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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The recent success of large pre-trained language models (PLMs) heavily hingeson massive labeled data, which typically produces inferior performance inlow-resource scenarios. To remedy this dilemma, we study self-training as oneof the predominant semi-supervised learning (SSL) approaches, which utilizeslarge-scale unlabeled data to generate synthetic examples. However, too manynoisy labels will hurt the model performance, and the self-training procedurerequires multiple training iterations making it more expensive if all the modelparameters of the PLM are updated. This paper presents UPET, a novelUncertainty-aware Parameter-Efficient self-Training framework to effectivelyand efficiently address the labeled data scarcity issue. Specifically, weincorporate Monte Carlo (MC) dropout in Bayesian neural network (BNN) toperform uncertainty estimation for the teacher model and then judiciouslyselect reliable pseudo-labeled examples based on confidence and certainty.During the student training, we introduce multiple parameter-efficient learning(PEL) paradigms that allow the optimization of only a small percentage ofparameters. We also propose a novel Easy-Hard Contrastive Tuning to enhance therobustness and generalization. Extensive experiments over multiple downstreamtasks demonstrate that UPET achieves a substantial improvement in terms ofperformance and efficiency. Our codes and data are released at https://github.com/wjn1996/UPET.

Uncertainty-aware Parameter-Efficient Self-training for Semi-supervised Language Understanding