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

As a promising solution of reducing annotation cost, training multi-labelmodels with partial positive labels (MLR-PPL), in which merely few positivelabels are known while other are missing, attracts increasing attention. Due tothe absence of any negative labels, previous works regard unknown labels asnegative and adopt traditional MLR algorithms. To reject noisy labels, recentworks regard large loss samples as noise but ignore the semantic correlationdifferent multi-label images. In this work, we propose to explore semanticcorrelation among different images to facilitate the MLR-PPL task.Specifically, we design a unified framework, Category-Adaptive Label Discoveryand Noise Rejection, that discovers unknown labels and rejects noisy labels foreach category in an adaptive manner. The framework consists of twocomplementary modules: (1) Category-Adaptive Label Discovery module firstmeasures the semantic similarity between positive samples and then complementunknown labels with high similarities; (2) Category-Adaptive Noise Rejectionmodule first computes the sample weights based on semantic similarities fromdifferent samples and then discards noisy labels with low weights. Besides, wepropose a novel category-adaptive threshold updating that adaptively adjuststhe threshold, to avoid the time-consuming manual tuning process. Extensiveexperiments demonstrate that our proposed method consistently outperformscurrent leading algorithms.

Category-Adaptive Label Discovery and Noise Rejection for Multi-label Image Recognition with Partial Positive Labels