Without loss of generality, we assume in this proof that all variables
are declared as evidence variables.
To prove this soundness theorem, all we need to show is that each Q-DAG
potential will evaluate to its corresponding probabilistic potential under
all possible evidence. Formally, for any cluster and variables
,
the matrices of which are assigned to
, we need to show that
[IMAGE ]
for a given evidence .
Once we establish this, we are guaranteed that
will
evaluate to the probability
because the application
of
and
in the Q-DAG algorithm is isomorphic to the
application of * and + in the probabilistic algorithm, respectively.
To prove Equation 1, we will extend the Q-DAG node evaluator
to mappings in the standard way. That is, if
is a mapping
from instantiations to Q-DAG nodes, then
is defined as follows:
That is, we simply apply the Q-DAG node evaluator to the range of mapping .
Note that will then be equal to
.
Therefore,
Note also that by definition of , we have
that
equals
. Therefore,
Therefore,