In the HMM formulation where z is hidden state and x is observed
In the forward case, I see it represented as such:
$\alpha_{k}(z_{k})=P(z_{k},x_{1:k})=\sum_{z_{k-1}}P(z_{k},z_{k-1},x_{1:k})$
but in the backward case I see it as such
$\beta_{k}(z_{k})=P(x_{k+1:n}|z_{k})=\sum_{z_{k+1}}P(x_{k+1:n},z_{k+1}|z_{k})$
I thought that in the probability rules, we can factor in a new term based on this expression:
$P(A,B)=\sum_{C}P(A,C|B)$
If so, how are both of those expressions even expanded? Shoudn't the forward rule then be:
$\alpha_{k}(z_{k})=P(z_{k},x_{1:k})=\sum_{z_{k-1}}P(z_{k},z_{k-1}|x_{1:k})$
and the backward rule be not possible to factorize based on the above rule? Is my rule wrong?