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

Pretrained language models have achieved super-human performances on manyMachine Reading Comprehension (MRC) benchmarks. Nevertheless, their relativeinability to defend against adversarial attacks has spurred skepticism abouttheir natural language understanding. In this paper, we ask whether trainingwith unanswerable questions in SQuAD 2.0 can help improve the robustness of MRCmodels against adversarial attacks. To explore that question, we fine-tunethree state-of-the-art language models on either SQuAD 1.1 or SQuAD 2.0 andthen evaluate their robustness under adversarial attacks. Our experimentsreveal that current models fine-tuned on SQuAD 2.0 do not initially appear tobe any more robust than ones fine-tuned on SQuAD 1.1, yet they reveal a measureof hidden robustness that can be leveraged to realize actual performance gains.Furthermore, we find that the robustness of models fine-tuned on SQuAD 2.0extends to additional out-of-domain datasets. Finally, we introduce a newadversarial attack to reveal artifacts of SQuAD 2.0 that current MRC models arelearning.

The Impacts of Unanswerable Questions on the Robustness of Machine Reading Comprehension Models