<s>
Bayesian	B-General_Concept
programming	I-General_Concept
is	O
a	O
formalism	O
and	O
a	O
methodology	O
for	O
having	O
a	O
technique	O
to	O
specify	O
probabilistic	O
models	O
and	O
solve	O
problems	O
when	O
less	O
than	O
the	O
necessary	O
information	O
is	O
available	O
.	O
</s>
<s>
a	O
physical	O
device	O
,	O
but	O
an	O
inference	O
engine	O
to	O
automate	O
probabilistic	O
reasoning	O
—	O
a	O
kind	O
of	O
Prolog	B-Language
for	O
probability	O
instead	O
of	O
logic	O
.	O
</s>
<s>
Bayesian	B-General_Concept
programming	I-General_Concept
is	O
a	O
formal	O
and	O
concrete	O
implementation	O
of	O
this	O
"	O
robot	O
"	O
.	O
</s>
<s>
Bayesian	B-General_Concept
programming	I-General_Concept
may	O
also	O
be	O
seen	O
as	O
an	O
algebraic	O
formalism	O
to	O
specify	O
graphical	O
models	O
such	O
as	O
,	O
for	O
instance	O
,	O
Bayesian	O
networks	O
,	O
dynamic	O
Bayesian	O
networks	O
,	O
Kalman	O
filters	O
or	O
hidden	O
Markov	O
models	O
.	O
</s>
<s>
Indeed	O
,	O
Bayesian	B-General_Concept
Programming	I-General_Concept
is	O
more	O
general	O
than	O
Bayesian	O
networks	O
and	O
has	O
a	O
power	O
of	O
expression	O
equivalent	O
to	O
probabilistic	O
factor	O
graphs	O
.	O
</s>
<s>
An	O
important	O
feature	O
of	O
Bayesian	B-General_Concept
Programming	I-General_Concept
is	O
this	O
capacity	O
to	O
use	O
questions	O
to	O
other	O
Bayesian	O
programs	O
as	O
components	O
of	O
the	O
definition	O
of	O
a	O
new	O
Bayesian	O
program	O
.	O
</s>
<s>
The	O
purpose	O
of	O
Bayesian	B-Application
spam	I-Application
filtering	I-Application
is	O
to	O
eliminate	O
junk	O
e-mails	O
.	O
</s>
<s>
Using	O
these	O
words	O
without	O
taking	O
the	O
order	O
into	O
account	O
is	O
commonly	O
called	O
a	O
bag	B-General_Concept
of	I-General_Concept
words	I-General_Concept
model	I-General_Concept
.	O
</s>
<s>
This	O
is	O
the	O
naive	B-General_Concept
Bayes	I-General_Concept
assumption	O
and	O
this	O
makes	O
this	O
spam	O
filter	O
a	O
naive	B-General_Concept
Bayes	I-General_Concept
model	I-General_Concept
.	O
</s>
<s>
This	O
kind	O
of	O
assumption	O
is	O
known	O
as	O
the	O
naive	B-General_Concept
Bayes	I-General_Concept
 '	I-General_Concept
assumption	I-General_Concept
.	O
</s>
<s>
Each	O
of	O
the	O
forms	O
may	O
be	O
specified	O
using	O
Laplace	O
rule	O
of	O
succession	O
(	O
this	O
is	O
a	O
pseudocounts-based	O
smoothing	O
technique	O
to	O
counter	O
the	O
zero-frequency	B-Algorithm
problem	I-Algorithm
of	O
words	O
never-seen-before	O
)	O
:	O
</s>
<s>
Both	O
methods	O
could	O
be	O
combined	O
:	O
the	O
system	O
could	O
start	O
with	O
initial	O
standard	O
values	O
of	O
these	O
parameters	O
issued	O
from	O
a	O
generic	O
database	O
,	O
then	O
some	O
incremental	B-Algorithm
learning	I-Algorithm
customizes	O
the	O
classifier	O
to	O
each	O
individual	O
user	O
.	O
</s>
<s>
The	O
Bayesian	B-Application
spam	I-Application
filter	I-Application
program	O
is	O
completely	O
defined	O
by	O
:	O
</s>
<s>
Bayesian	B-Application
filters	I-Application
(	O
often	O
called	O
Recursive	O
Bayesian	O
estimation	O
)	O
are	O
generic	O
probabilistic	O
models	O
for	O
time	O
evolving	O
processes	O
.	O
</s>
<s>
Bayesian	B-Application
filters	I-Application
have	O
a	O
very	O
interesting	O
recursive	O
property	O
,	O
which	O
contributes	O
greatly	O
to	O
their	O
attractiveness	O
.	O
</s>
<s>
Hidden	O
Markov	O
models	O
(	O
HMMs	O
)	O
are	O
another	O
very	O
popular	O
specialization	O
of	O
Bayesian	B-Application
filters	I-Application
.	O
</s>
<s>
called	O
the	O
Viterbi	B-Algorithm
algorithm	I-Algorithm
.	O
</s>
<s>
Since	O
2000	O
,	O
Bayesian	B-General_Concept
programming	I-General_Concept
has	O
been	O
used	O
to	O
develop	O
both	O
robotics	O
applications	O
and	O
life	O
sciences	O
models	O
.	O
</s>
<s>
In	O
robotics	O
,	O
bayesian	B-General_Concept
programming	I-General_Concept
was	O
applied	O
to	O
autonomous	O
robotics	O
,	O
robotic	O
CAD	B-Application
systems	O
,	O
advanced	O
driver-assistance	O
systems	O
,	O
robotic	O
arm	O
control	O
,	O
mobile	O
robotics	O
,	O
human-robot	O
interaction	O
,	O
human-vehicle	O
interaction	O
(	O
Bayesian	O
autonomous	O
driver	O
models	O
)	O
video	O
game	O
avatar	O
programming	O
and	O
training	O
and	O
real-time	O
strategy	O
games	O
(	O
AI	O
)	O
.	O
</s>
<s>
In	O
life	O
sciences	O
,	O
bayesian	B-General_Concept
programming	I-General_Concept
was	O
used	O
in	O
vision	O
to	O
reconstruct	O
shape	O
from	O
motion	O
,	O
to	O
model	O
visuo-vestibular	O
interaction	O
and	O
to	O
study	O
saccadic	O
eye	O
movements	O
;	O
in	O
speech	O
perception	O
and	O
control	O
to	O
study	O
early	O
speech	O
acquisition	O
and	O
the	O
emergence	O
of	O
articulatory-acoustic	O
systems	O
;	O
and	O
to	O
model	O
handwriting	O
perception	O
and	O
control	O
.	O
</s>
<s>
Bayesian	O
program	O
learning	O
has	O
potential	O
applications	O
voice	B-Application
recognition	I-Application
and	O
synthesis	O
,	O
image	O
recognition	O
and	O
natural	O
language	O
processing	O
.	O
</s>
<s>
The	O
comparison	O
between	O
probabilistic	O
approaches	O
(	O
not	O
only	O
bayesian	B-General_Concept
programming	I-General_Concept
)	O
and	O
possibility	O
theories	O
continues	O
to	O
be	O
debated	O
.	O
</s>
<s>
Extended	O
classical	O
programming	O
languages	O
include	O
logical	O
languages	O
as	O
proposed	O
in	O
Probabilistic	B-Application
Horn	I-Application
Abduction	I-Application
,	O
Independent	O
Choice	O
Logic	O
,	O
PRISM	O
,	O
and	O
ProbLog	O
which	O
proposes	O
an	O
extension	O
of	O
Prolog	B-Language
.	O
</s>
<s>
It	O
can	O
also	O
be	O
extensions	O
of	O
functional	B-Language
programming	I-Language
languages	I-Language
(	O
essentially	O
Lisp	B-Language
and	O
Scheme	B-Language
)	O
such	O
as	O
IBAL	O
or	O
CHURCH	O
.	O
</s>
<s>
The	O
purpose	O
of	O
Bayesian	B-General_Concept
programming	I-General_Concept
is	O
different	O
.	O
</s>
<s>
Bayesian	B-General_Concept
programming	I-General_Concept
attempts	O
to	O
replace	O
classical	O
languages	O
with	O
a	O
programming	O
approach	O
based	O
on	O
probability	O
that	O
considers	O
incompleteness	O
and	O
uncertainty	O
.	O
</s>
<s>
The	O
precise	O
comparison	O
between	O
the	O
semantics	B-Application
and	O
power	O
of	O
expression	O
of	O
Bayesian	O
and	O
probabilistic	O
programming	O
is	O
an	O
open	O
question	O
.	O
</s>
