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 Tools annual

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 monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Knowledge graph (KG) link prediction is a fundamental task in artificialintelligence, with applications in natural language processing, informationretrieval, and biomedicine. Recently, promising results have been achieved byleveraging cross-modal information in KGs, using ensembles that combineknowledge graph embeddings (KGEs) and contextual language models (LMs).However, existing ensembles are either (1) not consistently effective in termsof ranking accuracy gains or (2) impractically inefficient on larger datasetsdue to the combinatorial explosion problem of pairwise ranking with deeplanguage models. In this paper, we propose a novel tiered ranking architectureCascadER to maintain the ranking accuracy of full ensembling while improvingefficiency considerably. CascadER uses LMs to rerank the outputs of moreefficient base KGEs, relying on an adaptive subset selection scheme aimed atinvoking the LMs minimally while maximizing accuracy gain over the KGE.Extensive experiments demonstrate that CascadER improves MRR by up to 9 pointsover KGE baselines, setting new state-of-the-art performance on four benchmarkswhile improving efficiency by one or more orders of magnitude over competitivecross-modal baselines. Our empirical analyses reveal that diversity of modelsacross modalities and preservation of individual models' confidence signalshelp explain the effectiveness of CascadER, and suggest promising directionsfor cross-modal cascaded architectures. Code and pretrained models areavailable at https://github.com/tsafavi/cascader.

CascadER: Cross-Modal Cascading for Knowledge Graph Link Prediction