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

Social networks are frequently polluted by rumors, which can be detected byadvanced models such as graph neural networks. However, the models arevulnerable to attacks and understanding the vulnerabilities is critical torumor detection in practice. To discover subtle vulnerabilities, we design apowerful attacking algorithm to camouflage rumors in social networks based onreinforcement learning that can interact with and attack any black-boxdetectors. The environment has exponentially large state spaces, high-ordergraph dependencies, and delayed noisy rewards, making the state-of-the-artend-to-end approaches difficult to learn features as large learning costs andexpressive limitation of graph deep models. Instead, we design domain-specificfeatures to avoid learning features and produce interpretable attack policies.To further speed up policy optimization, we devise: (i) a credit assignmentmethod that decomposes delayed rewards to atomic attacking actions proportionalto the their camouflage effects on target rumors; (ii) a time-dependent controlvariate to reduce reward variance due to large graphs and many attacking steps,supported by the reward variance analysis and a Bayesian analysis of theprediction distribution. On three real world datasets of rumor detection tasks,we demonstrate: (i) the effectiveness of the learned attacking policy comparedto rule-based attacks and current end-to-end approaches; (ii) the usefulness ofthe proposed credit assignment strategy and variance reduction components;(iii) the interpretability of the policy when generating strong attacks via thecase study.

Interpretable and Effective Reinforcement Learning for Attacking against Graph-based Rumor Detection