Fat-shattering dimension of -fold maxima
Journal of machine learning research (JMLR), 2021
Abstract
We provide improved estimates on the fat-shattering dimension of the -fold maximum of real-valued function classes. The latter consists of all ways of choosing functions, one from each of the classes, and computing their pointwise maximum. The bound is stated in terms of the fat-shattering dimensions of the component classes. For linear and affine function classes, we provide a considerably sharper upper bound and a matching lower bound, achieving, in particular, an optimal dependence on . Along the way, we point out and correct a number of erroneous claims in the literature.
View on arXivComments on this paper
