Event-by-Event Simulation of Quantum Phenomena I just recently stumbled over 
http://rugth30.phys.rug.nl/dlm/
and
http://www.sbfisica.org.br/bjp/files/v38_26.pdf
As the title suggests these are presentations of mathematical models capable of simulating quantum "effects" on an event-by-event basis using "Deterministic Learning Machine" Models.
What is the state of discussion regarding this? Is this bad science? or really accepted alternative way to tackle the pure output side of the experiments?
What does this add to the available interpretations of quantum-phenomena?
 A: First the easy part: this is not bad science and it is a valid explanation of current experiments.  Now, in my personal opinion, the authors did not do the best job presenting these results to the quantum foundations community -- I'm referring to the simulation of Bell inequality violation experiments here.  In the later papers (here is a list) in this series, however, it is explicitly stated which so-called "loopholes" the authors are exploiting.  
So, to be clear, the authors have a valid local hidden variable model of real experiments, not of the hypothetical ideal experiments for which Bell's original theorem applies.  However, other authors have generalized the theorem to apply to more experimentally realistic situations.  For example, the relevant one for these models is the inclusion of the "time-coincidence" loophole by Larsson and Gill (initiated in this paper).
Explicitly constructed models which exploit the loopholes in Bell experiments are not really an active area of research.  My impression is the vast majority of the quantum community believes that there will eventually be a "loophole-free" test which refutes them all.  Moreover, since most hidden variable models which exploit loopholes are contrived to to do just that, they do not give us any deeper insights into nature -- and hence offer no contribution to interpretations of quantum theory.  
