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

Training with fewer annotations is a key issue for applying deep models tovarious practical domains. To date, semi-supervised learning has achieved greatsuccess in training with few annotations. However, confirmation bias increasesdramatically as the number of annotations decreases making it difficult tocontinue reducing the number of annotations. Based on the observation that thequality of pseudo-labels early in semi-supervised training plays an importantrole in mitigating confirmation bias, in this paper we propose an activeself-semi-supervised learning (AS3L) framework. AS3L bootstraps semi-supervisedmodels with prior pseudo-labels (PPL), where PPL is obtained by labelpropagation over self-supervised features. We illustrate that the accuracy ofPPL is not only affected by the quality of features, but also by the selectionof the labeled samples. We develop active learning and label propagationstrategies to obtain better PPL. Consequently, our framework can significantlyimprove the performance of models in the case of few annotations while reducingthe training time. Experiments on four semi-supervised learning benchmarksdemonstrate the effectiveness of the proposed methods. Our method outperformsthe baseline method by an average of 7\% on the four datasets and outperformsthe baseline method in accuracy while taking about 1/3 of the training time.

Training from a Better Start Point: Active Self-Semi-Supervised Learning for Few Labeled Samples