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

Zero-shot cross-lingual named entity recognition (NER) aims at transferringknowledge from annotated and rich-resource data in source languages tounlabeled and lean-resource data in target languages. Existing mainstreammethods based on the teacher-student distillation framework ignore the rich andcomplementary information lying in the intermediate layers of pre-trainedlanguage models, and domain-invariant information is easily lost duringtransfer. In this study, a mixture of short-channel distillers (MSD) method isproposed to fully interact the rich hierarchical information in the teachermodel and to transfer knowledge to the student model sufficiently andefficiently. Concretely, a multi-channel distillation framework is designed forsufficient information transfer by aggregating multiple distillers as amixture. Besides, an unsupervised method adopting parallel domain adaptation isproposed to shorten the channels between the teacher and student models topreserve domain-invariant features. Experiments on four datasets across ninelanguages demonstrate that the proposed method achieves new state-of-the-artperformance on zero-shot cross-lingual NER and shows great generalization andcompatibility across languages and fields.

WIDER & CLOSER: Mixture of Short-channel Distillers for Zero-shot Cross-lingual Named Entity Recognition