The probability calculations of quantum mechanics are generally interpreted in the frequentist viewpoint, where small sample sizes are meaningless. With the current experimental ability to manipulate individual particles, there is a need for a Bayesian interpretation, where probabilities for even a single sample are meaningful. Using graphical models, it is possible to create such a formulation that naturally incorporates weak measurements. In addition, these models allow observers with access to different sets of garbled records of the measurement outcomes to make self-consistent predictions.