Improving Unsupervised Network Alignment with Matched Neighborhood Consistency

Network alignment (NA) that identifies equivalent nodes across networks is an effective tool for integrating knowledge from multiple networks. The state-of-the-art NA methods learn inter-network node similarities based on labeled anchor links, which are costly, time-consuming, and difficult to acquire. Therefore, a few unsupervised network alignment (UNA) methods propose solving NA problems without anchor links. However, most existing UNA methods rely on discriminative attributes to capture nodes’ similarities and are hard to obtain optimal one-to-one alignments. Toward these issues, this article proposes a novel method named HackGAN to solve the UNA problem solely based on the structural information. Specifically, HackGAN represents nodes with embeddings based on an unsupervised graph neural network (GNN) to capture their global and local structural features. After that, it initializes mapping functions to transform the embedding spaces of different networks into the same vector space by iteratively solving the Wasserstein–Procrustes problem. The mapping functions are then refined by an adversarial model with cycle-consistency and Sinkhorn distance losses to obtain optimized one-to-one mappings. Based on the distances between mapped embeddings, accurate and robust results are obtained with a collective alignment algorithm. Experimental comparisons on both synthetic and real-world datasets demonstrate the superiority of HackGAN.

Graph Neural Networks (GNNs) have achieved significant success in addressing node classification tasks. However, the effectiveness of traditional GNNs degrades on heterophilic graphs, where connected nodes often belong to different labels or properties. While recent work has introduced mechanisms to improve GNN performance under heterophily, certain key limitations still exist. Most existing models apply a fixed aggregation depth across all nodes, overlooking the fact that nodes may require different propagation depths based on their local homophily levels and neighborhood structures. Moreover, many methods are tailored to either homophilic or heterophilic settings, lacking the flexibility to generalize across both regimes. To address these challenges, we develop a theoretical framework that links local structural and label characteristics to information propagation dynamics at the node level. Our analysis shows that optimal aggregation depth varies across nodes and is critical for preserving class-discriminative information. Guided by this insight, we propose a novel adaptive-depth GNN architecture that dynamically selects node-specific aggregation depths using theoretically grounded metrics. Our method seamlessly adapts to both homophilic and heterophilic patterns within a unified model. Extensive experiments demonstrate that our approach consistently enhances the performance of standard GNN backbones across diverse benchmarks.

Graph Neural Networks (GNNs) have proven effective and typically benefit from pre-training on accessible graphs to enhance performance on tasks with limited labeled data. However, existing GNNs are constrained by the ``one-domain-one-model'' limitation, which restricts their effectiveness across diverse graph domains. In this paper, we tackle this problem by developing a method called Multi-Domain Pre-training for a Unified GNN Model (MDP-GNN). This method is based on the philosophical notion that everything is interconnected, suggesting that a latent meta-domain exists to encompass the diverse graph domains and their interconnections. MDP-GNN seeks to identify and utilize this meta-domain to train a unified GNN model through three core strategies. Firstly, it integrates node feature semantics from different domains to create unified representations. Secondly, it employs a bi-level learning strategy to build a domain-synthesized network that identifies latent connections to facilitate cross-domain knowledge transfer. Thirdly, it uses Wasserstein distance to map diverse domains into the common meta-domain for graph distribution alignment. We validate the effectiveness of MDP-GNN through theoretical analysis and extensive experiments on four real-world graph datasets, showing its superiority in enhancing GNN performance across diverse domains.

Network alignment (NA) offers a comprehensive way to build associations between different networks by identifying shared nodes. While the majority of current NA methods rely on the topological consistency assumption, which posits that shared nodes across different networks typically have similar local structures or neighbors, we argue that anchor nodes, which play a pivotal role in NA, face a more challenging scenario that is often overlooked. In this paper, we conduct extensive statistical analysis across networks to investigate the connection status of labeled anchor node pairs and categorize them into four situations. Based on our analysis, we propose an end-to-end network alignment framework that uses node representations as a distribution rather than a point vector to better handle the structural diversity of networks. To mitigate the influence of specific nodes, we introduce a mask mechanism during the representation learning process. In addition, we utilize meta-learning to generalize the learned information on labeled anchor node pairs to other node pairs. Finally, we perform comprehensive experiments on both real-world and synthetic datasets to confirm the efficacy of our proposed method. The experimental results demonstrate that the proposed model outperforms the state-of-the-art methods significantly.

Graph neural networks (GNNs) have revolutionized deep learning on non-Euclidean data domains, and are extensively used in fields such as social media and recommendation systems. However, complex relational data structures such as hypergraphs, pose challenges for GNNs in terms of their ability to model, embed, and learn relational complexities of multigraphs. Most GNNs focus on capturing flat local neighborhoods of a node thus failing to account for structural properties of multi-relational graphs. This paper introduces Hypergraph Link Prediction (HLP), a novel approach of encoding the multilink structure of graphs. HLP allows pooling operations to incorporate a 360 degrees overview of a node interaction profile, by learning local neighborhood and global hypergraph structure simultaneously. Global graph information is injected into node representations, such that unique global structural patterns of every node are encoded at the node level. HLP leverages the augmented hypergraph adjacency matrix to incorporate the depth of the hypergraph in the convolutional layers. The model is applied to the task of predicting multi-drug interactions, by modeling relations between pairs of drugs as a hypergraph. The existence and the type of drug interactions between the same pair of drugs are mapped as multiple edges, and can be inferred by learning the multigraph local and global structure concurrently. To account for molecular graph properties of a drug, additional drug chemical graph structural fingerprints are included as node attributes.

Network alignment is the matching of two networks with corresponding nodes that belong to the same user or entity. The most common application is to analyze which accounts belong to the same user in two social networks. Most of existing techniques rely on matrix factorization so that they cannot be scaled to large-scale networks, are constrained by strict constraints, and cannot learn node embedding without a training set. In this paper, we propose an unsupervised network alignment model based on multi-level graph attention networks. The model uses multi-level graph attention network to learn the embedded representation of nodes, satisfying attribute and structure constraints of alignment. Augmented learning process is proposed to simulate attribute noise and structural noise to improve adaptability of the model. Extensive experiments on real datasets show that the proposed model performs better than the state-of-the-art network alignment model. We also demonstrate the robustness of the proposed model.

Knowledge graph (KG) as the source of side information has been proven to be useful to alleviate the data sparsity and cold start. Existing methods usually exploit the semantic relations between entities by learning structural or semantic paths information. However, they ignore the difficulty of information fusion and network alignment when constructing knowledge graph from different domains, and do not take temporal context into account. To address the limitations of existing methods, we propose a novel High-order semantic Relations-based Temporal Recommendation (HRTR), which captures the joint effects of high-order semantic relations in Collaborative Knowledge Graph (CKG) and temporal context. Firstly, it automatically extracts different order connectivities to represent semantic relations between entities from CKG. Then, we define a joint learning model to capture high-quality representations of users, items, and their attributes by employing TransE and recurrent neural network, which captures not only structural information, but also sequence information by encoding semantic paths, and to take their representations as the users’/items’ long-term static features. Next, we respectively employ LSTM and attention machine to capture the users’ and items’ short-term dynamic preferences. At last, the long-short term features are seamlessly fused into recommender system. We conduct extensive experiments on real-world datasets and the evaluation results show that HRTR achieves significant superiority over several state-of-the-art baselines.

Deep learning on graphs is an upcoming area of study. This chapter provides an introduction to graph neural networks (GNNs), a type of neural network that is designed to process data represented in the form of graphs. First, it summarizes the explanation of deep learning on graphs. The fundamental concepts of graph neural networks, as well as GNN theories, are then explained. In this chapter, different types of graph neural network (GNN) are also explained. At the end, the applications of graph neural network where GNN is used and for what purpose it is going to be used are explained. This also explores the various applications of GNNs in fields such as social network analysis, recommendation systems, drug discovery, computer vision, and natural language processing. With the increasing prevalence of graph data, GNNs are becoming increasingly important and will likely continue to play a significant role in many fields in the future.

Behavioral theory assumes that leaders can be identified by their daily behaviors. Social network analysis helps to understand behavioral patterns within their social networks. This work considers leaders as the managerial personnel of the organization and differentiates managements from non-managerial staff by their behavior with five different types of interactions with PageRank and their attributes in modern organizations. PageRank and word embedding using word2vec with phrases from features are adopted to extract new features for the identification of managerial staff. Both traditional machine learning methods and graph neural networks are utilized with real-world data from an Austrian IT company called Knapp System Integration. Our experimental results show that the proposed new features extracted using PageRank with different types of interactions and word2vec with phrases significantly improve the identification accuracy. We also propose to use graph neural networks as an effective learning algorithm to identify managers from organizations. Our approach can identify managerial staff with an accuracy of around 80%, which demonstrates that managers could be identified through social network analysis. By analyzing the behaviors of members, the proposed method is effective as a performance appraisal tool for organizations. The study facilitates sustainable management by helping organizations to retain managerial talents or to invite potential talents to join the management team.

Analogous to genomic sequence alignment, biological network alignment identifies conserved regions between networks of different species. Then, function can be transferred from well- to poorly-annotated species between aligned network regions. Network alignment typically encompasses two algorithmic components: node cost function (NCF), which measures similarities between nodes in different networks, and alignment strategy (AS), which uses these similarities to rapidly identify high-scoring alignments. Different methods use both different NCFs and different ASs. Thus, it is unclear whether the superiority of a method comes from its NCF, its AS, or both. We already showed on state-of-the-art methods, MI-GRAAL and IsoRankN, that combining NCF of one method and AS of another method can give a new superior method. Here, we evaluate MI-GRAAL against a newer approach, GHOST, by mixing-and-matching the methods' NCFs and ASs to potentially further improve alignment quality. While doing so, we approach important questions that have not been asked systematically thus far. First, we ask how much of the NCF information should come from protein sequence data compared to network topology data. Existing methods determine this parameter more-less arbitrarily, which could affect alignment quality. Second, when topological information is used in NCF, we ask how large the size of the neighborhoods of the compared nodes should be. Existing methods assume that the larger the neighborhood size, the better. Our findings are as follows. MI-GRAAL's NCF is superior to GHOST's NCF, while the performance of the methods' ASs is data-dependent. Thus, for data on which GHOST's AS is superior to MI-GRAAL's AS, the combination of MI-GRAAL's NCF and GHOST's AS represents a new superior method. Also, which amount of sequence information is used within NCF does not affect alignment quality, while the inclusion of topological information is crucial for producing good alignments. Finally, larger neighborhood sizes are preferred, but often, it is the second largest size that is superior. Using this size instead of the largest one would decrease computational complexity. Taken together, our results represent general recommendations for a fair evaluation of network alignment methods and in particular of two-stage NCF-AS approaches.

Network alignment finds node correspondences across multiple networks, where the alignment accuracy is of crucial importance because of its profound impact on downstream applications. The vast majority of existing works focus on how to best utilize the topology and attribute information of the input networks as well as the anchor links when available. Nonetheless, it has not been well studied on how to boost the alignment performance through actively obtaining high-quality and informative anchor links, with a few exceptions. The sparse literature on active network alignment introduces the human in the loop to label some seed node correspondence (i.e., anchor links), which are informative from the perspective of querying the most uncertain node given few potential matchings. However, the direct influence of the intrinsic network attribute information on the alignment results has largely remained unknown. In this paper, we tackle this challenge and propose an active network alignment method (Attent) to identify the best nodes to query. The key idea of the proposed method is to leverage effective and efficient influence functions defined over the alignment solution to evaluate the goodness of the candidate nodes for query. Our proposed query strategy bears three distinct advantages, including (1) effectiveness, being able to accurately quantify the influence of the candidate nodes on the alignment results; (2) efficiency, scaling linearly with 15 − 17 × speed-up over the straight-forward implementation without any quality loss; (3) generality, consistently improving alignment performance of a variety of network alignment algorithms.

The burgeoning volume of graph data presents significant computational challenges in training graph neural networks (GNNs), critically impeding their efficiency in various applications. To tackle this challenge, graph condensation (GC) has emerged as a promising acceleration solution, focusing on the synthesis of a compact yet representative graph for efficiently training GNNs while retaining performance. Despite the potential to promote scalable use of GNNs, existing GC methods are limited to aligning the condensed graph with merely the observed static graph distribution. This limitation significantly restricts the generalization capacity of condensed graphs, particularly in adapting to dynamic distribution changes. In real-world scenarios, however, graphs are dynamic and constantly evolving, with new nodes and edges being continually integrated. Consequently, due to the limited generalization capacity of condensed graphs, applications that employ GC for efficient GNN training end up with sub-optimal GNNs when confronted with evolving graph structures and distributions in dynamic real-world situations. To overcome this issue, we propose open-world graph condensation (OpenGC), a robust GC framework that integrates structure-aware distribution shift to simulate evolving graph patterns and exploit the temporal environments for invariance condensation. This approach is designed to extract temporal invariant patterns from the original graph, thereby enhancing the generalization capabilities of the condensed graph and, subsequently, the GNNs trained on it. Furthermore, to support the periodic re-condensation and expedite condensed graph updating in life-long graph learning, OpenGC reconstructs the sophisticated optimization scheme with kernel ridge regression and non-parametric graph convolution, significantly accelerating the condensation process while ensuring the exact solutions. Extensive experiments on both real-world and synthetic evolving graphs demonstrate that OpenGC outperforms state-of-the-art (SOTA) GC methods in adapting to dynamic changes in open-world graph environments.

Traditional Graph Neural Network (GNN) learning patterns can only achieve optimal performance under the assumption of independent and identically distributed data. However, in real-world scenarios, numerous unpredictable factors give rise to the Out-of-Distribution (OOD) problem. Moreover, this distribution discrepancy may lead to unreliable predictions by GNNs models within unknown OOD domains. Therefore, this paper proposes a machine learning method to enhance the stability of GNN cross-site classification from the perspective of stable learning. The aim is to extract genuine causal features while eliminating spurious causal features. By introducing a feature sample weighting decorrelation technique in the random Fourier transform space and combining it with a baseline GNN model, a Stable-GNN model and a constrained sampling weight gradient update algorithm are designed. Its theoretical proof indicates that this algorithm can ensure the decrease in the loss, thus remedying the shortcomings of the original algorithm in weight updating. While guaranteeing predictive performance on the training distribution data, it also reduces prediction bias on data from unseen test distributions. Compared with the GNN model and the Stable-GNN model (S-GNN), the experimental results of the S-GNN model demonstrate that it not only surpasses the current state-of-the-art GNN models, but also offers a flexible framework for strengthening existing GNNs.

A realistic model extraction attack against graph neural networks

Graph Neural Networks (GNNs) have become indispensable tools for learning from graph-structured data, catering to various applications such as social network analysis and fraud detection for financial services. At the heart of these networks are the edges, which are crucial in guiding GNN models' predictions. In many scenarios, these edges represent sensitive information, such as personal associations or financial dealings, which require privacy assurance. However, their contributions to GNN model predictions may, in turn, be exploited by the adversary to compromise their privacy. Motivated by these conflicting requirements, this paper investigates edge privacy in contexts where adversaries possess only black-box access to the target GNN model, restricted further by access controls, preventing direct insights into arbitrary node outputs. Moreover, we are the first to extensively examine situations where the target graph continuously evolves—a common trait of many real-world graphs. In this setting, we present a range of attacks that leverage the message-passing mechanism of GNNs. We evaluated the effectiveness of our attacks using nine real-world datasets, encompassing both static and dynamic graphs, across four different GNN architectures. The results demonstrate that our attack outperforms existing methods across various GNN architectures, consistently achieving an F1 score of at least 0.8 in static scenarios. Furthermore, our attack retains robustness in dynamic graph scenarios, maintaining F1 scores up to 0.8, unlike previous methods that only achieve F1 scores around 0.2.