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

Natural Language Inference (NLI) datasets contain annotation artefactsresulting in spurious correlations between the natural language utterances andtheir respective entailment classes. These artefacts are exploited by neuralnetworks even when only considering the hypothesis and ignoring the premise,leading to unwanted biases. Belinkov et al. (2019b) proposed tackling thisproblem via adversarial training, but this can lead to learned sentencerepresentations that still suffer from the same biases. We show that the biascan be reduced in the sentence representations by using an ensemble ofadversaries, encouraging the model to jointly decrease the accuracy of thesedifferent adversaries while fitting the data. This approach produces morerobust NLI models, outperforming previous de-biasing efforts when generalisedto 12 other datasets (Belinkov et al., 2019a; Mahabadi et al., 2020). Inaddition, we find that the optimal number of adversarial classifiers depends onthe dimensionality of the sentence representations, with larger sentencerepresentations being more difficult to de-bias while benefiting from using agreater number of adversaries.

Avoiding the Hypothesis-Only Bias in Natural Language Inference via Ensemble Adversarial Training