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

Partial-label learning (PLL) is an important weakly supervised learningproblem, which allows each training example to have a candidate label setinstead of a single ground-truth label. Identification-based methods have beenwidely explored to tackle label ambiguity issues in PLL, which regard the truelabel as a latent variable to be identified. However, identifying the truelabels accurately and completely remains challenging, causing noise in pseudolabels during model training. In this paper, we propose a new method calledCroSel, which leverages historical prediction information from models toidentify true labels for most training examples. First, we introduce a crossselection strategy, which enables two deep models to select true labels ofpartially labeled data for each other. Besides, we propose a novel consistentregularization term called co-mix to avoid sample waste and tiny noise causedby false selection. In this way, CroSel can pick out the true labels of mostexamples with high precision. Extensive experiments demonstrate the superiorityof CroSel, which consistently outperforms previous state-of-the-art methods onbenchmark datasets. Additionally, our method achieves over 90\% accuracy andquantity for selecting true labels on CIFAR-type datasets under varioussettings.

CroSel: Cross Selection of Confident Pseudo Labels for Partial-Label Learning