Published:
Summary: we study a novel problem of unnoticeable graph backdoor attacks with limited attack budget. The proposed method can achieve high attack rate on large-scale datasets with samll costs.
Published:
Summary: we study a novel problem of developing robust GNNs on noisy graphs with limited labeled nodes. We propose to learn a denoised and dense graph, which can down-weight or eliminate noisy edges and facilitate message passing of GNNs to alleviate the issue of limited labeled nodes.
Published:
Summary: We give a taxonomy of the trustworthy GNNs in privacy, robustness, fairness, and explainability. For each aspect, we categorize existing works into various categories, give general frameworks in each category, describe the details of representative and state-of-the-art methods, and discuss future works.
Published:
Summary: we study a novel problem of unnoticeable graph backdoor attacks with limited attack budget. The proposed method can achieve high attack rate on large-scale datasets with samll costs.
Published:
Summary: we study a novel problem of developing robust GNNs on noisy graphs with limited labeled nodes. We propose to learn a denoised and dense graph, which can down-weight or eliminate noisy edges and facilitate message passing of GNNs to alleviate the issue of limited labeled nodes.
Published:
Summary: we study a novel problem of unnoticeable graph backdoor attacks with limited attack budget. The proposed method can achieve high attack rate on large-scale datasets with samll costs.
Published:
Summary: we study a novel problem of developing robust GNNs on noisy graphs with limited labeled nodes. We propose to learn a denoised and dense graph, which can down-weight or eliminate noisy edges and facilitate message passing of GNNs to alleviate the issue of limited labeled nodes.
Published:
Summary: We give a taxonomy of the trustworthy GNNs in privacy, robustness, fairness, and explainability. For each aspect, we categorize existing works into various categories, give general frameworks in each category, describe the details of representative and state-of-the-art methods, and discuss future works.