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

Commonsense knowledge is critical in human reading comprehension. Whilemachine comprehension has made significant progress in recent years, theability in handling commonsense knowledge remains limited. Synonyms are one ofthe most widely used commonsense knowledge. Constructing adversarial dataset isan important approach to find weak points of machine comprehension models andsupport the design of solutions. To investigate machine comprehension models'ability in handling the commonsense knowledge, we created a Question and AnswerDataset with common knowledge of Synonyms (QADS). QADS are questions generatedbased on SQuAD 2.0 by applying commonsense knowledge of synonyms. The synonymsare extracted from WordNet. Words often have multiple meanings and synonyms. Weused an enhanced Lesk algorithm to perform word sense disambiguation toidentify synonyms for the context. ELECTRA achieves the state-of-art result onthe SQuAD 2.0 dataset in 2019. With scale, ELECTRA can achieve similarperformance as BERT does. However, QADS shows that ELECTRA has little abilityto handle commonsense knowledge of synonyms. In our experiment, ELECTRA-smallcan achieve 70% accuracy on SQuAD 2.0, but only 20% on QADS. ELECTRA-large didnot perform much better. Its accuracy on SQuAD 2.0 is 88% but droppedsignificantly to 26% on QADS. In our earlier experiments, BERT, although alsofailed badly on QADS, was not as bad as ELECTRA. The result shows that eventop-performing NLP models have little ability to handle commonsense knowledgewhich is essential in reading comprehension.

Commonsense knowledge adversarial dataset that challenges ELECTRA