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

Compositional generalization allows efficient learning and human-likeinductive biases. Since most research investigating compositionalgeneralization in NLP is done on English, important questions remainunderexplored. Do the necessary compositional generalization abilities differacross languages? Can models compositionally generalize cross-lingually? As afirst step to answering these questions, recent work used neural machinetranslation to translate datasets for evaluating compositional generalizationin semantic parsing. However, we show that this entails critical semanticdistortion. To address this limitation, we craft a faithful rule-basedtranslation of the MCWQ dataset from English to Chinese and Japanese. Even withthe resulting robust benchmark, which we call MCWQ-R, we show that thedistribution of compositions still suffers due to linguistic divergences, andthat multilingual models still struggle with cross-lingual compositionalgeneralization. Our dataset and methodology will be useful resources for thestudy of cross-lingual compositional generalization in other tasks.

On Evaluating Multilingual Compositional Generalization with Translated Datasets