Probabilistic target propagation

Bengio and Frasconi [4] propose a probabilistic approach for propagating targets. With $n$ so-called ``state networks'', at a given time, their system can be in one of only $n$ different discrete states. Parameters are adjusted using the expectation-maximization algorithm. But to solve problems that require a significant amount of memory to store contextual information, such systems would require an unacceptable number of states (i.e., state networks).