We present the first adaptive strategy for active learning in the setting of classification with smooth decision boundary. The problem of adaptivity (to unknown distributional parameters) has remained opened since the seminal work of Castro and Nowak (2007), which first established (active learning) rates for this setting. While some recent advances on this problem establish adaptive rates in t...

This paper proposes adaptable boundary sets as a new version-space representation. It is shown that adaptable boundary sets can be adjusted to a version-space representation between boundary sets [5, 6, 7] and instance-based boundary sets [11, 12, 13]. This allows the choice of proper version-space representations to be realised dynamically during the learning phase.

Deep networks have recently been shown to be vulnerable to universal perturbations: there exist very small image-agnostic perturbations that cause most natural images to be misclassified by such classifiers. In this paper, we propose a quantitative analysis of the robustness of classifiers to universal perturbations, and draw a formal link between the robustness to universal perturbations, and ...

We describe a method for constructing rational quadratic patch boundary curves for scattered data in B3. The method has quadric boundary precision; if the given point and normal data are extracted from a quadric, then the boundary curves will lie on this quad&. Each boundary curve is a conic section represented in the rational BCzier representation.

in this paper, the boundary-based estimation of pressure distribution in the cardiovascular system is investigated using two dimensional flow images. the conventional methods of non-invasive estimation of pressure distribution in the cardiovascular flow domain use the differential form of governing equations. this study evaluates the advantages of using the integral form of the equations in the...

A Howell design of side s and order 2n, or more briefly an H(s, 2n), is an s x s array in which each cell is either empty or contains an unordered pair of elements from some (2n)-set V such that (1) every element of V occurs in precisely one cell of each row and each column, and (2) every unordered pair of elements from V is in at most one cell of the array. It follows immediately from the defi...