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

When an NLP model is trained on text data from one time period and tested ordeployed on data from another, the resulting temporal misalignment can degradeend-task performance. In this work, we establish a suite of eight diverse tasksacross different domains (social media, science papers, news, and reviews) andperiods of time (spanning five years or more) to quantify the effects oftemporal misalignment. Our study is focused on the ubiquitous setting where apretrained model is optionally adapted through continued domain-specificpretraining, followed by task-specific finetuning. We establish a suite oftasks across multiple domains to study temporal misalignment in modern NLPsystems. We find stronger effects of temporal misalignment on task performancethan have been previously reported. We also find that, while temporaladaptation through continued pretraining can help, these gains are smallcompared to task-specific finetuning on data from the target time period. Ourfindings motivate continued research to improve temporal robustness of NLPmodels.

Time Waits for No One! Analysis and Challenges of Temporal Misalignment