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

Evaluating the trustworthiness of a model's prediction is essential fordifferentiating between `right for the right reasons' and `right for the wrongreasons'. Identifying textual spans that determine the target label, known asfaithful rationales, usually relies on pipeline approaches or reinforcementlearning. However, such methods either require supervision and thus costlyannotation of the rationales or employ non-differentiable models. We propose adifferentiable training-framework to create models which output faithfulrationales on a sentence level, by solely applying supervision on the targettask. To achieve this, our model solves the task based on each rationaleindividually and learns to assign high scores to those which solved the taskbest. Our evaluation on three different datasets shows competitive resultscompared to a standard BERT blackbox while exceeding a pipeline counterpart'sperformance in two cases. We further exploit the transparent decision-makingprocess of these models to prefer selecting the correct rationales by applyingdirect supervision, thereby boosting the performance on the rationale-level.

Why do you think that? Exploring Faithful Sentence-Level Rationales Without Supervision