All Dates/Times are Australian Eastern Standard Time (AEST)

Technical Program

Paper Detail

Paper IDD7-S6-T3.3
Paper Title Active Binary Classification of Random Fields
Authors Arpan Mukherjee, Ali Tajer, Rensselaer Polytechnic Institute, United States; Pin-Yu Chen, Payel Das, IBM Thomas J. Watson Research Center, United States
Session D7-S6-T3: Testing
Chaired Session: Tuesday, 20 July, 23:40 - 00:00
Engagement Session: Wednesday, 21 July, 00:00 - 00:20
Abstract Consider a sequence of $n$ random variables $\bX=(X_1,\cdots,X_n)$ forming a random field (RF). $\bX$ is assumed to be generated according to one of the two possible classes of probability measures $\mcP\triangleq \big\{\P_i:i\in\{1,\cdots,m\}\big\}$ and $\mcQ\triangleq \big\{\mathbb{Q}_i:i\in\{1,\cdots,m\}\big\}$. There are upto $s$ realizations of the random variable $X_i$, for $i\in\{1,\cdots,m\}$. This paper addresses the following two questions : 1) Given a target classification reliability, what is the minimum number of samples required (out of $ns$) to classify $\bX$? 2) What is the optimal sampling order? This paper addresses these questions in the asymptote of large $n$.