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

Neural networks trained with ERM (empirical risk minimization) sometimeslearn unintended decision rules, in particular when their training data isbiased, i.e., when training labels are strongly correlated with undesirablefeatures. To prevent a network from learning such features, recent methodsaugment training data such that examples displaying spurious correlations(i.e., bias-aligned examples) become a minority, whereas the other,bias-conflicting examples become prevalent. However, these approaches aresometimes difficult to train and scale to real-world data because they rely ongenerative models or disentangled representations. We propose an alternativebased on mixup, a popular augmentation that creates convex combinations oftraining examples. Our method, coined SelecMix, applies mixup to contradictingpairs of examples, defined as showing either (i) the same label but dissimilarbiased features, or (ii) different labels but similar biased features.Identifying such pairs requires comparing examples with respect to unknownbiased features. For this, we utilize an auxiliary contrastive model with thepopular heuristic that biased features are learned preferentially duringtraining. Experiments on standard benchmarks demonstrate the effectiveness ofthe method, in particular when label noise complicates the identification ofbias-conflicting examples.

SelecMix: Debiased Learning by Contradicting-pair Sampling