Graph neural networks (GNNs) have emerged as the mainstream paradigm for graph representation learning due to their effective message aggregation. However, this advantage also amplifies biases inherent in graph topology, raising fairness concerns. Existing fairness-aware GNNs provide satisfactory performance on fairness metrics such as Statistical Parity and Equal Opportunity while maintaining acceptable accuracy trade-offs. Unfortunately, we observe that this pursuit of fairness metrics neglects the GNN's ability to predict negative labels, which renders their predications with extremely high False Positive Rates (FPRs), resulting in negative effects in high-risk scenarios. To this end, we advocate that classification performance should be carefully calibrated while improving fairness, rather than simply constraining accuracy loss. Furthermore, we propose Fair GNN via Structural Entropy (FairGSE), a novel framework that maximizes two-dimensional structural entropy (2D-SE) to improve fairness without neglecting false positives. Experiments on several real-world datasets show FairGSE reduces FPR by 39% vs. state-of-the-art fairness-aware GNNs, with comparable fairness improvement.

Recommendation systems enhance user experience and retention by offering personalized content. As they increasingly influence social resource allocation (e.g., job recommendations), ensuring fair recommendations is becoming essential. Fairness notions in recommendation systems are mainly divided into individual and group fairness, and existing researchers notice the necessity of achieving both of them. However, the research indicates that there exists a trade-off relationship between them due to the way of defining individual fairness. Hence, existing methods attempt to redefine individual fairness. Unfortunately, they overlook the similarity comparison, which is essential for individual fairness. To address this issue, we redefine individual fairness from a ranking perspective. Based on this, we propose a framework that enhances both individual and group fairness. This framework includes a group fairness module, an individual fairness module, and a utility module. Extensive experiments show that our framework achieves a good balance between group fairness, individual fairness, and utility.

Graph-structured data is ubiquitous in today's connected world, driving extensive research in graph analysis. Graph Neural Networks (GNNs) have shown great success in this field, leading to growing interest in developing fair GNNs for critical applications. However, most existing fair GNNs focus on statistical fairness notions, which may be insufficient when dealing with statistical anomalies. Hence, motivated by the causal theory, there has been growing attention to mitigating root causes of unfairness utilizing graph counterfactuals. Unfortunately, existing methods for generating graph counterfactuals invariably require the sensitive attribute. Nevertheless, in many real-world applications, it is usually infeasible to obtain sensitive attributes due to privacy or legal issues, which challenge existing methods. In this paper, we propose a framework named Fairwos (improving Fairness withQut sensitive attributes). In particular, we first propose a mechanism to generate pseudo-sensitive attributes to remedy the problem of missing sensitive attributes, and then design a strategy for finding graph counterfactuals from the real dataset. To train fair GNNs, we propose a method to ensure that the embeddings from the original data are consistent with those from the graph counterfactuals, and dynamically adjust the weight of each pseudo-sensitive attribute to balance its contribution to fairness and utility. Furthermore, we theoretically demonstrate that minimizing the relation between these pseudo-sensitive attributes and the prediction can enable the fairness of GNNs. Experimental results on six real-world datasets show that our approach outperforms state-of-the-art methods in balancing utility and fairness.

Graph Neural Networks (GNNs) have achieved great success in node classification tasks. However, GNNs often suffer from class imbalance, which causes the underrepresentation of minority classes and subsequently gives rise to fairness issues. Most existing methods usually only enrich the information of minority class nodes from the perspective of quantity imbalance or topological imbalance, and therefore suffer from fairness concerns in class-imbalanced graphs with both imbalances. To address these problems, we propose a novel framework called the Imbalance-Aware Graph Neural Network (IAGNN). We design a new method to mitigate quantity imbalance. Specifically, we maximize the agreement between node representations in two augmented graph views to learn node representations. Then, we combine the supervised signals of the labeled data to train the classifier, thereby augmenting the minority class by predicted minority-class nodes of unlabeled nodes with high confidence. To solve topological imbalance, we propose a diffusion-based information enhancement method to further enrich the representation of minority-class nodes. Extensive experimental results on three real-world datasets demonstrate that IAGNN outperforms state-of-the-art baseline methods in terms of utility and fairness. The source code is available at https://github.com/Y7Lau/IAGNN.

Harmful meme detection presents a significant multimodal challenge that necessitates contextual background knowledge and comprehensive inference. Although some research studies have been related to harmful meme detection in English, detecting harmful memes in Chinese is also an unresolved issue. In this paper, to bridge this gap, we constructed a Chinese harmful meme detection dataset, named CHMEMES. Furthermore, existing multimodal alignment methods have shown poor performance in tasks involving harmful meme detection, where there is a mismatch between the image and text components. To improve the task, we propose a multimodal framework Semantic Contrastive Alignment fRamEwork (SCARE), which enables fully representing both cross-modal and intra-modal information. For cross-modal information, we introduce a cross-modal contrast alignment objective to maximize the mutual information between image and text. For intra-modal information, we design a new intra-modal contrast objective to achieve more robust visual and textual representation learning. Moreover, we present a simple yet efficient vision prompt tuning paradigm for parameter-efficient harmful meme detection. We conduct extensive experiments on the constructed Chinese dataset and the existing English dataset. Experimental results show that our method outperforms state-of-the-art baselines in harmful meme detection.