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

Learning sentence embeddings in an unsupervised manner is fundamental innatural language processing. Recent common practice is to couple pre-trainedlanguage models with unsupervised contrastive learning, whose success relies onaugmenting a sentence with a semantically-close positive instance to constructcontrastive pairs. Nonetheless, existing approaches usually depend on amono-augmenting strategy, which causes learning shortcuts towards theaugmenting biases and thus corrupts the quality of sentence embeddings. Astraightforward solution is resorting to more diverse positives from amulti-augmenting strategy, while an open question remains about how tounsupervisedly learn from the diverse positives but with uneven augmentingqualities in the text field. As one answer, we propose a novel Peer-ContrastiveLearning (PCL) with diverse augmentations. PCL constructs diverse contrastivepositives and negatives at the group level for unsupervised sentenceembeddings. PCL performs peer-positive contrast as well as peer-networkcooperation, which offers an inherent anti-bias ability and an effective way tolearn from diverse augmentations. Experiments on STS benchmarks verify theeffectiveness of PCL against its competitors in unsupervised sentenceembeddings.

PCL: Peer-Contrastive Learning with Diverse Augmentations for Unsupervised Sentence Embeddings