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

Semantic relation prediction aims to mine the implicit relationships betweenobjects in heterogeneous graphs, which consist of different types of objectsand different types of links. In real-world scenarios, new semantic relationsconstantly emerge and they typically appear with only a few labeled data. Sincea variety of semantic relations exist in multiple heterogeneous graphs, thetransferable knowledge can be mined from some existing semantic relations tohelp predict the new semantic relations with few labeled data. This inspires anovel problem of few-shot semantic relation prediction across heterogeneousgraphs. However, the existing methods cannot solve this problem because theynot only require a large number of labeled samples as input, but also focus ona single graph with a fixed heterogeneity. Targeting this novel and challengingproblem, in this paper, we propose a Meta-learning based Graph neural networkfor Semantic relation prediction, named MetaGS. Firstly, MetaGS decomposes thegraph structure between objects into multiple normalized subgraphs, then adoptsa two-view graph neural network to capture local heterogeneous information andglobal structure information of these subgraphs. Secondly, MetaGS aggregatesthe information of these subgraphs with a hyper-prototypical network, which canlearn from existing semantic relations and adapt to new semantic relations.Thirdly, using the well-initialized two-view graph neural network andhyper-prototypical network, MetaGS can effectively learn new semantic relationsfrom different graphs while overcoming the limitation of few labeled data.Extensive experiments on three real-world datasets have demonstrated thesuperior performance of MetaGS over the state-of-the-art methods.

Few-Shot Semantic Relation Prediction across Heterogeneous Graphs