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/

Zendy Plus

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

Zendy Tools

Zendy Open

There is a recent surge of interest in using attention as explanation ofmodel predictions, with mixed evidence on whether attention can be used assuch. While attention conveniently gives us one weight per input token and iseasily extracted, it is often unclear toward what goal it is used asexplanation. We find that often that goal, whether explicitly stated or not, isto find out what input tokens are the most relevant to a prediction, and thatthe implied user for the explanation is a model developer. For this goal anduser, we argue that input saliency methods are better suited, and that thereare no compelling reasons to use attention, despite the coincidence that itprovides a weight for each input. With this position paper, we hope to shiftsome of the recent focus on attention to saliency methods, and for authors toclearly state the goal and user for their explanations.

The elephant in the interpretability room: Why use attention as explanation when we have saliency methods?