Exact Inference of Hidden Structure from Sample Data in Noisy-OR Networks
This addresses a fundamental challenge in graphical models for researchers and practitioners, offering an exact solution in specific cases, though it is incremental as it applies to restricted settings.
The paper tackles the problem of learning hidden structure from sample data in noisy-OR networks, showing that in restricted settings, it is possible to perfectly reconstruct the hidden structure solely based on observed data.
In the literature on graphical models, there has been increased attention paid to the problems of learning hidden structure (see Heckerman [H96] for survey) and causal mechanisms from sample data [H96, P88, S93, P95, F98]. In most settings we should expect the former to be difficult, and the latter potentially impossible without experimental intervention. In this work, we examine some restricted settings in which perfectly reconstruct the hidden structure solely on the basis of observed sample data.