In this light, I'm not sure that your challenge makes sense. Bayes nets are a way of expressing a properties of probabilistic models, rather than a model unto themselves.

Valid point, but I think in practice it's possible to identify a model as one of some specific family such as "Bayes Net", "Neural Network", "MaxEnt", etc.

Perhaps a particular Bayes net together with a certain choice of conditional probability function for each arc and a certain choice of inference algorithm would constitute a model.

Right, the point is that the challenger can make any reasonable choice for these unspecified components. Ideally someone would say: here is the data set; I'm modeling it using the method described in such-and-such paper; here are some minor revisions to the method of the paper to make it useful in this case; here are the results.