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

Language models (LMs) excel in in-distribution (ID) scenarios where train andtest data are independent and identically distributed. However, theirperformance often degrades in real-world applications like argument mining.Such degradation happens when new topics emerge, or other text domains andlanguages become relevant. To assess LMs' generalization abilities in suchout-of-distribution (OOD) scenarios, we simulate such distribution shifts bydeliberately withholding specific instances for testing, as from the socialmedia domain or the topic Solar Energy.  Unlike prior studies focusing on specific shifts and metrics in isolation, wecomprehensively analyze OOD generalization. We define three metrics to pinpointgeneralization flaws and propose eleven classification tasks covering topic,domain, and language shifts. Overall, we find superior performance ofprompt-based fine-tuning, notably when train and test splits primarily differsemantically. Simultaneously, in-context learning is more effective thanprompt-based or vanilla fine-tuning for tasks when training data embodies heavydiscrepancies in label distribution compared to testing data. This reveals acrucial drawback of gradient-based learning: it biases LMs regarding suchstructural obstacles.

Bridging Topic, Domain, and Language Shifts: An Evaluation of Comprehensive Out-of-Distribution Scenarios