Empirical likelihood in non-standard settings Ian McKeague Florida State University  Empirical likelihood is a powerful technique for obtaining confidence regions based on nonparametric likelihood ratios.  This talk surveys various "non-standard" settings in which   empirical likelihood methods are potentially useful but in which theoretical justification has (until recently) been lacking.  Problems involving high dimensional data, plug-in estimates of infinite dimensional nuisance parameters, and cube-root asymptotics are discussed.