Traditionally, conditional probability is framed as a distribution function given some scalar:
[tex]\Pr(A|B) = \frac{\Pr(A, B)}{\Pr(B)} = \frac{\Pr(B|A)\Pr(A)}{\Pr(B)}[/tex]
However, what if we don’t have a specific value of [tex]B[/tex]? What if instead, we have some distribution over [tex]B[/tex] that is different from the marginal distribution of [tex]B[/tex] when the joint distribution is integrated? More concretely, [...]