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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recent work has suggested post-hoc explainers might be ineffective fordetecting spurious correlations in Deep Neural Networks (DNNs). However, weshow there are serious weaknesses with the existing evaluation frameworks forthis setting. Previously proposed metrics are extremely difficult to interpretand are not directly comparable between explainer methods. To alleviate theseconstraints, we propose a new evaluation methodology, Explainer DivergenceScores (EDS), grounded in an information theory approach to evaluateexplainers. EDS is easy to interpret and naturally comparable acrossexplainers. We use our methodology to compare the detection performance ofthree different explainers - feature attribution methods, influential examplesand concept extraction, on two different image datasets. We discover post-hocexplainers often contain substantial information about a DNN's dependence onspurious artifacts, but in ways often imperceptible to human users. Thissuggests the need for new techniques that can use this information to betterdetect a DNN's reliance on spurious correlations.

Explainer Divergence Scores (EDS): Some Post-Hoc Explanations May be Effective for Detecting Unknown Spurious Correlations