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

The problem of counterfactual visual explanations is considered. A new familyof discriminant explanations is introduced. These produce heatmaps thatattribute high scores to image regions informative of a classifier predictionbut not of a counter class. They connect attributive explanations, which arebased on a single heat map, to counterfactual explanations, which account forboth predicted class and counter class. The latter are shown to be computableby combination of two discriminant explanations, with reversed class pairs. Itis argued that self-awareness, namely the ability to produce classificationconfidence scores, is important for the computation of discriminantexplanations, which seek to identify regions where it is easy to discriminatebetween prediction and counter class. This suggests the computation ofdiscriminant explanations by the combination of three attribution maps. Theresulting counterfactual explanations are optimization free and thus muchfaster than previous methods. To address the difficulty of their evaluation, aproxy task and set of quantitative metrics are also proposed. Experiments underthis protocol show that the proposed counterfactual explanations outperform thestate of the art while achieving much higher speeds, for popular networks. In ahuman-learning machine teaching experiment, they are also shown to improve meanstudent accuracy from chance level to 95\%.

SCOUT: Self-aware Discriminant Counterfactual Explanations