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

The advent of pre-trained Language Models (LMs) has markedly advanced naturallanguage processing, but their efficacy in out-of-distribution (OOD) scenariosremains a significant challenge. Computational argumentation (CA), modelinghuman argumentation processes, is a field notably impacted by these challengesbecause complex annotation schemes and high annotation costs naturally lead toresources barely covering the multiplicity of available text sources andtopics. Due to this data scarcity, generalization to data from uncoveredcovariant distributions is a common challenge for CA tasks like stancedetection or argument classification. This work systematically assesses LMs'capabilities for such OOD scenarios. While previous work targets specific OODtypes like topic shifts or OOD uniformly, we address three prevalent OODscenarios in CA: topic shift, domain shift, and language shift. Our findingschallenge the previously asserted general superiority of in-context learning(ICL) for OOD. We find that the efficacy of such learning paradigms varies withthe type of OOD. Specifically, while ICL excels for domain shifts, prompt-basedfine-tuning surpasses for topic shifts. To sum up, we navigate theheterogeneity of OOD scenarios in CA and empirically underscore the potentialof base-sized LMs in overcoming these challenges.

How to Handle Different Types of Out-of-Distribution Scenarios in Computational Argumentation? A Comprehensive and Fine-Grained Field Study