
to NN O O
subjective NN O I-OtherScientificTerm
scores NN O I-OtherScientificTerm
. NN O O

All NN O O
measures NN O I-Generic
gave NN O O
less NN O O
accurate NN O O
predictions NN O I-Task
of NN O I-Task
intelligibility NN O I-Task
for NN O O
synthetic NN O I-Material
speech NN O I-Material
than NN O O
have NN O O
previously NN O O
been NN O O
found NN O O
for NN O O
natural NN O I-Material
speech NN O I-Material
; NN O O
in NN O O
particular NN O O
the NN O O
SII NN O I-Metric
measure NN O I-Metric
. NN O O

In NN O O
additional NN O O
experiments NN O O
, NN O O
we NN O O
processed NN O O
the NN O O
synthesized NN O I-Material
speech NN O I-Material
by NN O O
an NN O O
ideal NN O I-Method
binary NN O I-Method
mask NN O I-Method
before NN O O
adding NN O O
noise NN O O
. NN O O

The NN O O
Glimpse NN O I-Metric
measure NN O I-Metric
gave NN O O
the NN O O
most NN O O
accurate NN O O
intelligibility NN O I-Task
predictions NN O I-Task
in NN O O
this NN O O
situation NN O O
. NN O O

-DOCSTART- NN O O

A NN O O
'' NN O I-Method
graphics NN O I-Method
for NN O I-Method
vision NN O I-Method
'' NN O I-Method
approach NN O I-Method
is NN O O
proposed NN O O
to NN O O
address NN O O
the NN O O
problem NN O O
of NN O O
reconstruction NN O I-Task
from NN O O
a NN O O
large NN O I-Material
and NN O I-Material
imperfect NN O I-Material
data NN O I-Material
set NN O I-Material
: NN O O
reconstruction NN O I-Task
on NN O O
demand NN O O
by NN O O
tensor NN O I-Method
voting NN O I-Method
, NN O O
or NN O O
ROD-TV NN O I-Method
. NN O O

ROD-TV NN O I-Method
simultaneously NN O O
delivers NN O O
good NN O O
efficiency NN O I-Metric
and NN O O
robust-ness NN O I-Metric
, NN O O
by NN O O
adapting NN O O
to NN O O
a NN O O
continuum NN O O
of NN O O
primitive NN O I-OtherScientificTerm
connectivity NN O I-OtherScientificTerm
, NN O O
view NN O I-OtherScientificTerm
dependence NN O I-OtherScientificTerm
, NN O O
and NN O O
levels NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
detail NN O I-OtherScientificTerm
-LRB- NN O I-OtherScientificTerm
LOD NN O I-OtherScientificTerm
-RRB- NN O I-OtherScientificTerm
. NN O O

Locally NN O I-OtherScientificTerm
inferred NN O I-OtherScientificTerm
surface NN O I-OtherScientificTerm
elements NN O I-OtherScientificTerm
are NN O O
robust NN O O
to NN O O
noise NN O I-OtherScientificTerm
and NN O O
better NN O O
capture NN O O
local NN O I-OtherScientificTerm
shapes NN O I-OtherScientificTerm
. NN O O

By NN O O
inferring NN O O
per-vertex NN O I-OtherScientificTerm
normals NN O I-OtherScientificTerm
at NN O O
sub-voxel NN O I-Metric
precision NN O I-Metric
on NN O O
the NN O O
fly NN O O
, NN O O
we NN O O
can NN O O
achieve NN O O
interpolative NN O I-Task
shading NN O I-Task
. NN O O

Since NN O O
these NN O O
missing NN O O
details NN O O
can NN O O
be NN O O
recovered NN O O
at NN O O
the NN O O
current NN O O
level NN O O
of NN O O
detail NN O O
, NN O O
our NN O O
result NN O O
is NN O O
not NN O O
upper NN O O
bounded NN O O
by NN O O
the NN O O
scanning NN O I-OtherScientificTerm
resolution NN O I-OtherScientificTerm
. NN O O

By NN O O
relaxing NN O O
the NN O O
mesh NN O I-OtherScientificTerm
connectivity NN O I-OtherScientificTerm
requirement NN O I-OtherScientificTerm
, NN O O
we NN O O
extend NN O O
ROD-TV NN O I-Method
and NN O O
propose NN O O
a NN O O
simple NN O O
but NN O O
effective NN O O
multiscale NN O I-Method
feature NN O I-Method
extraction NN O I-Method
algorithm NN O I-Method
. NN O O

ROD-TV NN O I-Method
consists NN O O
of NN O O
a NN O O
hierarchical NN O I-Method
data NN O I-Method
structure NN O I-Method
that NN O O
encodes NN O O
different NN O O
levels NN O O
of NN O O
detail NN O O
. NN O O

The NN O O
local NN O I-Method
reconstruction NN O I-Method
algorithm NN O I-Method
is NN O O
tensor NN O I-Method
voting NN O I-Method
. NN O O

It NN O I-Generic
is NN O O
applied NN O O
on NN O O
demand NN O O
to NN O O
the NN O O
visible NN O O
subset NN O O
of NN O O
data NN O O
at NN O O
a NN O O
desired NN O O
level NN O O
of NN O O
detail NN O O
, NN O O
by NN O O
traversing NN O I-Method
the NN O I-Method
data NN O I-Method
hierarchy NN O I-Method
and NN O O
collecting NN O I-Method
tensorial NN O I-Method
support NN O I-Method
in NN O O
a NN O O
neighborhood NN O O
. NN O O

We NN O O
compare NN O O
our NN O O
approach NN O I-Generic
and NN O O
present NN O O
encouraging NN O O
results NN O O
. NN O O

-DOCSTART- NN O O

Both NN O O
rhetorical NN O I-OtherScientificTerm
structure NN O I-OtherScientificTerm
and NN O O
punctuation NN O I-OtherScientificTerm
have NN O O
been NN O O
helpful NN O O
in NN O O
discourse NN O I-Task
processing NN O I-Task
. NN O O

Based NN O O
on NN O O
a NN O O
corpus NN O I-Task
annotation NN O I-Task
project NN O I-Task
, NN O O
this NN O O
paper NN O O
reports NN O O
the NN O O
discursive NN O I-Task
usage NN O I-Task
of NN O I-Task
6 NN O I-Task
Chinese NN O I-OtherScientificTerm
punctuation NN O I-OtherScientificTerm
marks NN O I-OtherScientificTerm
in NN O O
news NN O I-Material
commentary NN O I-Material
texts NN O I-Material
: NN O O
Colon NN O I-OtherScientificTerm
, NN O O
Dash NN O I-OtherScientificTerm
, NN O O
Ellipsis NN O I-OtherScientificTerm
, NN O O
Exclamation NN O I-OtherScientificTerm
Mark NN O I-OtherScientificTerm
, NN O O
Question NN O I-OtherScientificTerm
Mark NN O I-OtherScientificTerm
, NN O O
and NN O O
Semicolon NN O I-OtherScientificTerm
. NN O O

The NN O O
rhetorical NN O I-OtherScientificTerm
patterns NN O I-OtherScientificTerm
of NN O O
these NN O O
marks NN O I-Generic
are NN O O
compared NN O O
against NN O O
patterns NN O I-OtherScientificTerm
around NN O I-OtherScientificTerm
cue NN O I-OtherScientificTerm
phrases NN O I-OtherScientificTerm
in NN O O
general NN O O
. NN O O

Results NN O O
show NN O O
that NN O O
these NN O O
Chinese NN O I-OtherScientificTerm
punctuation NN O I-OtherScientificTerm
marks NN O I-OtherScientificTerm
, NN O O
though NN O O
fewer NN O O
in NN O O
number NN O O
than NN O O
cue NN O I-OtherScientificTerm
phrases NN O I-OtherScientificTerm
, NN O O
are NN O O
easy NN O O
to NN O O
identify NN O O
, NN O O
have NN O O
strong NN O O
correlation NN O O
with NN O O
certain NN O O
relations NN O O
, NN O O
and NN O O
can NN O O
be NN O O
used NN O O
as NN O O
distinctive NN O O
indicators NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
nuclearity NN O I-OtherScientificTerm
in NN O O
Chinese NN O I-Material
texts NN O I-Material
. NN O O

-DOCSTART- NN O O

The NN O O
features NN O I-OtherScientificTerm
based NN O O
on NN O O
Markov NN O I-Method
random NN O I-Method
field NN O I-Method
-LRB- NN O I-Method
MRF NN O I-Method
-RRB- NN O I-Method
models NN O I-Method
are NN O O
usually NN O O
sensitive NN O O
to NN O O
the NN O O
rotation NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
image NN O I-OtherScientificTerm
textures NN O I-OtherScientificTerm
. NN O O

This NN O O
paper NN O O
develops NN O O
an NN O O
anisotropic NN O I-Method
circular NN O I-Method
Gaussian NN O I-Method
MRF NN O I-Method
-LRB- NN O I-Method
ACGMRF NN O I-Method
-RRB- NN O I-Method
model NN O I-Method
for NN O O
modelling NN O I-Task
rotated NN O I-Task
image NN O I-Task
textures NN O I-Task
and NN O O
retrieving NN O I-Task
rotation-invariant NN O I-Task
texture NN O I-Task
features NN O I-Task
. NN O O

To NN O O
overcome NN O O
the NN O O
singularity NN O I-OtherScientificTerm
problem NN O I-OtherScientificTerm
of NN O O
the NN O O
least NN O I-Method
squares NN O I-Method
estimate NN O I-Method
-LRB- NN O I-Method
LSE NN O I-Method
-RRB- NN O I-Method
method NN O I-Method
, NN O O
an NN O O
approximate NN O I-Method
least NN O I-Method
squares NN O I-Method
estimate NN O I-Method
-LRB- NN O I-Method
ALSE NN O I-Method
-RRB- NN O I-Method
method NN O I-Method
is NN O O
proposed NN O O
to NN O O
estimate NN O O
the NN O O
parameters NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
the NN O I-OtherScientificTerm
ACGMRF NN O I-Method
model NN O I-Method
. NN O O

The NN O O
rotation-invariant NN O I-OtherScientificTerm
features NN O I-OtherScientificTerm
can NN O O
be NN O O
obtained NN O O
from NN O O
the NN O O
parameters NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
the NN O I-OtherScientificTerm
ACGMRF NN O I-Method
model NN O I-Method
by NN O O
the NN O O
one-dimensional NN O I-Method
-LRB- NN O I-Method
1-D NN O I-Method
-RRB- NN O I-Method
discrete NN O I-Method
Fourier NN O I-Method
transform NN O I-Method
-LRB- NN O I-Method
DFT NN O I-Method
-RRB- NN O I-Method
. NN O O

Significantly NN O O
improved NN O O
accuracy NN O I-Metric
can NN O O
be NN O O
achieved NN O O
by NN O O
applying NN O O
the NN O O
rotation-invariant NN O I-OtherScientificTerm
features NN O I-OtherScientificTerm
to NN O O
classify NN O O
SAR NN O I-OtherScientificTerm
-LRB- NN O I-OtherScientificTerm
synthetic NN O I-OtherScientificTerm
aperture NN O I-OtherScientificTerm
radar NN O I-OtherScientificTerm
-RRB- NN O O
sea NN O O
ice NN O O
and NN O O
Brodatz NN O O
imagery NN O O
. NN O O

-DOCSTART- NN O O

Despite NN O O
much NN O O
recent NN O O
progress NN O O
on NN O O
accurate NN O O
semantic NN O I-Task
role NN O I-Task
labeling NN O I-Task
, NN O O
previous NN O O
work NN O O
has NN O O
largely NN O O
used NN O O
independent NN O I-Method
classifiers NN O I-Method
, NN O O
possibly NN O O
combined NN O O
with NN O O
separate NN O O
label NN O I-Method
sequence NN O I-Method
models NN O I-Method
via NN O O
Viterbi NN O I-Method
decoding NN O I-Method
. NN O O

This NN O O
stands NN O O
in NN O O
stark NN O O
contrast NN O O
to NN O O
the NN O O
linguistic NN O O
observation NN O O
that NN O O
a NN O O
core NN O I-OtherScientificTerm
argument NN O I-OtherScientificTerm
frame NN O I-OtherScientificTerm
is NN O O
a NN O O
joint NN O O
structure NN O O
, NN O O
with NN O O
strong NN O O
dependencies NN O O
between NN O O
arguments NN O O
. NN O O

We NN O O
show NN O O
how NN O O
to NN O O
build NN O O
a NN O O
joint NN O I-Method
model NN O I-Method
of NN O I-Method
argument NN O I-Method
frames NN O I-Method
, NN O O
incorporating NN O O
novel NN O O
features NN O I-OtherScientificTerm
that NN O O
model NN O O
these NN O O
interactions NN O O
into NN O O
discriminative NN O I-Method
log-linear NN O I-Method
models NN O I-Method
. NN O O

This NN O O
system NN O I-Generic
achieves NN O O
an NN O O
error NN O I-Metric
reduction NN O I-Metric
of NN O O
22 NN O O
% NN O O
on NN O O
all NN O O
arguments NN O O
and NN O O
32 NN O O
% NN O O
on NN O O
core NN O O
arguments NN O O
over NN O O
a NN O O
state-of-the NN O O
art NN O O
independent NN O I-Method
classifier NN O I-Method
for NN O O
gold-standard NN O I-OtherScientificTerm
parse NN O I-OtherScientificTerm
trees NN O I-OtherScientificTerm
on NN O O
PropBank NN O I-Material
. NN O O

-DOCSTART- NN O O

One NN O O
of NN O O
the NN O O
major NN O O
problems NN O O
one NN O O
is NN O O
faced NN O O
with NN O O
when NN O O
decomposing NN O O
words NN O O
into NN O O
their NN O O
constituent NN O O
parts NN O O
is NN O O
ambiguity NN O I-OtherScientificTerm
: NN O O
the NN O O
generation NN O I-Task
of NN O O
multiple NN O O
analyses NN O O
for NN O O
one NN O O
input NN O O
word NN O O
, NN O O
many NN O O
of NN O O
which NN O O
are NN O O
implausible NN O O
. NN O O

In NN O O
order NN O O
to NN O O
deal NN O O
with NN O O
ambiguity NN O I-OtherScientificTerm
, NN O O
the NN O O
MORphological NN O I-Method
PArser NN O I-Method
MORPA NN O I-Method
is NN O O
provided NN O O
with NN O O
a NN O O
probabilistic NN O I-Method
context-free NN O I-Method
grammar NN O I-Method
-LRB- NN O I-Method
PCFG NN O I-Method
-RRB- NN O I-Method
, NN O O
i.e. NN O O
it NN O I-Generic
combines NN O O
a NN O O
`` NN O I-Method
conventional NN O I-Method
'' NN O I-Method
context-free NN O I-Method
morphological NN O I-Method
grammar NN O I-Method
to NN O O
filter NN O O
out NN O O
ungrammatical NN O I-OtherScientificTerm
segmentations NN O I-OtherScientificTerm
with NN O O
a NN O O
probability-based NN O I-Method
scoring NN O I-Method
function NN O I-Method
which NN O O
determines NN O O
the NN O O
likelihood NN O O
of NN O O
each NN O O
successful NN O O
parse NN O I-Task
. NN O O

Consequently NN O O
, NN O O
remaining NN O O
analyses NN O O
can NN O O
be NN O O
ordered NN O O
along NN O O
a NN O O
scale NN O O
of NN O O
plausibility NN O O
. NN O O

Test NN O O
performance NN O O
data NN O O
will NN O O
show NN O O
that NN O O
a NN O O
PCFG NN O I-Method
yields NN O O
good NN O O
results NN O O
in NN O O
morphological NN O I-Task
parsing NN O I-Task
. NN O O

MORPA NN O I-Method
is NN O O
a NN O O
fully NN O O
implemented NN O O
parser NN O I-Method
developed NN O O
for NN O O
use NN O O
in NN O O
a NN O O
text-to-speech NN O I-Task
conversion NN O I-Task
system NN O I-Task
. NN O O

-DOCSTART- NN O O

This NN O O
paper NN O O
describes NN O O
the NN O O
framework NN O O
of NN O O
a NN O O
Korean NN O I-Task
phonological NN O I-Task
knowledge NN O I-Task
base NN O I-Task
system NN O I-Task
using NN O O
the NN O O
unification-based NN O I-Method
grammar NN O I-Method
formalism NN O I-Method
: NN O O
Korean NN O I-Method
Phonology NN O I-Method
Structure NN O I-Method
Grammar NN O I-Method
-LRB- NN O I-Method
KPSG NN O I-Method
-RRB- NN O I-Method
. NN O O

The NN O O
approach NN O I-Generic
of NN O O
KPSG NN O I-Method
provides NN O O
an NN O O
explicit NN O O
development NN O O
model NN O O
for NN O O
constructing NN O O
a NN O O
computational NN O O
phonological NN O I-Task
system NN O I-Task
: NN O O
speech NN O I-Task
recognition NN O I-Task
and NN O I-Task
synthesis NN O I-Task
system NN O I-Task
. NN O O

We NN O O
show NN O O
that NN O O
the NN O O
proposed NN O O
approach NN O I-Generic
is NN O O
more NN O O
describable NN O O
than NN O O
other NN O O
approaches NN O I-Generic
such NN O O
as NN O O
those NN O I-Generic
employing NN O O
a NN O O
traditional NN O O
generative NN O I-Method
phonological NN O I-Method
approach NN O I-Method
. NN O O

-DOCSTART- NN O O

In NN O O
some NN O O
auction NN O I-Task
domains NN O I-Task
, NN O O
there NN O O
is NN O O
uncertainty NN O O
regarding NN O O
the NN O O
final NN O O
availability NN O O
of NN O O
the NN O O
goods NN O O
being NN O O
auctioned NN O O
off NN O O
. NN O O

For NN O O
example NN O O
, NN O O
a NN O O
government NN O O
may NN O O
auction NN O O
off NN O O
spectrum NN O O
from NN O O
its NN O O
public NN O O
safety NN O O
network NN O O
, NN O O
but NN O O
it NN O O
may NN O O
need NN O O
this NN O O
spectrum NN O O
back NN O O
in NN O O
times NN O O
of NN O O
emergency NN O O
. NN O O

In NN O O
such NN O O
a NN O O
domain NN O I-Generic
, NN O O
standard NN O O
combinatorial NN O I-Method
auctions NN O I-Method
perform NN O O
poorly NN O O
because NN O O
they NN O I-Generic
lead NN O O
to NN O O
violations NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
individual NN O I-OtherScientificTerm
rationality NN O I-OtherScientificTerm
-LRB- NN O I-OtherScientificTerm
IR NN O I-OtherScientificTerm
-RRB- NN O I-OtherScientificTerm
, NN O O
even NN O O
in NN O O
expectation NN O O
, NN O O
and NN O O
to NN O O
very NN O O
low NN O O
efficiency NN O O
. NN O O

In NN O O
this NN O O
paper NN O O
, NN O O
we NN O O
study NN O O
the NN O O
design NN O I-Task
of NN O I-Task
core-selecting NN O I-Task
payment NN O I-Task
rules NN O I-Task
for NN O O
such NN O O
domains NN O I-Generic
. NN O O

Surprisingly NN O O
, NN O O
we NN O O
show NN O O
that NN O O
in NN O O
this NN O O
new NN O O
domain NN O I-Generic
, NN O O
there NN O O
does NN O O
not NN O O
exist NN O O
a NN O O
payment NN O I-OtherScientificTerm
rule NN O I-OtherScientificTerm
with NN O O
is NN O O
guaranteed NN O O
to NN O O
be NN O O
ex-post NN O O
core-selecting NN O O
. NN O O

However NN O O
, NN O O
we NN O O
show NN O O
that NN O O
by NN O O
designing NN O O
rules NN O I-OtherScientificTerm
that NN O O
are NN O O
'' NN O O
execution-contingent NN O O
, NN O O
'' NN O O
i.e. NN O O
, NN O O
by NN O O
charging NN O O
payments NN O O
that NN O O
are NN O O
conditioned NN O O
on NN O O
the NN O O
realization NN O O
of NN O O
the NN O O
availability NN O O
of NN O O
the NN O O
goods NN O O
, NN O O
we NN O O
can NN O O
reduce NN O O
IR NN O I-OtherScientificTerm
violations NN O I-OtherScientificTerm
. NN O O

We NN O O
design NN O O
two NN O O
core-selecting NN O I-OtherScientificTerm
rules NN O I-OtherScientificTerm
that NN O O
always NN O O
satisfy NN O O
IR NN O I-OtherScientificTerm
in NN O O
expectation NN O O
. NN O O

To NN O O
study NN O O
the NN O O
performance NN O O
of NN O O
our NN O O
rules NN O I-Generic
we NN O O
perform NN O O
a NN O O
computational NN O I-Method
Bayes-Nash NN O I-Method
equilibrium NN O I-Method
analysis NN O I-Method
. NN O O

We NN O O
show NN O O
that NN O O
, NN O O
in NN O O
equilibrium NN O O
, NN O O
our NN O O
new NN O O
rules NN O I-Generic
have NN O O
better NN O O
incentives NN O O
, NN O O
higher NN O O
efficiency NN O O
, NN O O
and NN O O
a NN O O
lower NN O O
rate NN O I-Metric
of NN O I-Metric
ex-post NN O I-Metric
IR NN O I-Metric
violations NN O I-Metric
than NN O O
standard NN O O
core-selecting NN O I-OtherScientificTerm
rules NN O I-OtherScientificTerm
. NN O O

-DOCSTART- NN O O

In NN O O
this NN O O
paper NN O O
, NN O O
we NN O O
will NN O O
describe NN O O
a NN O O
search NN O I-Method
tool NN O I-Method
for NN O O
a NN O O
huge NN O O
set NN O O
of NN O O
ngrams NN O I-OtherScientificTerm
. NN O O

The NN O O
tool NN O I-Generic
supports NN O O
queries NN O O
with NN O O
an NN O O
arbitrary NN O O
number NN O O
of NN O O
wildcards NN O O
. NN O O

It NN O I-Generic
takes NN O O
a NN O O
fraction NN O O
of NN O O
a NN O O
second NN O O
for NN O O
a NN O O
search NN O O
, NN O O
and NN O O
can NN O O
provide NN O O
the NN O O
fillers NN O O
of NN O O
the NN O O
wildcards NN O O
. NN O O

The NN O O
system NN O I-Generic
runs NN O O
on NN O O
a NN O O
single NN O O
Linux NN O O
PC NN O O
with NN O O
reasonable NN O O
size NN O O
memory NN O I-OtherScientificTerm
-LRB- NN O O
less NN O O
than NN O O
4GB NN O O
-RRB- NN O O
and NN O O
disk NN O I-OtherScientificTerm
space NN O I-OtherScientificTerm
-LRB- NN O O
less NN O O
than NN O O
400GB NN O O
-RRB- NN O O
. NN O O

This NN O O
system NN O I-Generic
can NN O O
be NN O O
a NN O O
very NN O O
useful NN O O
tool NN O I-Generic
for NN O O
linguistic NN O I-Task
knowledge NN O I-Task
discovery NN O I-Task
and NN O O
other NN O O
NLP NN O I-Task
tasks NN O I-Task
. NN O O

-DOCSTART- NN O O

For NN O O
intelligent NN O I-Method
interactive NN O I-Method
systems NN O I-Method
to NN O O
communicate NN O O
with NN O O
humans NN O O
in NN O O
a NN O O
natural NN O O
manner NN O O
, NN O O
they NN O I-Generic
must NN O O
have NN O O
knowledge NN O O
about NN O O
the NN O O
system NN O O
users NN O O
. NN O O

This NN O O
paper NN O O
explores NN O O
the NN O O
role NN O O
of NN O O
user NN O I-Task
modeling NN O I-Task
in NN O O
such NN O O
systems NN O I-Method
. NN O O

It NN O O
begins NN O O
with NN O O
a NN O O
characterization NN O O
of NN O O
what NN O O
a NN O O
user NN O I-Method
model NN O I-Method
is NN O O
and NN O O
how NN O O
it NN O I-Generic
can NN O O
be NN O O
used NN O O
. NN O O

The NN O O
types NN O O
of NN O O
information NN O O
that NN O O
a NN O O
user NN O I-Method
model NN O I-Method
may NN O O
be NN O O
required NN O O
to NN O O
keep NN O O
about NN O O
a NN O O
user NN O O
are NN O O
then NN O O
identified NN O O
and NN O O
discussed NN O O
. NN O O

User NN O I-Method
models NN O I-Method
themselves NN O O
can NN O O
vary NN O O
greatly NN O O
depending NN O O
on NN O O
the NN O O
requirements NN O O
of NN O O
the NN O O
situation NN O O
and NN O O
the NN O O
implementation NN O O
, NN O O
so NN O O
several NN O O
dimensions NN O O
along NN O O
which NN O O
they NN O I-Generic
can NN O O
be NN O O
classified NN O O
are NN O O
presented NN O O
. NN O O

Since NN O O
acquiring NN O O
the NN O O
knowledge NN O O
for NN O O
a NN O O
user NN O I-Method
model NN O I-Method
is NN O O
a NN O O
fundamental NN O O
problem NN O O
in NN O O
user NN O I-Task
modeling NN O I-Task
, NN O O
a NN O O
section NN O O
is NN O O
devoted NN O O
to NN O O
this NN O O
topic NN O O
. NN O O

Next NN O O
, NN O O
the NN O O
benefits NN O O
and NN O O
costs NN O O
of NN O O
implementing NN O O
a NN O O
user NN O I-Method
modeling NN O I-Method
component NN O I-Method
for NN O O
a NN O O
system NN O I-Generic
are NN O O
weighed NN O O
in NN O O
light NN O O
of NN O O
several NN O O
aspects NN O O
of NN O O
the NN O O
interaction NN O O
requirements NN O O
that NN O O
may NN O O
be NN O O
imposed NN O O
by NN O O
the NN O O
system NN O I-Generic
. NN O O

Finally NN O O
, NN O O
the NN O O
current NN O O
state NN O O
of NN O O
research NN O O
in NN O O
user NN O I-Task
modeling NN O I-Task
is NN O O
summarized NN O O
, NN O O
and NN O O
future NN O O
research NN O O
topics NN O O
that NN O O
must NN O O
be NN O O
addressed NN O O
in NN O O
order NN O O
to NN O O
achieve NN O O
powerful NN O O
, NN O O
general NN O O
user NN O I-Method
modeling NN O I-Method
systems NN O I-Method
are NN O O
assessed NN O O
. NN O O

-DOCSTART- NN O O

Information NN O I-Method
extraction NN O I-Method
techniques NN O I-Method
automatically NN O O
create NN O O
structured NN O I-Material
databases NN O I-Material
from NN O O
unstructured NN O I-Material
data NN O I-Material
sources NN O I-Material
, NN O O
such NN O O
as NN O O
the NN O O
Web NN O I-Material
or NN O O
newswire NN O I-Material
documents NN O I-Material
. NN O O

Despite NN O O
the NN O O
successes NN O O
of NN O O
these NN O O
systems NN O I-Generic
, NN O O
accuracy NN O I-Metric
will NN O O
always NN O O
be NN O O
imperfect NN O O
. NN O O

For NN O O
many NN O O
reasons NN O O
, NN O O
it NN O O
is NN O O
highly NN O O
desirable NN O O
to NN O O
accurately NN O O
estimate NN O O
the NN O O
confidence NN O O
the NN O O
system NN O I-Generic
has NN O O
in NN O O
the NN O O
correctness NN O O
of NN O O
each NN O O
extracted NN O O
field NN O O
. NN O O

The NN O O
information NN O I-Method
extraction NN O I-Method
system NN O I-Method
we NN O O
evaluate NN O O
is NN O O
based NN O O
on NN O O
a NN O O
linear-chain NN O I-Method
conditional NN O I-Method
random NN O I-Method
field NN O I-Method
-LRB- NN O I-Method
CRF NN O I-Method
-RRB- NN O I-Method
, NN O O
a NN O O
probabilistic NN O I-Method
model NN O I-Method
which NN O O
has NN O O
performed NN O O
well NN O O
on NN O O
information NN O I-Task
extraction NN O I-Task
tasks NN O I-Task
because NN O O
of NN O O
its NN O O
ability NN O O
to NN O O
capture NN O O
arbitrary NN O I-OtherScientificTerm
, NN O I-OtherScientificTerm
overlapping NN O I-OtherScientificTerm
features NN O I-OtherScientificTerm
of NN O O
the NN O O
input NN O I-Generic
in NN O O
a NN O O
Markov NN O I-Method
model NN O I-Method
. NN O O

We NN O O
implement NN O O
several NN O O
techniques NN O I-Generic
to NN O O
estimate NN O O
the NN O O
confidence NN O O
of NN O O
both NN O O
extracted NN O I-Generic
fields NN O I-Generic
and NN O O
entire NN O O
multi-field NN O I-Material
records NN O I-Material
, NN O O
obtaining NN O O
an NN O O
average NN O I-Metric
precision NN O I-Metric
of NN O O
98 NN O O
% NN O O
for NN O O
retrieving NN O O
correct NN O O
fields NN O O
and NN O O
87 NN O O
% NN O O
for NN O O
multi-field NN O I-Material
records NN O I-Material
. NN O O

-DOCSTART- NN O O

In NN O O
this NN O O
paper NN O O
, NN O O
we NN O O
use NN O O
the NN O O
information NN O I-OtherScientificTerm
redundancy NN O I-OtherScientificTerm
in NN O I-OtherScientificTerm
multilingual NN O I-OtherScientificTerm
input NN O I-OtherScientificTerm
to NN O O
correct NN O O
errors NN O O
in NN O O
machine NN O I-Task
translation NN O I-Task
and NN O O
thus NN O O
improve NN O O
the NN O O
quality NN O O
of NN O O
multilingual NN O I-Task
summaries NN O I-Task
. NN O O

We NN O O
consider NN O O
the NN O O
case NN O O
of NN O O
multi-document NN O I-Task
summarization NN O I-Task
, NN O O
where NN O O
the NN O O
input NN O O
documents NN O O
are NN O O
in NN O O
Arabic NN O I-Material
, NN O O
and NN O O
the NN O O
output NN O O
summary NN O O
is NN O O
in NN O O
English NN O I-Material
. NN O O

Typically NN O O
, NN O O
information NN O O
that NN O O
makes NN O O
it NN O O
to NN O O
a NN O O
summary NN O O
appears NN O O
in NN O O
many NN O O
different NN O O
lexical-syntactic NN O I-OtherScientificTerm
forms NN O I-OtherScientificTerm
in NN O O
the NN O O
input NN O O
documents NN O O
. NN O O

Further NN O O
, NN O O
the NN O O
use NN O O
of NN O O
multiple NN O O
machine NN O I-Method
translation NN O I-Method
systems NN O I-Method
provides NN O O
yet NN O O
more NN O O
redundancy NN O O
, NN O O
yielding NN O O
different NN O O
ways NN O O
to NN O O
realize NN O O
that NN O O
information NN O O
in NN O O
English NN O I-Material
. NN O O

We NN O O
demonstrate NN O O
how NN O O
errors NN O O
in NN O O
the NN O O
machine NN O I-Task
translations NN O I-Task
of NN O O
the NN O O
input NN O O
Arabic NN O I-Material
documents NN O I-Material
can NN O O
be NN O O
corrected NN O O
by NN O O
identifying NN O O
and NN O O
generating NN O O
from NN O O
such NN O O
redundancy NN O O
, NN O O
focusing NN O O
on NN O O
noun NN O O
phrases NN O O
. NN O O

-DOCSTART- NN O O

In NN O O
this NN O O
paper NN O O
, NN O O
we NN O O
propose NN O O
a NN O O
new NN O O
approach NN O I-Generic
to NN O O
generate NN O O
oriented NN O I-Task
object NN O I-Task
proposals NN O I-Task
-LRB- NN O I-Task
OOPs NN O I-Task
-RRB- NN O I-Task
to NN O O
reduce NN O O
the NN O O
detection NN O I-Metric
error NN O I-Metric
caused NN O O
by NN O O
various NN O O
orientations NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
the NN O I-OtherScientificTerm
object NN O I-OtherScientificTerm
. NN O O

To NN O O
this NN O O
end NN O O
, NN O O
we NN O O
propose NN O O
to NN O O
efficiently NN O O
locate NN O O
object NN O I-OtherScientificTerm
regions NN O I-OtherScientificTerm
according NN O O
to NN O O
pixelwise NN O I-OtherScientificTerm
object NN O I-OtherScientificTerm
probability NN O I-OtherScientificTerm
, NN O O
rather NN O O
than NN O O
measuring NN O O
the NN O O
objectness NN O I-OtherScientificTerm
from NN O O
a NN O O
set NN O O
of NN O O
sampled NN O O
windows NN O O
. NN O O

We NN O O
formulate NN O O
the NN O O
proposal NN O I-Task
generation NN O I-Task
problem NN O I-Task
as NN O O
a NN O O
generative NN O I-Method
proba-bilistic NN O I-Method
model NN O I-Method
such NN O O
that NN O O
object NN O I-OtherScientificTerm
proposals NN O I-OtherScientificTerm
of NN O O
different NN O O
shapes NN O I-OtherScientificTerm
-LRB- NN O O
i.e. NN O O
, NN O O
sizes NN O I-OtherScientificTerm
and NN O O
orientations NN O I-OtherScientificTerm
-RRB- NN O O
can NN O O
be NN O O
produced NN O O
by NN O O
locating NN O O
the NN O O
local NN O I-OtherScientificTerm
maximum NN O I-OtherScientificTerm
likelihoods NN O I-OtherScientificTerm
. NN O O

The NN O O
new NN O O
approach NN O I-Generic
has NN O O
three NN O O
main NN O O
advantages NN O O
. NN O O

First NN O O
, NN O O
it NN O I-Generic
helps NN O O
the NN O O
object NN O I-Method
detector NN O I-Method
handle NN O O
objects NN O O
of NN O O
different NN O O
orientations NN O I-OtherScientificTerm
. NN O O

Second NN O O
, NN O O
as NN O O
the NN O O
shapes NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
the NN O I-OtherScientificTerm
proposals NN O I-OtherScientificTerm
may NN O O
vary NN O O
to NN O O
fit NN O O
the NN O O
objects NN O O
, NN O O
the NN O O
resulting NN O O
proposals NN O O
are NN O O
tighter NN O O
than NN O O
the NN O O
sampling NN O I-OtherScientificTerm
windows NN O I-OtherScientificTerm
with NN O O
fixed NN O O
sizes NN O O
. NN O O

Third NN O O
, NN O O
it NN O I-Generic
avoids NN O O
massive NN O I-Method
window NN O I-Method
sampling NN O I-Method
, NN O O
and NN O O
thereby NN O O
reducing NN O O
the NN O O
number NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
proposals NN O I-OtherScientificTerm
while NN O O
maintaining NN O O
a NN O O
high NN O O
recall NN O I-Metric
. NN O O

Experiments NN O O
on NN O O
the NN O O
PASCAL NN O I-Material
VOC NN O I-Material
2007 NN O I-Material
dataset NN O I-Material
show NN O O
that NN O O
the NN O O
proposed NN O O
OOP NN O I-Method
outperforms NN O O
the NN O O
state-of-the-art NN O I-Generic
fast NN O I-Generic
methods NN O I-Generic
. NN O O

Further NN O O
experiments NN O O
show NN O O
that NN O O
the NN O O
rotation NN O I-OtherScientificTerm
invariant NN O I-OtherScientificTerm
property NN O I-OtherScientificTerm
helps NN O O
a NN O O
class-specific NN O I-Method
object NN O I-Method
detector NN O I-Method
achieve NN O O
better NN O O
performance NN O O
than NN O O
the NN O O
state-of-the-art NN O O
proposal NN O I-Method
generation NN O I-Method
methods NN O I-Method
in NN O O
either NN O O
object NN O I-Material
rotation NN O I-Material
scenarios NN O I-Material
or NN O O
general NN O I-Material
scenarios NN O I-Material
. NN O O

Generating NN O O
OOPs NN O I-Task
is NN O O
very NN O O
fast NN O O
and NN O O
takes NN O O
only NN O O
0.5 NN O O
s NN O O
per NN O O
image NN O O
. NN O O

-DOCSTART- NN O O

This NN O O
paper NN O O
describes NN O O
three NN O O
relatively NN O O
domain-independent NN O I-Generic
capabilities NN O I-Generic
recently NN O O
added NN O O
to NN O O
the NN O O
Paramax NN O I-Method
spoken NN O I-Method
language NN O I-Method
understanding NN O I-Method
system NN O I-Method
: NN O O
non-monotonic NN O I-Task
reasoning NN O I-Task
, NN O O
implicit NN O I-Task
reference NN O I-Task
resolution NN O I-Task
, NN O O
and NN O O
database NN O I-Task
query NN O I-Task
paraphrase NN O I-Task
. NN O O

In NN O O
addition NN O O
, NN O O
we NN O O
discuss NN O O
the NN O O
results NN O O
of NN O O
the NN O O
February NN O I-Material
1992 NN O I-Material
ATIS NN O I-Material
benchmark NN O I-Material
tests NN O I-Material
. NN O O

We NN O O
describe NN O O
a NN O O
variation NN O O
on NN O O
the NN O O
standard NN O O
evaluation NN O O
metric NN O O
which NN O O
provides NN O O
a NN O O
more NN O O
tightly NN O O
controlled NN O O
measure NN O O
of NN O O
progress NN O O
. NN O O

Finally NN O O
, NN O O
we NN O O
briefly NN O O
describe NN O O
an NN O O
experiment NN O O
which NN O O
we NN O O
have NN O O
done NN O O
in NN O O
extending NN O O
the NN O O
n-best NN O I-Method
speech/language NN O I-Method
integration NN O I-Method
architecture NN O I-Method
to NN O O
improving NN O O
OCR NN O I-Metric
accuracy NN O I-Metric
. NN O O

-DOCSTART- NN O O

We NN O O
investigate NN O O
the NN O O
problem NN O O
of NN O O
fine-grained NN O I-Task
sketch-based NN O I-Task
image NN O I-Task
retrieval NN O I-Task
-LRB- NN O I-Task
SBIR NN O I-Task
-RRB- NN O I-Task
, NN O O
where NN O O
free-hand NN O I-OtherScientificTerm
human NN O I-OtherScientificTerm
sketches NN O I-OtherScientificTerm
are NN O O
used NN O O
as NN O O
queries NN O O
to NN O O
perform NN O O
instance-level NN O I-Task
retrieval NN O I-Task
of NN O I-Task
images NN O I-Task
. NN O O

This NN O O
is NN O O
an NN O O
extremely NN O O
challenging NN O O
task NN O O
because NN O O
-LRB- NN O O
i NN O O
-RRB- NN O O
visual NN O I-OtherScientificTerm
comparisons NN O I-OtherScientificTerm
not NN O O
only NN O O
need NN O O
to NN O O
be NN O O
fine-grained NN O O
but NN O O
also NN O O
executed NN O O
cross-domain NN O O
, NN O O
-LRB- NN O O
ii NN O O
-RRB- NN O O
free-hand NN O I-OtherScientificTerm
-LRB- NN O I-OtherScientificTerm
finger NN O I-OtherScientificTerm
-RRB- NN O I-OtherScientificTerm
sketches NN O I-OtherScientificTerm
are NN O O
highly NN O O
abstract NN O O
, NN O O
making NN O O
fine-grained NN O I-Task
matching NN O I-Task
harder NN O O
, NN O O
and NN O O
most NN O O
importantly NN O O
-LRB- NN O O
iii NN O O
-RRB- NN O O
annotated NN O I-Material
cross-domain NN O I-Material
sketch-photo NN O I-Material
datasets NN O I-Material
required NN O O
for NN O O
training NN O O
are NN O O
scarce NN O O
, NN O O
challenging NN O O
many NN O O
state-of-the-art NN O O
machine NN O I-Method
learning NN O I-Method
techniques NN O I-Method
. NN O O

In NN O O
this NN O O
paper NN O O
, NN O O
for NN O O
the NN O O
first NN O O
time NN O O
, NN O O
we NN O O
address NN O O
all NN O O
these NN O O
challenges NN O O
, NN O O
providing NN O O
a NN O O
step NN O O
towards NN O O
the NN O O
capabilities NN O O
that NN O O
would NN O O
underpin NN O O
a NN O O
commercial NN O O
sketch-based NN O I-Task
image NN O I-Task
retrieval NN O I-Task
application NN O I-Task
. NN O O

We NN O O
introduce NN O O
a NN O O
new NN O O
database NN O O
of NN O O
1,432 NN O O
sketch-photo NN O O
pairs NN O O
from NN O O
two NN O O
categories NN O O
with NN O O
32,000 NN O O
fine-grained NN O I-OtherScientificTerm
triplet NN O I-OtherScientificTerm
ranking NN O I-OtherScientificTerm
annotations NN O I-OtherScientificTerm
. NN O O

We NN O O
then NN O O
develop NN O O
a NN O O
deep NN O I-Method
triplet-ranking NN O I-Method
model NN O I-Method
for NN O O
instance-level NN O I-Task
SBIR NN O I-Task
with NN O O
a NN O O
novel NN O O
data NN O I-Method
augmentation NN O I-Method
and NN O O
staged NN O I-Method
pre-training NN O I-Method
strategy NN O I-Method
to NN O O
alleviate NN O O
the NN O O
issue NN O O
of NN O O
insufficient NN O I-Material
fine-grained NN O I-Material
training NN O I-Material
data NN O I-Material
. NN O O

Extensive NN O O
experiments NN O O
are NN O O
carried NN O O
out NN O O
to NN O O
contribute NN O O
a NN O O
variety NN O O
of NN O O
insights NN O O
into NN O O
the NN O O
challenges NN O O
of NN O O
data NN O I-OtherScientificTerm
sufficiency NN O I-OtherScientificTerm
and NN O O
over-fitting NN O I-OtherScientificTerm
avoidance NN O I-OtherScientificTerm
when NN O O
training NN O O
deep NN O I-Method
networks NN O I-Method
for NN O O
fine-grained NN O I-Task
cross-domain NN O I-Task
ranking NN O I-Task
tasks NN O I-Task
. NN O O

-DOCSTART- NN O O

In NN O O
this NN O O
paper NN O O
we NN O O
target NN O O
at NN O O
generating NN O O
generic NN O I-OtherScientificTerm
action NN O I-OtherScientificTerm
proposals NN O I-OtherScientificTerm
in NN O O
unconstrained NN O I-Material
videos NN O I-Material
. NN O O

Each NN O O
action NN O I-OtherScientificTerm
proposal NN O I-OtherScientificTerm
corresponds NN O O
to NN O O
a NN O O
temporal NN O I-OtherScientificTerm
series NN O I-OtherScientificTerm
of NN O I-OtherScientificTerm
spatial NN O I-OtherScientificTerm
bounding NN O I-OtherScientificTerm
boxes NN O I-OtherScientificTerm
, NN O O
i.e. NN O O
, NN O O
a NN O O
spatio-temporal NN O I-OtherScientificTerm
video NN O I-OtherScientificTerm
tube NN O I-OtherScientificTerm
, NN O O
which NN O O
has NN O O
a NN O O
good NN O O
potential NN O O
to NN O O
locate NN O O
one NN O O
human NN O I-OtherScientificTerm
action NN O I-OtherScientificTerm
. NN O O

Assuming NN O O
each NN O O
action NN O O
is NN O O
performed NN O O
by NN O O
a NN O O
human NN O O
with NN O O
meaningful NN O O
motion NN O O
, NN O O
both NN O O
appearance NN O I-OtherScientificTerm
and NN O I-OtherScientificTerm
motion NN O I-OtherScientificTerm
cues NN O I-OtherScientificTerm
are NN O O
utilized NN O O
to NN O O
measure NN O O
the NN O O
ac-tionness NN O I-Metric
of NN O O
the NN O O
video NN O I-Material
tubes NN O I-Material
. NN O O

After NN O O
picking NN O O
those NN O O
spatiotem-poral NN O O
paths NN O O
of NN O O
high NN O O
actionness NN O I-Metric
scores NN O I-Metric
, NN O O
our NN O O
action NN O I-Task
proposal NN O I-Task
generation NN O I-Task
is NN O O
formulated NN O O
as NN O O
a NN O O
maximum NN O I-Task
set NN O I-Task
coverage NN O I-Task
problem NN O I-Task
, NN O O
where NN O O
greedy NN O I-Method
search NN O I-Method
is NN O O
performed NN O O
to NN O O
select NN O O
a NN O O
set NN O O
of NN O O
action NN O I-OtherScientificTerm
proposals NN O I-OtherScientificTerm
that NN O O
can NN O O
maximize NN O O
the NN O O
overall NN O O
actionness NN O I-Metric
score NN O I-Metric
. NN O O

Compared NN O O
with NN O O
existing NN O O
action NN O I-Method
proposal NN O I-Method
approaches NN O I-Method
, NN O O
our NN O O
action NN O I-OtherScientificTerm
proposals NN O I-OtherScientificTerm
do NN O O
not NN O O
rely NN O O
on NN O O
video NN O I-Method
segmentation NN O I-Method
and NN O O
can NN O O
be NN O O
generated NN O O
in NN O O
nearly NN O O
real-time NN O O
. NN O O

Experimental NN O O
results NN O O
on NN O O
two NN O O
challenging NN O O
datasets NN O I-Generic
, NN O O
MSRII NN O I-Material
and NN O O
UCF NN O I-Material
101 NN O I-Material
, NN O O
validate NN O O
the NN O O
superior NN O O
performance NN O O
of NN O O
our NN O O
action NN O I-OtherScientificTerm
proposals NN O I-OtherScientificTerm
as NN O O
well NN O O
as NN O O
competitive NN O O
results NN O O
on NN O O
action NN O I-Task
detection NN O I-Task
and NN O I-Task
search NN O I-Task
. NN O O

-DOCSTART- NN O O

This NN O O
paper NN O O
reports NN O O
recent NN O O
research NN O O
into NN O O
methods NN O I-Generic
for NN O O
creating NN O I-Task
natural NN O I-Task
language NN O I-Task
text NN O I-Task
. NN O O

A NN O O
new NN O O
processing NN O I-Generic
paradigm NN O I-Generic
called NN O O
Fragment-and-Compose NN O I-Method
has NN O O
been NN O O
created NN O O
and NN O O
an NN O O
experimental NN O O
system NN O O
implemented NN O O
in NN O O
it NN O O
. NN O O

The NN O O
knowledge NN O O
to NN O O
be NN O O
expressed NN O O
in NN O O
text NN O O
is NN O O
first NN O O
divided NN O O
into NN O O
small NN O O
propositional NN O O
units NN O O
, NN O O
which NN O O
are NN O O
then NN O O
composed NN O O
into NN O O
appropriate NN O O
combinations NN O O
and NN O O
converted NN O O
into NN O O
text NN O O
. NN O O

KDS NN O I-Method
-LRB- NN O I-Method
Knowledge NN O I-Method
Delivery NN O I-Method
System NN O I-Method
-RRB- NN O I-Method
, NN O O
which NN O O
embodies NN O O
this NN O O
paradigm NN O I-Generic
, NN O O
has NN O O
distinct NN O O
parts NN O O
devoted NN O O
to NN O O
creation NN O O
of NN O O
the NN O O
propositional NN O O
units NN O O
, NN O O
to NN O O
organization NN O O
of NN O O
the NN O O
text NN O O
, NN O O
to NN O O
prevention NN O O
of NN O O
excess NN O O
redundancy NN O O
, NN O O
to NN O O
creation NN O O
of NN O O
combinations NN O O
of NN O O
units NN O O
, NN O O
to NN O O
evaluation NN O O
of NN O O
these NN O O
combinations NN O O
as NN O O
potential NN O O
sentences NN O O
, NN O O
to NN O O
selection NN O O
of NN O O
the NN O O
best NN O O
among NN O O
competing NN O O
combinations NN O O
, NN O O
and NN O O
to NN O O
creation NN O O
of NN O O
the NN O O
final NN O O
text NN O O
. NN O O

The NN O O
Fragment-and-Compose NN O I-Method
paradigm NN O I-Method
and NN O O
the NN O O
computational NN O I-Generic
methods NN O I-Generic
of NN O O
KDS NN O I-Method
are NN O O
described NN O O
. NN O O

-DOCSTART- NN O O

This NN O O
paper NN O O
explores NN O O
the NN O O
issue NN O O
of NN O O
using NN O O
different NN O O
co-occurrence NN O I-OtherScientificTerm
similarities NN O I-OtherScientificTerm
between NN O O
terms NN O O
for NN O O
separating NN O O
query NN O I-Generic
terms NN O I-Generic
that NN O O
are NN O O
useful NN O O
for NN O O
retrieval NN O I-Task
from NN O O
those NN O I-Generic
that NN O O
are NN O O
harmful NN O O
. NN O O

The NN O O
hypothesis NN O O
under NN O O
examination NN O O
is NN O O
that NN O O
useful NN O I-Generic
terms NN O I-Generic
tend NN O O
to NN O O
be NN O O
more NN O O
similar NN O O
to NN O O
each NN O O
other NN O O
than NN O O
to NN O O
other NN O O
query NN O I-Generic
terms NN O I-Generic
. NN O O

Preliminary NN O O
experiments NN O O
with NN O O
similarities NN O I-Generic
computed NN O O
using NN O O
first-order NN O I-OtherScientificTerm
and NN O I-OtherScientificTerm
second-order NN O I-OtherScientificTerm
co-occurrence NN O I-OtherScientificTerm
seem NN O O
to NN O O
confirm NN O O
the NN O O
hypothesis NN O O
. NN O O

Term NN O I-OtherScientificTerm
similarities NN O I-OtherScientificTerm
could NN O O
then NN O O
be NN O O
used NN O O
for NN O O
determining NN O O
which NN O O
query NN O O
terms NN O O
are NN O O
useful NN O O
and NN O O
best NN O O
reflect NN O O
the NN O O
user NN O O
's NN O O
information NN O O
need NN O O
. NN O O

A NN O O
possible NN O O
application NN O O
would NN O O
be NN O O
to NN O O
use NN O O
this NN O O
source NN O O
of NN O O
evidence NN O O
for NN O O
tuning NN O O
the NN O O
weights NN O O
of NN O O
the NN O O
query NN O O
terms NN O O
. NN O O

-DOCSTART- NN O O

We NN O O
propose NN O O
a NN O O
new NN O O
phrase-based NN O I-Method
translation NN O I-Method
model NN O I-Method
and NN O O
decoding NN O I-Method
algorithm NN O I-Method
that NN O O
enables NN O O
us NN O O
to NN O O
evaluate NN O O
and NN O O
compare NN O O
several NN O O
, NN O O
previously NN O O
proposed NN O O
phrase-based NN O I-Method
translation NN O I-Method
models NN O I-Method
. NN O O

Within NN O O
our NN O O
framework NN O I-Generic
, NN O O
we NN O O
carry NN O O
out NN O O
a NN O O
large NN O O
number NN O O
of NN O O
experiments NN O O
to NN O O
understand NN O O
better NN O O
and NN O O
explain NN O O
why NN O O
phrase-based NN O I-Method
models NN O I-Method
outperform NN O O
word-based NN O I-Method
models NN O I-Method
. NN O O

Our NN O O
empirical NN O O
results NN O O
, NN O O
which NN O O
hold NN O O
for NN O O
all NN O O
examined NN O O
language NN O O
pairs NN O O
, NN O O
suggest NN O O
that NN O O
the NN O O
highest NN O O
levels NN O O
of NN O O
performance NN O O
can NN O O
be NN O O
obtained NN O O
through NN O O
relatively NN O O
simple NN O O
means NN O I-Generic
: NN O O
heuristic NN O I-Method
learning NN O I-Method
of NN O I-Method
phrase NN O I-Method
translations NN O I-Method
from NN O O
word-based NN O I-Method
alignments NN O I-Method
and NN O O
lexical NN O I-Method
weighting NN O I-Method
of NN O I-Method
phrase NN O I-Method
translations NN O I-Method
. NN O O

Surprisingly NN O O
, NN O O
learning NN O O
phrases NN O O
longer NN O O
than NN O O
three NN O O
words NN O O
and NN O O
learning NN O O
phrases NN O O
from NN O O
high-accuracy NN O I-Method
word-level NN O I-Method
alignment NN O I-Method
models NN O I-Method
does NN O O
not NN O O
have NN O O
a NN O O
strong NN O O
impact NN O O
on NN O O
performance NN O O
. NN O O

Learning NN O O
only NN O O
syntactically NN O O
motivated NN O O
phrases NN O O
degrades NN O O
the NN O O
performance NN O O
of NN O O
our NN O O
systems NN O I-Generic
. NN O O

-DOCSTART- NN O O

Color NN O O
is NN O O
known NN O O
to NN O O
be NN O O
highly NN O O
discriminative NN O O
for NN O O
many NN O O
object NN O I-Task
recognition NN O I-Task
tasks NN O I-Task
, NN O O
but NN O O
is NN O O
difficult NN O O
to NN O O
infer NN O O
from NN O O
uncontrolled NN O I-Material
images NN O I-Material
in NN O O
which NN O O
the NN O O
illuminant NN O I-OtherScientificTerm
is NN O O
not NN O O
known NN O O
. NN O O

Traditional NN O O
methods NN O I-Generic
for NN O O
color NN O I-Task
constancy NN O I-Task
can NN O O
improve NN O O
surface NN O I-Method
re-flectance NN O I-Method
estimates NN O I-Method
from NN O O
such NN O O
uncalibrated NN O I-Material
images NN O I-Material
, NN O O
but NN O O
their NN O O
output NN O O
depends NN O O
significantly NN O O
on NN O O
the NN O O
background NN O I-OtherScientificTerm
scene NN O I-OtherScientificTerm
. NN O O

In NN O O
many NN O O
recognition NN O I-Task
and NN O I-Task
retrieval NN O I-Task
applications NN O I-Task
, NN O O
we NN O O
have NN O O
access NN O O
to NN O O
image NN O O
sets NN O O
that NN O O
contain NN O O
multiple NN O O
views NN O O
of NN O O
the NN O O
same NN O O
object NN O O
in NN O O
different NN O O
environments NN O O
; NN O O
we NN O O
show NN O O
in NN O O
this NN O O
paper NN O O
that NN O O
correspondences NN O O
between NN O O
these NN O O
images NN O O
provide NN O O
important NN O O
constraints NN O O
that NN O O
can NN O O
improve NN O O
color NN O I-Task
constancy NN O I-Task
. NN O O

We NN O O
introduce NN O O
the NN O O
multi-view NN O I-Task
color NN O I-Task
constancy NN O I-Task
problem NN O I-Task
, NN O O
and NN O O
present NN O O
a NN O O
method NN O I-Generic
to NN O O
recover NN O O
estimates NN O I-Task
of NN O I-Task
underlying NN O I-Task
surface NN O I-Task
re-flectance NN O I-Task
based NN O O
on NN O O
joint NN O O
estimation NN O O
of NN O O
these NN O O
surface NN O I-OtherScientificTerm
properties NN O I-OtherScientificTerm
and NN O O
the NN O O
illuminants NN O I-OtherScientificTerm
present NN O O
in NN O O
multiple NN O O
images NN O O
. NN O O

The NN O O
method NN O I-Generic
can NN O O
exploit NN O O
image NN O I-OtherScientificTerm
correspondences NN O I-OtherScientificTerm
obtained NN O O
by NN O O
various NN O O
alignment NN O I-Method
techniques NN O I-Method
, NN O O
and NN O O
we NN O O
show NN O O
examples NN O O
based NN O O
on NN O O
matching NN O I-OtherScientificTerm
local NN O I-OtherScientificTerm
region NN O I-OtherScientificTerm
features NN O I-OtherScientificTerm
. NN O O

Our NN O O
results NN O O
show NN O O
that NN O O
multi-view NN O I-OtherScientificTerm
constraints NN O I-OtherScientificTerm
can NN O O
significantly NN O O
improve NN O O
estimates NN O I-Task
of NN O I-Task
both NN O I-Task
scene NN O I-Task
illuminants NN O I-Task
and NN O I-Task
object NN O I-Task
color NN O I-Task
-LRB- NN O I-Task
surface NN O I-Task
reflectance NN O I-Task
-RRB- NN O I-Task
when NN O O
compared NN O O
to NN O O
a NN O O
baseline NN O I-Method
single-view NN O I-Method
method NN O I-Method
. NN O O

-DOCSTART- NN O O

We NN O O
describe NN O O
a NN O O
dialogue NN O I-Method
system NN O I-Method
that NN O O
works NN O O
with NN O O
its NN O O
interlocutor NN O O
to NN O O
identify NN O O
objects NN O O
. NN O O

Our NN O O
contributions NN O O
include NN O O
a NN O O
concise NN O I-Method
, NN O I-Method
modular NN O I-Method
architecture NN O I-Method
with NN O O
reversible NN O O
processes NN O O
of NN O O
understanding NN O I-Task
and NN O O
generation NN O I-Task
, NN O O
an NN O O
information-state NN O I-Method
model NN O I-Method
of NN O I-Method
reference NN O I-Method
, NN O O
and NN O O
flexible NN O O
links NN O O
between NN O O
semantics NN O I-OtherScientificTerm
and NN O O
collaborative NN O I-Task
problem NN O I-Task
solving NN O I-Task
. NN O O

