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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Graph neural networks (GNNs) are widely used in many applications. However,their robustness against adversarial attacks is criticized. Prior studies showthat using unnoticeable modifications on graph topology or nodal features cansignificantly reduce the performances of GNNs. It is very challenging to designrobust graph neural networks against poisoning attack and several efforts havebeen taken. Existing work aims at reducing the negative impact from adversarialedges only with the poisoned graph, which is sub-optimal since they fail todiscriminate adversarial edges from normal ones. On the other hand, cleangraphs from similar domains as the target poisoned graph are usually availablein the real world. By perturbing these clean graphs, we create supervisedknowledge to train the ability to detect adversarial edges so that therobustness of GNNs is elevated. However, such potential for clean graphs isneglected by existing work. To this end, we investigate a novel problem ofimproving the robustness of GNNs against poisoning attacks by exploring cleangraphs. Specifically, we propose PA-GNN, which relies on a penalizedaggregation mechanism that directly restrict the negative impact of adversarialedges by assigning them lower attention coefficients. To optimize PA-GNN for apoisoned graph, we design a meta-optimization algorithm that trains PA-GNN topenalize perturbations using clean graphs and their adversarial counterparts,and transfers such ability to improve the robustness of PA-GNN on the poisonedgraph. Experimental results on four real-world datasets demonstrate therobustness of PA-GNN against poisoning attacks on graphs. Code and data areavailable here: https://github.com/tangxianfeng/PA-GNN.

Transferring Robustness for Graph Neural Network Against Poisoning Attacks