<s>
A	O
confabulation	B-Algorithm
,	O
also	O
known	O
as	O
a	O
false	O
,	O
degraded	O
,	O
or	O
corrupted	O
memory	O
,	O
is	O
a	O
stable	O
pattern	O
of	O
activation	O
in	O
an	O
artificial	B-Architecture
neural	I-Architecture
network	I-Architecture
or	O
neural	O
assembly	O
that	O
does	O
not	O
correspond	O
to	O
any	O
previously	O
learned	O
patterns	O
.	O
</s>
<s>
The	O
same	O
term	O
is	O
also	O
applied	O
to	O
the	O
(	O
nonartificial	O
)	O
neural	O
mistake-making	O
process	O
leading	O
to	O
a	O
false	O
memory	O
(	O
confabulation	B-Algorithm
)	O
.	O
</s>
<s>
In	O
this	O
,	O
confabulations	B-Algorithm
relate	O
to	O
pathologically	O
induced	O
neural	O
activation	O
patterns	O
depart	O
from	O
direct	O
experience	O
and	O
learned	O
relationships	O
.	O
</s>
<s>
In	O
computational	B-Application
modeling	I-Application
of	O
such	O
damage	O
,	O
related	O
brain	O
pathologies	O
such	O
as	O
dyslexia	O
and	O
hallucination	B-General_Concept
result	O
from	O
simulated	O
lesioning	O
and	O
neuron	O
death	O
.	O
</s>
<s>
Forms	O
of	O
confabulation	B-Algorithm
in	O
which	O
missing	O
or	O
incomplete	O
information	O
is	O
incorrectly	O
filled	O
in	O
by	O
the	O
brain	O
are	O
generally	O
modelled	O
by	O
the	O
well	O
known	O
neural	B-Architecture
network	I-Architecture
process	O
called	O
pattern	O
completion	O
.	O
</s>
<s>
Confabulation	B-Algorithm
is	O
central	O
to	O
a	O
theory	O
of	O
cognition	O
and	O
consciousness	O
by	O
S	O
.	O
L	O
.	O
Thaler	O
in	O
which	O
thoughts	O
and	O
ideas	O
originate	O
in	O
both	O
biological	O
and	O
synthetic	O
neural	B-Architecture
networks	I-Architecture
as	O
false	O
or	O
degraded	O
memories	O
nucleate	O
upon	O
various	O
forms	O
of	O
neuronal	O
and	O
synaptic	B-Application
fluctuations	O
and	O
damage	O
.	O
</s>
<s>
Such	O
novel	O
patterns	O
of	O
neural	O
activation	O
are	O
promoted	O
to	O
ideas	O
as	O
other	O
neural	B-Architecture
nets	I-Architecture
perceive	O
utility	O
or	O
value	O
to	O
them	O
(	O
i.e.	O
,	O
the	O
thalamo-cortical	O
loop	O
)	O
.	O
</s>
<s>
The	O
exploitation	O
of	O
these	O
false	O
memories	O
by	O
other	O
artificial	B-Architecture
neural	I-Architecture
networks	I-Architecture
forms	O
the	O
basis	O
of	O
inventive	O
artificial	B-Application
intelligence	I-Application
systems	O
currently	O
utilized	O
in	O
product	O
design	O
,	O
materials	O
discovery	O
and	O
improvisational	O
military	O
robots	O
.	O
</s>
<s>
The	O
concept	O
of	O
such	O
opportunistic	O
confabulation	B-Algorithm
grew	O
out	O
of	O
experiments	O
with	O
artificial	B-Architecture
neural	I-Architecture
networks	I-Architecture
that	O
simulated	O
brain	O
cell	O
apoptosis	O
.	O
</s>
<s>
The	O
term	O
confabulation	B-Algorithm
is	O
also	O
used	O
by	O
Robert	O
Hecht-Nielsen	O
in	O
describing	O
inductive	O
reasoning	O
accomplished	O
via	O
Bayesian	O
networks	O
.	O
</s>
<s>
Confabulation	B-Algorithm
is	O
used	O
to	O
select	O
the	O
expectancy	O
of	O
the	O
concept	O
that	O
follows	O
a	O
particular	O
context	O
.	O
</s>
<s>
However	O
,	O
most	O
events	O
and	O
concepts	O
occur	O
in	O
multiple	O
,	O
conflicting	O
contexts	O
and	O
so	O
confabulation	B-Algorithm
yields	O
a	O
consensus	O
of	O
an	O
expected	O
event	O
that	O
may	O
only	O
be	O
minimally	O
more	O
likely	O
than	O
many	O
other	O
events	O
.	O
</s>
<s>
Confabulation	B-Algorithm
grew	O
out	O
of	O
vector	O
analysis	O
of	O
data	O
retrieval	O
like	O
that	O
of	O
latent	O
semantic	O
analysis	O
and	O
support	B-Algorithm
vector	I-Algorithm
machines	I-Algorithm
.	O
</s>
