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

Background: Neural networks produce biased classification results due tocorrelation bias (they learn correlations between their inputs and outputs toclassify samples, even when those correlations do not representcause-and-effect relationships).  Objective: This study introduces a fully unsupervised method of mitigatingcorrelation bias, demonstrated with sentiment classification on COVID-19 socialmedia data.  Methods: Correlation bias in sentiment classification often arises inconversations about controversial topics. Therefore, this study usesadversarial learning to contrast clusters based on sentiment classificationlabels, with clusters produced by unsupervised topic modeling. This discouragesthe neural network from learning topic-related features that produce biasedclassification results.  Results: Compared to a baseline classifier, neural contrastive clusteringapproximately doubles accuracy on bias-prone sentences for human-labeledCOVID-19 social media data, without adversely affecting the classifier'soverall F1 score. Despite being a fully unsupervised approach, neuralcontrastive clustering achieves a larger improvement in accuracy on bias-pronesentences than a supervised masking approach.  Conclusions: Neural contrastive clustering reduces correlation bias insentiment text classification. Further research is needed to exploregeneralizing this technique to other neural network architectures andapplication domains.

Neural Contrastive Clustering: Fully Unsupervised Bias Reduction for Sentiment Classification