English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Facial analysis models are increasingly used in applications that haveserious impacts on people's lives, ranging from authentication to surveillancetracking. It is therefore critical to develop techniques that can revealunintended biases in facial classifiers to help guide the ethical use of facialanalysis technology. This work proposes a framework called \textit{imagecounterfactual sensitivity analysis}, which we explore as a proof-of-concept inanalyzing a smiling attribute classifier trained on faces of celebrities. Theframework utilizes counterfactuals to examine how a classifier's predictionchanges if a face characteristic slightly changes. We leverage recent advancesin generative adversarial networks to build a realistic generative model offace images that affords controlled manipulation of specific imagecharacteristics. We then introduce a set of metrics that measure the effect ofmanipulating a specific property on the output of the trained classifier.Empirically, we find several different factors of variation that affect thepredictions of the smiling classifier. This proof-of-concept demonstratespotential ways generative models can be leveraged for fine-grained analysis ofbias and fairness.

Image Counterfactual Sensitivity Analysis for Detecting Unintended Bias