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/

Zendy Plus

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

Zendy Tools

Zendy Open

Semi-supervised learning (SSL) is one of the dominant approaches to addressthe annotation bottleneck of supervised learning. Recent SSL methods caneffectively leverage a large repository of unlabeled data to improveperformance while relying on a small set of labeled data. One common assumptionin most SSL methods is that the labeled and unlabeled data are from the samedata distribution. However, this is hardly the case in many real-worldscenarios, which limits their applicability. In this work, instead, we attemptto solve the challenging open-world SSL problem that does not make such anassumption. In the open-world SSL problem, the objective is to recognizesamples of known classes, and simultaneously detect and cluster samplesbelonging to novel classes present in unlabeled data. This work introducesOpenLDN that utilizes a pairwise similarity loss to discover novel classes.Using a bi-level optimization rule this pairwise similarity loss exploits theinformation available in the labeled set to implicitly cluster novel classsamples, while simultaneously recognizing samples from known classes. Afterdiscovering novel classes, OpenLDN transforms the open-world SSL problem into astandard SSL problem to achieve additional performance gains using existing SSLmethods. Our extensive experiments demonstrate that OpenLDN outperforms thecurrent state-of-the-art methods on multiple popular classification benchmarkswhile providing a better accuracy/training time trade-off.

OpenLDN: Learning to Discover Novel Classes for Open-World Semi-Supervised Learning