The Distributed Census Problem (DCP) for an observer T can be for- malized as a tuple (Ns; I;Ps;R;W;Ms; T;O), where:

  • Ns is the set of neighborhoods Ns = (1;...;d),
  • I is the set of person identities,
  • O is the organization (constitution), defining the eligibility
  • Ps is a set of peers with identities from set I, each peer P having a different interpretation of O: P(O)
  • R is the set of residence declarations (constituent items)
  • W is the set of witness stances
  • Ms is a model of the relation between the ground truth I* and Ns, I, P, R and W, as believed by the observer T (e.g., a certain belief network) I* (the ground truth), each having an identity from the set I. The problem is to approximate the I* that best explains Ns, I, P, R and W based on the model Ms.