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

Given a black-box classification model and an unlabeled evaluation datasetfrom some application domain, efficient strategies need to be developed toevaluate the model. Random sampling allows a user to estimate metrics likeaccuracy, precision, and recall, but may not provide insight to high-confidenceerrors. High-confidence errors are rare events for which the model is highlyconfident in its prediction, but is wrong. Such errors can represent costlymistakes and should be explicitly searched for. In this paper we propose ageneralization to the Adversarial Distance search that leverages concepts fromadversarial machine learning to identify predictions for which a classifier maybe overly confident. These predictions are useful instances to sample whenlooking for high-confidence errors because they are prone to a higher rate oferror than expected. Our generalization allows Adversarial Distance to beapplied to any classifier or data domain. Experimental results show that thegeneralized method finds errors at rates greater than expected given theconfidence of the sampled predictions, and outperforms competing methods.

Generalized Adversarial Distances to Efficiently Discover Classifier Errors