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

Real data collected from different applications that have additionaltopological structures and connection information are amenable to berepresented as a weighted graph. Considering the node labeling problem, GraphNeural Networks (GNNs) is a powerful tool, which can mimic experts' decision onnode labeling. GNNs combine node features, connection patterns, and graphstructure by using a neural network to embed node information and pass itthrough edges in the graph. We want to identify the patterns in the input dataused by the GNN model to make a decision and examine if the model works as wedesire. However, due to the complex data representation and non-lineartransformations, explaining decisions made by GNNs is challenging. In thiswork, we propose new graph features' explanation methods to identify theinformative components and important node features. Besides, we propose apipeline to identify the key factors used for node classification. We use fourdatasets (two synthetic and two real) to validate our methods. Our resultsdemonstrate that our explanation approach can mimic data patterns used for nodeclassification by human interpretation and disentangle different features inthe graphs. Furthermore, our explanation methods can be used for understandingdata, debugging GNN models, and examine model decisions.

Explain Graph Neural Networks to Understand Weighted Graph Features in Node Classification