Trustworthy and Secure Graph Modeling

Trustworthy and Secure Graph Modeling is critical for ensuring the robust operation of graph intelligence systems in complex environments. This field is dedicated to enhancing the resilience of models against structural perturbations and adversarial attacks through robust generalization techniques, while constructing rigorous security protection frameworks to prevent privacy leakage. Its core objective lies in trustworthiness enhancement, which involves quantifying and improving the transparency and reliability of model decisions through interpretability research and causal inference. By integrating methods such as differential privacy and federated learning, this technology effectively mitigates the risks associated with black-box algorithms, providing high-reliability decision support for the defense of critical infrastructure.