Probabilistic and combinatorial models for quantized group testing
Wednesday, February 15, 2012 - 11:30am - 12:00pm
We consider a novel group testing framework where defectives obey a probabilistic model in which the number and set of defectives is governed by a graphical model. We furthermore assume that the defectives have importance factors that influence their strength on the test outcomes and the accuracy with which the outcomes are read. This kind of scenario arises, for example, in MAC channels with networked users using different communication powers, or in DNA pooling schemes which involve large families.