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

In unsupervised feature learning, sample specificity based methods ignore theinter-class information, which deteriorates the discriminative capability ofrepresentation models. Clustering based methods are error-prone to explore thecomplete class boundary information due to the inevitable class inconsistentsamples in each cluster. In this work, we propose a novel clustering basedmethod, which, by iteratively excluding class inconsistent samples duringprogressive cluster formation, alleviates the impact of noise samples in asimple-yet-effective manner. Our approach, referred to as Progressive ClusterPurification (PCP), implements progressive clustering by gradually reducing thenumber of clusters during training, while the sizes of clusters continuouslyexpand consistently with the growth of model representation capability. With awell-designed cluster purification mechanism, it further purifies clusters byfiltering noise samples which facilitate the subsequent feature learning byutilizing the refined clusters as pseudo-labels. Experiments on commonly usedbenchmarks demonstrate that the proposed PCP improves baseline method withsignificant margins. Our code will be available athttps://github.com/zhangyifei0115/PCP.

Progressive Cluster Purification for Unsupervised Feature Learning