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

Semi-supervised learning (SSL) tackles the label missing problem by enablingthe effective usage of unlabeled data. While existing SSL methods focus on thetraditional setting, a practical and challenging scenario called label MissingNot At Random (MNAR) is usually ignored. In MNAR, the labeled and unlabeleddata fall into different class distributions resulting in biased labelimputation, which deteriorates the performance of SSL models. In this work,class transition tracking based Pseudo-Rectifying Guidance (PRG) is devised forMNAR. We explore the class-level guidance information obtained by the Markovrandom walk, which is modeled on a dynamically created graph built over theclass tracking matrix. PRG unifies the historical information of classdistribution and class transitions caused by the pseudo-rectifying procedure tomaintain the model's unbiased enthusiasm towards assigning pseudo-labels to allclasses, so as the quality of pseudo-labels on both popular classes and rareclasses in MNAR could be improved. Finally, we show the superior performance ofPRG across a variety of MNAR scenarios, outperforming the latest SSL approachescombining bias removal solutions by a large margin. Code and model weights areavailable at https://github.com/NJUyued/PRG4SSL-MNAR.

Towards Semi-supervised Learning with Non-random Missing Labels