Speaker
Venkata Gandikota
(Syracuse University)
Description
Group testing is the study of pooling strategies that allow the identification of a small set of k defective items among a population of n using a small number of pooled tests. State-of-the-art testing schemes have shown that \Theta(k log n) schemes are both necessary and sufficient for the purpose which provides large gains when k is small (sublinear in n). However, these schemes are not resilient to deletion noise. In this work, we explore group testing algorithms resilient to deletion channels. We provide lower bounds, sufficient conditions, construction of matrices that meet the sufficiency condition and a decoding algorithm to recover the set of all k defective items.
