<s>
In	O
artificial	B-Application
intelligence	I-Application
research	I-Application
,	O
commonsense	B-General_Concept
knowledge	I-General_Concept
consists	O
of	O
facts	O
about	O
the	O
everyday	O
world	O
,	O
such	O
as	O
"	O
Lemons	O
are	O
sour	O
"	O
,	O
or	O
"	O
Cows	O
say	O
moo	O
"	O
,	O
that	O
all	O
humans	O
are	O
expected	O
to	O
know	O
.	O
</s>
<s>
The	O
first	O
AI	B-Application
program	O
to	O
address	O
common	B-General_Concept
sense	I-General_Concept
knowledge	I-General_Concept
was	O
Advice	B-Application
Taker	I-Application
in	O
1959	O
by	O
John	O
McCarthy	O
.	O
</s>
<s>
Commonsense	B-General_Concept
knowledge	I-General_Concept
can	O
underpin	O
a	O
commonsense	O
reasoning	O
process	O
,	O
to	O
attempt	O
inferences	O
such	O
as	O
"	O
You	O
might	O
bake	O
a	O
cake	O
because	O
you	O
want	O
people	O
to	O
eat	O
the	O
cake.	O
"	O
</s>
<s>
A	O
natural	B-Language
language	I-Language
processing	I-Language
process	O
can	O
be	O
attached	O
to	O
the	O
commonsense	B-General_Concept
knowledge	I-General_Concept
base	I-General_Concept
to	O
allow	O
the	O
knowledge	O
base	O
to	O
attempt	O
to	O
answer	B-Algorithm
questions	I-Algorithm
about	O
the	O
world	O
.	O
</s>
<s>
Common	B-General_Concept
sense	I-General_Concept
knowledge	I-General_Concept
also	O
helps	O
to	O
solve	O
problems	O
in	O
the	O
face	O
of	O
incomplete	O
information	O
.	O
</s>
<s>
Using	O
widely	O
held	O
beliefs	O
about	O
everyday	O
objects	O
,	O
or	O
common	B-General_Concept
sense	I-General_Concept
knowledge	I-General_Concept
,	O
AI	B-Application
systems	O
make	O
common	O
sense	O
assumptions	O
or	O
default	B-Application
assumptions	I-Application
about	O
the	O
unknown	O
similar	O
to	O
the	O
way	O
people	O
do	O
.	O
</s>
<s>
In	O
an	O
AI	B-Application
system	O
or	O
in	O
English	O
,	O
this	O
is	O
expressed	O
as	O
"	O
Normally	O
P	O
holds	O
"	O
,	O
"	O
Usually	O
P	O
"	O
or	O
"	O
Typically	O
P	O
so	O
Assume	O
P	O
"	O
.	O
</s>
<s>
As	O
more	O
knowledge	O
of	O
the	O
world	O
is	O
discovered	O
or	O
learned	O
over	O
time	O
,	O
the	O
AI	B-Application
system	O
can	O
revise	O
its	O
assumptions	O
about	O
Tweety	O
using	O
a	O
truth	O
maintenance	O
process	O
.	O
</s>
<s>
Commonsense	O
reasoning	O
simulates	O
the	O
human	O
ability	O
to	O
use	O
commonsense	B-General_Concept
knowledge	I-General_Concept
to	O
make	O
presumptions	O
about	O
the	O
type	O
and	O
essence	O
of	O
ordinary	O
situations	O
they	O
encounter	O
every	O
day	O
,	O
and	O
to	O
change	O
their	O
"	O
minds	O
"	O
should	O
new	O
information	O
come	O
to	O
light	O
.	O
</s>
<s>
The	O
ability	O
to	O
explain	O
cause	O
and	O
effect	O
is	O
an	O
important	O
aspect	O
of	O
explainable	O
AI	B-Application
.	O
</s>
<s>
Compared	O
with	O
humans	O
,	O
all	O
existing	O
computer	O
programs	O
that	O
attempt	O
human-level	O
AI	B-Application
perform	O
extremely	O
poorly	O
on	O
modern	O
"	O
commonsense	O
reasoning	O
"	O
benchmark	O
tests	O
such	O
as	O
the	O
Winograd	O
Schema	O
Challenge	O
.	O
</s>
<s>
The	O
problem	O
of	O
attaining	O
human-level	O
competency	O
at	O
"	O
commonsense	B-General_Concept
knowledge	I-General_Concept
"	O
tasks	O
is	O
considered	O
to	O
probably	O
be	O
"	O
AI	B-General_Concept
complete	I-General_Concept
"	O
(	O
that	O
is	O
,	O
solving	O
it	O
would	O
require	O
the	O
ability	O
to	O
synthesize	O
a	O
fully	O
human-level	O
intelligence	O
)	O
,	O
although	O
some	O
oppose	O
this	O
notion	O
and	O
believe	O
compassionate	O
intelligence	O
is	O
also	O
required	O
for	O
human-level	O
AI	B-Application
.	O
</s>
<s>
Common	O
sense	O
reasoning	O
has	O
been	O
applied	O
successfully	O
in	O
more	O
limited	O
domains	O
such	O
as	O
natural	B-Language
language	I-Language
processing	I-Language
and	O
automated	O
diagnosis	O
or	O
analysis	O
.	O
</s>
<s>
Compiling	O
comprehensive	O
knowledge	O
bases	O
of	O
commonsense	O
assertions	O
(	O
CSKBs	O
)	O
is	O
a	O
long-standing	O
challenge	O
in	O
AI	B-Application
research	I-Application
.	O
</s>
<s>
From	O
early	O
expert-driven	O
efforts	O
like	O
CYC	B-Language
and	O
WordNet	O
,	O
significant	O
advances	O
were	O
achieved	O
via	O
the	O
crowdsourced	O
OpenMind	O
Commonsense	O
project	O
,	O
which	O
lead	O
to	O
the	O
crowdsourced	O
ConceptNet	O
KB	O
.	O
</s>
<s>
Challenges	O
also	O
remain	O
on	O
the	O
representation	O
of	O
commonsense	B-General_Concept
knowledge	I-General_Concept
:	O
Most	O
CSKB	O
projects	O
follow	O
a	O
triple	O
data	O
model	O
,	O
which	O
is	O
not	O
necessarily	O
best	O
suited	O
for	O
breaking	O
more	O
complex	O
natural	O
language	O
assertions	O
.	O
</s>
<s>
At	O
Lawrence	O
Livermore	O
National	O
Laboratory	O
,	O
common	B-General_Concept
sense	I-General_Concept
knowledge	I-General_Concept
was	O
used	O
in	O
an	O
intelligent	B-General_Concept
software	I-General_Concept
agent	I-General_Concept
to	O
detect	O
violations	O
of	O
a	O
comprehensive	O
nuclear	O
test	O
ban	O
treaty	O
.	O
</s>
