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

Modern Natural Language Processing (NLP) models are known to be sensitive toinput perturbations and their performance can decrease when applied toreal-world, noisy data. However, it is still unclear why models are less robustto some perturbations than others. In this work, we test the hypothesis thatthe extent to which a model is affected by an unseen textual perturbation(robustness) can be explained by the learnability of the perturbation (definedas how well the model learns to identify the perturbation with a small amountof evidence). We further give a causal justification for the learnabilitymetric. We conduct extensive experiments with four prominent NLP models --TextRNN, BERT, RoBERTa and XLNet -- over eight types of textual perturbationson three datasets. We show that a model which is better at identifying aperturbation (higher learnability) becomes worse at ignoring such aperturbation at test time (lower robustness), providing empirical support forour hypothesis.

Interpreting the Robustness of Neural NLP Models to Textual Perturbations