Discovering Hidden Variables in Noisy-Or Networks using Quartet Tests

We give a polynomial-time algorithm for provably learning the structure and pa-rameters of bipartite noisy-or Bayesian networks of binary variables where the top layer is completely hidden. Unsupervised learning of these models is a form of discrete factor analysis, enabling the discovery of hidden variables and their causal relationships with observed data. We obtain an efficient learning algorithm for a family of Bayesian networks that we call quartet-learnable. For each latent variable, the existence of a singly-coupled quartet allows us to uniquely identify and learn all parameters involving that latent variable. We give a proof of the poly-nomial sample complexity of our learning algorithm, and experimentally compare it to variational EM.