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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

In recent years, the development of large pretrained language models, such asBERT and GPT, significantly improved information extraction systems on varioustasks, including relation classification. State-of-the-art systems are highlyaccurate on scientific benchmarks. A lack of explainability is currently acomplicating factor in many real-world applications. Comprehensible systems arenecessary to prevent biased, counterintuitive, or harmful decisions.  We introduce semantic extents, a concept to analyze decision patterns for therelation classification task. Semantic extents are the most influential partsof texts concerning classification decisions. Our definition allows similarprocedures to determine semantic extents for humans and models. We provide anannotation tool and a software framework to determine semantic extents forhumans and models conveniently and reproducibly. Comparing both reveals thatmodels tend to learn shortcut patterns from data. These patterns are hard todetect with current interpretability methods, such as input reductions. Ourapproach can help detect and eliminate spurious decision patterns during modeldevelopment. Semantic extents can increase the reliability and security ofnatural language processing systems. Semantic extents are an essential step inenabling applications in critical areas like healthcare or finance. Moreover,our work opens new research directions for developing methods to explain deeplearning models.

Explaining Relation Classification Models with Semantic Extents