Stochastic Systems Group
		Home Group Members Publications Research Alumni Mailing List Demos Mission Statement Programs Seminars Library

This talk speculates on connections between two traditionally separated categories of problems, namely those addressed by belief propagation algorithms and those considered in team decision theory. Both types of problems (i) seek to tractably compute summarizing statistics, or estimates, of hidden random variables based on related observables, (ii) are formulated on an underlying graph, and (iii) have natural extensions to distributed sensing applications. Even so, the respective problems are defined by entirely different interpretations of the given graph and, in turn, the respective analyses result in fundamentally different notions of a satisfactory solution. We summarize initial attempts to reconcile these differences, concluding with more new questions than new answers in both problem categories.

Problems with this site should be emailed to jonesb@mit.edu