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

Weak supervision enables efficient development of training sets by reducingthe need for ground truth labels. However, the techniques that make weaksupervision attractive -- such as integrating any source of signal to estimateunknown labels -- also entail the danger that the produced pseudolabels arehighly biased. Surprisingly, given everyday use and the potential for increasedbias, weak supervision has not been studied from the point of view of fairness.We begin such a study, starting with the observation that even when a fairmodel can be built from a dataset with access to ground-truth labels, thecorresponding dataset labeled via weak supervision can be arbitrarily unfair.To address this, we propose and empirically validate a model for sourceunfairness in weak supervision, then introduce a simple counterfactualfairness-based technique that can mitigate these biases. Theoretically, we showthat it is possible for our approach to simultaneously improve both accuracyand fairness -- in contrast to standard fairness approaches that suffer fromtradeoffs. Empirically, we show that our technique improves accuracy on weaksupervision baselines by as much as 32\% while reducing demographic parity gapby 82.5\%. A simple extension of our method aimed at maximizing performanceproduces state-of-the-art performance in five out of ten datasets in the WRENCHbenchmark.

Mitigating Source Bias for Fairer Weak Supervision