This paper provides a systematic approach to semiparametric identification that is based on statistical information as measure of its “quality.” Identification can be regular or irregular, depending whether the Fisher for parameter positive zero, respectively. I first characterize these cases in models with densities linear an infinite-dimensional parameter. then introduce novel “generalized in...

in this paper, the geometric distribution is considered. the means, variances, and covariances of its order statistics are derived. the fisher information in any set of order statistics in any distribution can be represented as a sum of fisher information in at most two order statistics. it is shown that, for the geometric distribution, it can be further simplified to a sum of fisher informatio...

This paper studies the over-parameterization of deep neural networks using Fisher Information Matrix from information geometry. We identify several surprising trends in structure its eigenspectrum, and how this relates to eigenspectrum data correlation matrix. topology predictions model develop a "model reduction'' method for networks. ongoing investigation hypothesizes certain universal FIM th...