The cold-start recommendation has been challenging due to the limited historical interactions for new users and new items. Recently, methods based on meta-learning and graph neural networks have been effective in this problem. However, these methods mainly focus on the missing user-item interactions in cold-start scenarios, overlooking the missing of user/item feature information, which significantly limits the quality and effectiveness of node embeddings. To address this issue, we propose a new method called Feature-Structure Adaptive Completion Graph Neural Network (FS-GNN), which is designed to tackle the cold-start problem by simultaneously addressing the missing feature and structure information in a bipartite graph composed of users and items. Specifically, we first design a trainable feature completion module that leverages the knowledge emergence abilities of large language models to enhance node embedding and mitigate the impact of missing features. Then, we incorporate a three-channel structure completion module to simultaneously complete the structures among users-users, items-items, as well as users-items. Finally, we adaptively integrate the feature and structure completion modules in an end-to-end fashion, so as to minimize cross-module interference when completing features and structures simultaneously. This generates more comprehensive and robust embeddings for users and items in recommendation tasks. Experimental results on multiple public benchmark datasets demonstrate significant improvements in our proposed FS-GNN in cold-start scenarios, outperforming or being competitive with state-of-the-art methods.

Fraud detection that aims to discern frauds from the majority of benigns has become an increasingly prominent research field. Recently, Graph Neural Networks (GNNs) have been widely applied in graph-based fraud detection due to their outstanding data analysis and mining capabilities. However, owing to the inherent homophily-heterophily mixture and class imbalance of fraud graphs, most GNNs with homophily assumption inevitably suffer from local abnormal signal loss during information propagation, posing significant challenges in situations where frauds are rare and valuable. To address the aforementioned issues, we present a novel dynamic neighborhood modeling via node-subgraph contrastive learning for graph-based fraud detection, dubbed DCL-GFD. Specifically, we first design a node abnormality estimation module from the perspective of feature, which analyses the likelihood of a node belonging to fraud or benign by comparing the feature similarity between the target node and its corresponding subgraph. We then present a dynamic neighborhood modeling mechanism guided by the abnormal probability of a node to adaptively group and aggregate neighborhood information. By this means, the target node can effectively aggregate the neighbor information from the perspective of fraud or benign, thereby preserving as much fraud characteristics that occupy minority population as possible. Extensive experiments across four real-world fraud detection datasets demonstrate the superiority and effectiveness of our proposed DCL-GFD over state-of-the-art baselines.

The widespread adoption of online recruitment platforms has highlighted the challenge of achieving efficient and personalized person-job fit. Existing approaches focus on interaction patterns or textual content but fail to leverage the rich relationships among candidates, thus limiting their ability to capture the full spectrum of candidates' characteristics. In this paper, we propose a novel model, FitCLM, which integrates graph representation learning and large language models (LLMs) to enable a comprehensive understanding of candidate features. FitCLM first constructs a candidate relationship graph based on historical recruitment data, uncovering potential connections and similarities among candidates. It then utilizes LLMs to process historical job-candidate matching data, producing fine seman-tic representations. Finally, through self-supervised dual-view ranking optimization, FitCLM facilitates collaborative learning between the interaction graph and the relationship graph, while cross-modal contrastive learning aligns graph-based and LLM-generated embeddings. Extensive experimental results demonstrate that FitCLM outperforms state-of-the-art collaborative filtering and content-based baseline methods on NDCG@5 and MRR@5 metrics across two real-world recruitment datasets, with notable advantages in data-sparse scenarios. This study offers new insights into achieving accurate and efficient person-job fit, highlighting the significant potential of LLMs for generating contextual information from recruitment data.

In recent years, the Graph Transformer has demonstrated superiority on various graph-level tasks by facilitating global interactions among nodes. However, as for node-level tasks, the existing Graph Transformer cannot perform as well as expected. Actually, a node in a real-world graph does not necessarily have relationships with every other node, and this global interaction weakens node features. This raises a fundamental question: should we partition out an appropriate interaction channel based on graph structure so that noisy and irrelevant information will be filtered and every node can aggregate information in the optimal channel? We first perform a series of experiments on manually created graphs with varying homophily ratios. Surprisingly, we observe that different graph structures indeed require distinct optimal interaction channels. This leads us to ask whether we can develop a partitioning rule that ensures each node interacts with relevant and valuable targets. To overcome this challenge, we propose a novel Graph Transformer named Multi-channel Graphormer. The model is evaluated on six network datasets with different homophily ratios for the node classification task. Moreover, comprehensive experiments are conducted on two real datasets for the recommendation task. Experimental results show that the Multi-channel Graphormer surpasses state-of-the-art baselines, demonstrating superior performance.

Self-supervised tasks show significant advantages for node representation learning in recommender systems. This core idea of self-supervised task-based recommender systems depends on data augmentation to generate multi-view representations. However, there are two key challenges that are not well explored in existing self-supervised tasks: i) Restricted by the structure of the graph-based CF paradigm itself, the classical graph comparison learning architecture ignores the global structural information on the user-item interaction graph. ii) In a key part of existing contrast learning-random graph data enhancement schemes can significantly deteriorate model performance. To address these challenges, we propose a new hypergraph collaborative filtering with adaptive augmentation framework(HCFAA). It captures both local and global collaborative relationships on the user-item graph through a hypergraph-enhanced joint learning architecture. In particular, the designed adaptive structure-guided model ignores the noise introduced on unimportant edges, and thus learns the critical node information on the user-item graph. Comprehensive experimental studies on the Amazon dataset show that the method is effective, which provides an optimization scheme with a new perspective for the problems of key node loss in graph data enhancement and loss of higher-order structural information in GNN. The source code of our model can be available onhttps://github.com/RSnewbie/RS/tree/master/HCFAA.