Meng Cao

Pre-trained Graph Neural Networks (GNNs) have demonstrated remarkable performance in graph mining tasks, yet they often amplify societal biases against protected demographic groups. Existing fairness-aware approaches primarily address discrimination based on a single sensitive attribute (e.g., gender or race), overlooking real-world scenarios where individuals possess multiple overlapping demographic characteristics, leading to unfair treatment of underrepresented subgroups. Moreover, incorporating extra fairness constraints into pre-trained GNNs usually requires full model retraining, which is computationally expensive and often impractical. To address these limitations, we propose a novel Multi-attribute Fairness-aware Prompt Tuning framework named MultiFPT. Our approach operates in two key stages: in the graph prompt learning stage, MultiFPT injects fairness-aware structural and feature prompts into pre-trained GNN inputs; in the adapter tuning stage, a lightweight adapter regularized by the Hilbert–Schmidt Independence Criterion (HSIC) enforces statistical independence between node representations and multiple sensitive attributes. Experiments on real-world datasets demonstrate that MultiFPT significantly improves multi-attribute fairness, reducing bias by approximately 30% on average in node classification while maintaining competitive predictive performance compared to state-of-the-art baselines.

The recent video grounding works attempt to introduce vanilla contrastive learning into video grounding. However, we claim that this naive solution is suboptimal. Contrastive learning requires two key properties: (1) alignment of features of similar samples, and (2) uniformity of the induced distribution of the normalized features on the hypersphere. Due to two annoying issues in video grounding: (1) the coexistence of some visual entities in both ground truth and other moments, i.e. semantic overlapping; (2) only a few moments in the video are annotated, i.e. sparse annotation dilemma, vanilla contrastive learning is unable to model the correlations between temporally distant moments and learned inconsistent video representations. Both characteristics lead to vanilla contrastive learning being unsuitable for video grounding. In this paper, we introduce Geodesic and Game Localization (G2L), a semantically aligned and uniform video grounding framework via geodesic and game theory. We quantify the correlations among moments leveraging the geodesic distance that guides the model to learn the correct cross-modal representations. Furthermore, from the novel perspective of game theory, we propose semantic Shapley interaction based on geodesic distance sampling to learn fine-grained semantic alignment in similar moments. Experiments on three benchmarks demonstrate the effectiveness of our method.

Short text classification has been a fundamental task in natural language processing, which benefits various applications, such as sentiment analysis, news tagging, and intent recommendation. However, classifying short texts is challenging due to the information sparsity in the text corpus. Besides, the performance of existing machine learning classification models largely relies on sufficient training data, yet labels can be scarce and expensive to obtain in real‐world text classification scenarios. In this article, we propose a novel self‐supervised short text classification method. Specifically, we first model the short text corpus as a heterogeneous graph to address the information sparsity problem. Then, we introduce a self‐attention‐based heterogeneous graph neural network model to learn short text embeddings. In addition, we adopt a self‐supervised learning framework to exploit internal and external similarities among short texts. Experiments on five real‐world short text benchmarks validate the effectiveness of our proposed method compared with the state‐of‐the‐art methods.

Heterogeneous Information Networks (HINs) are ubiquitous in our daily life, as they can describe complex interactions among various types of objects. Heterogeneous information network embedding aims to project the network elements of HINs into low-dimensional node representation vectors, which can facilitate effective analyze of HINs. To learn appropriate HIN embeddings, most existing methods usually adopt meta-paths to capture the heterogeneous semantic information, which require domain knowledge and trial-and-error to find suitable meta-paths. Besides, most existing methods fail to preserve high-order local structural information, and pay little attention to the implicit semantics in HINs. In this paper, we propose a local structural aware heterogeneous information network embedding model named LSA-HNE. Specifically, we first design a relational self-attention graph neural network model to aggregate heterogeneous information and automatically extract semantic similarity without using meta-paths. In addition, we employ a biased random walk based sampling method to extract the local structural information and preserve the implicit semantics in HINs. The experiments conducted on four real-world datasets show that our proposed model is effective compared with the state-of-the-art methods.

In general, online social networks include different types of nodes and edges, which means that online social networks are a type of heterogeneous information network. Link prediction is a very important research problem in heterogeneous social networks. The solution to this problem is generally to predict the possibility of a link between two nodes by extracting the characteristics of the nodes in the network. However, the information provided by a single network may not be sufficient, so useful information can be passed from other networks to assist in link prediction in the target network. This is called a partially aligned heterogeneous social network link prediction problem. In this paper, a method, called Meta-path and AUC optimization based Link Predictor(MALP), is proposed to predict the social links in the partially aligned social networks at the same time with a semi-supervised AUC optimization technology. Experimental results on real social network data show that our approach exhibits better predictive performance than other state-of-the-art methods.