Leslie M. Blaha

Ryan McCoppin, Colin Dawson, Sean M. Kennedy et al.

We introduce a metric for evaluating the robustness of a classifier, with particular attention to adversarial perturbations, in terms of expected functionality with respect to possible adversarial perturbations. A classifier is assumed to be non-functional (that is, has a functionality of zero) with respect to a perturbation bound if a conventional measure of performance, such as classification accuracy, is less than a minimally viable threshold when the classifier is tested on examples from that perturbation bound. Defining robustness in terms of an expected value is motivated by a domain general approach to robustness quantification.

Dustin L. Arendt, Leslie M. Blaha

Graphs change over time, and typically variations on the small multiples or animation pattern is used to convey this dynamism visually. However, both of these classical techniques have significant drawbacks, so a new approach, Storyline Visualization of Events on a Network (SVEN) is proposed. SVEN builds on storyline techniques, conveying nodes as contiguous lines over time. SVEN encodes time in a natural manner, along the horizontal axis, and optimizes the vertical placement of storylines to decrease clutter (line crossings, straightness, and bends) in the drawing. This paper demonstrates SVEN on several different flavors of real-world dynamic data, and outlines the remaining near-term future work.