Fraud detection on streaming transaction graphs (STGs) faces challenges on the catastrophic forgetting of previously learned fraud patterns when adapting to evolving patterns. Although some Graph Continual Learning (GCL) approaches mitigate this issue by storing and revisiting historical samples, practical storage constraints prevent them from fully preserving previous patterns. In this work, we propose STG-DGR, a streaming GNN model with diffusion-based generative replay that generates synthetic samples to retain previously learned patterns without storing real samples. The generation of replay samples for STGs faces two key challenges: (1) Heterogeneity challenge of generating STG samples with discrete adjacency table, user features, transaction features, and transaction timestamps. (2) Dependency challenge of capturing bottom-up dependencies across layers in STG samples. To address these challenges, STG-DGR integrates two novel components: (1) a Computational Subgraph Processor (CSP) that transforms heterogeneous STG samples into well-organized hierarchical subgraphs, and (2) a Diffusion-based Subgraph Generator (DSG) that captures the bottom-up dependencies using a novel Transformer-based Hierarchical Denoising Network (THDN), and generates synthetic replay samples that preserve these dependencies. Extensive experiments on four streaming fraud detection datasets demonstrate STG-DGR's superiority in reducing forgetting and improving accuracy over nineteen state-of-the-art baselines.

In the rapidly evolving web-based financial ecosystem where digital banking services become critical infrastructure for underserved communities, credit card fraud disproportionately affects vulnerable populations relying on financial platforms. However, previous studies overlook extreme data scarcity conditions, particularly at small-to-medium web banks that serve as crucial gateways for vulnerable communities. This paper addresses the fundamental challenge of building inclusive and secure financial systems operable at true web scale. To overcome this deficiency, we propose a novel web-based fairness-aware federated fraud detection model, CLARF, which utilizes the designed privacy-enhanced representation fusion and fraud-aware contrastive learning modules to enhance detection performance under conditions of data scarcity and label imbalance. Furthermore, current federated fraud detection systems critically neglect vulnerability to backdoor attacks, where malicious actors can implant hidden triggers during model aggregation, compromising system integrity. We propose a novel dynamic web Min-Max adversarial game framework where attackers employ hybrid multi-stage reinforcement learning with multi-dimensional reward mechanisms to dynamically evolve triggers that achieve excellent tradeoff between stealthiness and effectiveness. Defender adapts a closed-loop Selection-Evaluation-Suppression framework where high-reliability clients are selected via Fisher information to carry out reverse trigger engineering. Then clients' confidence scores are calculated as weights to minimize Attack Success Rate (ASR) during aggregation. Extensive experiments on six financial fraud datasets demonstrate the superiority of CLARF model and Min-Max adversarial game paradigm compared with multiple SOTA models.

In the global financial market, assessing bank credit ratings is essential for evaluating financial health, managing risk, and safeguarding systemic stability. While risk can transmit rapidly within the interbank lending network, timely incorporation of the latest financial disclosures to update bank ratings is vital in the swiftly evolving financial markets. However, existing approaches primarily conduct credit rating tasks using end-to-end models trained on historical financial data, thereby overlooking the staggered timing of financial disclosure from banks. Limited excavation of the credit rating records and the temporal distribution shifts existed in different financial periods still pose challenges to improving the accuracy of the credit rating tasks. To address these challenges, in this work we propose a Dual Pairwise pre-training and prompt Tuning framework with Aligned Prototypes (DPTAP) for interbank credit rating, which enables dynamic credit updates. Specifically, the dual pairwise pre-training strategy allows the framework to capture direction and distance discrepancies between rating categories. To alleviate the adverse impact of temporal distribution shifts in quarters, the latest financial features are prompted to dynamically map the patterns of the corresponding banks in the last quarter. Furthermore, we integrate rating guides from two consecutive quarters into a set of aligned prototypes to enhance supervision during the prompt tuning process. We conducted extensive experiments on a real-world bank dataset globally in the latest 8 years. The results demonstrate the superiority of our proposed framework over various competitive models, highlighting its notable capabilities in early warning and risk contagion forecasting.

Spoofing detection in financial trading is crucial, especially for identifying complex behaviors such as conspiracy spoofing. Traditional machine-learning approaches primarily focus on isolated node features, often overlooking the broader context of interconnected nodes. Graph-based techniques, particularly Graph Neural Networks (GNNs), have advanced the field by leveraging relational information effectively. However, in real-world spoofing detection datasets, trading behaviors exhibit dynamic, irregular patterns. Existing spoofing detection methods, though effective in some scenarios, struggle to capture the complexity of dynamic and diverse, evolving inter-node relationships. To address these challenges, we propose a novel framework called the Generative Dynamic Graph Model (GDGM), which models dynamic trading behaviors and the relationships among nodes to learn representations for conspiracy spoofing detection. Specifically, our approach incorporates the generative dynamic latent space to capture the temporal patterns and evolving market conditions. Raw trading data is first converted into time-stamped sequences. Then we model trading behaviors using the neural ordinary differential equations and gated recurrent units, to generate the representation incorporating temporal dynamics of spoofing patterns. Furthermore, pseudo-label generation and heterogeneous aggregation techniques are employed to gather relevant information and enhance the detection performance for conspiratorial spoofing behaviors. Experiments conducted on spoofing detection datasets demonstrate that our approach outperforms state-of-the-art models in detection accuracy. Additionally, our spoofing detection system has been successfully deployed in one of the largest global trading markets, further validating the practical applicability and performance of the proposed method.

Financial time series prediction is an important and challenging data mining task for quantitative investment. The inherent non-linearity, high noise, and susceptibility to various factors, such as macroeconomic conditions and market sentiment in the stock market, increase the difficulty of prediction. Existing financial industries mainly employ time series models or fundamental analysis methods for prediction. However, these methods fail to effectively capture the complex interrelationships between equity. In recent years, graph neural networks (GNNs), due to their powerful relational modeling capabilities, have been applied to stock prediction. However, with the advances of recent digital power, such as widely-used high-frequency trading techniques, existing graph-based methods still have shortcomings in effectively learning multi-granularity temporal relations as they cannot effectively learn the patterns in different frequencies, e.g., minute-level, daily, weekly, etc. Therefore, in this paper, we propose a multi-granularity graph augmented learning framework for interrelated financial time series forecasting. We first construct a temporal return relationship graph with multi-granularity financial time series, including weekly, daily, and minute-level, to comprehensively capture the dynamic relations of equities, including both medium-term trends and short-term fluctuations. Then, to further augment the node relations, we devise an attentional graph augment module to improve the graph learning with fundamental data, which are jointly optimized in the prediction layer. We conduct extensive empirical studies on multiple datasets from both the Chinese and U.S. stock markets. The results demonstrate that our proposed model consistently outperforms existing baseline methods across four key financial metrics, including ARR, ASR, CR, and IR, thereby validating its effectiveness and superiority. The model has been applied and empirically tested in commercial-grade trading platforms, further demonstrating its efficiency and robustness in real-world trading environments.

Graph neural networks (GNNs) are playing exciting roles in the application scenarios where features are hidden in information associations. Fraud prediction of online credit loan services (OCLSs) is such a typical scenario. But it has another rather critical challenge, i.e., the scarcity of data labels. Fortunately, GNNs can also cope with this problem due to their good ability of semi-supervised learning by mining structure and feature information within graphs. Nevertheless, the gain of internal information is often too limited to help GNNs handle the extreme deficiency of labels with high performance beyond the basic requirement of fraud prediction in OCLSs. Therefore, adding labels from the experts, such as manually adding labels through rules, has become a logical practice. However, the existing rule engines for OCLSs have the confliction problem among continuously accumulated rules. To address this issue, we propose a Snorkel-based Semi-Supervised GNN (S3GNN). Under S3GNN, we specially design an upgraded version of the rule engines, called Graph-Oriented Snorkel (GOS), a graph-specific extension of Snorkel, a widely used weakly supervised learning framework, to design rules by subject matter experts (SMEs) and resolve confliction. In particular, in the graph of an anti-fraud scenario, each node pair may have multiple different types of edges, so we propose the Multiple Edge-Types Based Attention mechanism. In general, for the heterogeneous information and multiple relations in the graph, we first obtain the embedding of applicant nodes by aggregating the representation of attribute nodes, and then use the attention mechanism to aggregate neighbor nodes on multiple meta-paths to get ultimate applicant node embedding. We conduct experiments over the real-life data of a large financial platform. The results demonstrate that S3GNN can outperform the state-of-the-art methods, including the method of pilot platform.

In networked-loans, guarantor enterprises have a legal duty to repay debt to the commercial bank when the guaranteed borrower enterprise defaults (fail to repay). During an economic recession, the risk of defaults may spread like wildfire -- the loan network structure could amplify both reach and impact; thus leading to a large-scale corporation defaults even systemic financial crises. The Central Bank urges advanced regulation technology to recognize and act on the contagion risk in order to avoid the ‘`gray rhino’'. Therefore, we present a novel approach to help the regulators quantify the systemic risk and provide stemming clues. In particular, we report a state-of-the-art graph neural network architecture (iConReg) for detecting and isolating of contagion risk in China's national-wide networked-loans. The overall accuracy of our model reaches over 91% of AUC, which considerably outperforms the compared benchmark methods. By isolating the top 1% of predicted high-risky nodes in the contagion chains, iConReg reports a significant shrink (averaged 25.8%) of loan default rates. Our presented approach opens up a new direction of using deep graph learning techniques to regulate the contagion risk of networked-loans, which enables the authorities to design more prompt prevention measures against systemic financial crises

Anti-money laundering (AML) is a classical data mining problem in finance applications. As well known, money laundering (ML) is critical to the effective operation of transnational and organized crime, which affects a country's economy, government, and social wellbeings. Financial services organizations facilitate the movement of money and have been enlisted by governments to assist with the detection and prevention of money laundering, which is a key tool in the fight to reduce crime and create sustainable economic development. In the application of AML, user identity and financial behavior data are widely used to detect laundering transactions. In recent years, an increasing number of money laundering activities have been conducted by organized criminal gangs while most existing works still treat the actions of each account as independent identity behavior without considering the group-level conspired interactions. Therefore, in this paper, we propose a group-aware deep graph learning-based approach for organized money-laundering detection. In particular, we design a community-centric encoder to represent the nodes and attributes in user transaction graphs and derive the adjacent gang behaviors. Then, we devise a scheme of local enhancement to accommodate nodes with similar transaction features, which are aggregated into gangs for downstream detection. Extensive experiments on the real-world dataset from one of the largest bank card alliances worldwide show that our proposed method outperforms state-of-the-art methods in both offline and online modes, showing the effectiveness of money laundering detection with group-aware deep graph learning.

As a typical latent factor model, Matrix Factorization (MF) has demonstrated its great effectiveness in recommender systems. Users and items are represented in a shared low-dimensional space so that the user preference can be modeled by linearly combining the item factor vector$V$using the user-specific coefficients$U$. From a generative model perspective,$U$and$V$are drawn from twoindependentGaussian distributions, which is not so faithful to the reality. Items are produced to maximally meet users’ requirements, which makes$U$and$V$strongly correlated. Meanwhile, the linear combination between$U$and$V$forces a bijection (one-to-one mapping), which thereby neglects the mutual correlation between the latent factors. In this paper, we address the upper drawbacks, and propose a new model, named Correlated Matrix Factorization (CMF). Technically, we apply Canonical Correlation Analysis (CCA) to map$U$and$V$into a new semantic space. Besides achieving the optimal fitting on the rating matrix, one component in each vector ($U$or$V$) is also tightly correlated with every single component in the other. We derive efficient inference and learning algorithms based on variational EM methods. The effectiveness of our proposed model is comprehensively verified on four public datasets. Experimental results show that our approach achieves competitive performance on both prediction accuracy and efficiency compared with the current state of the art.

Document network is a kind of intriguing dataset which can provide both topical (textual content) and topological (relational link) information. A key point in modeling such datasets is to discover proper denominators beneath the text and link. Most previous work introduces the assumption that documents closely linked with each other share common latent topics. However, the heterophily (i.e., tendency to link to different others) of nodes is neglected, which is pervasive in social networks. In this paper, we simultaneously incorporate community detection and topic modeling in a unified framework, and appeal to Canonical Correlation Analysis (CCA) to capture the latent semantic correlations between the two heterogeneous factors, community and topic. Despite of the homophily (i.e., tendency to link to similar others) or heterophily, CCA can properly capture the inherent correlations which fit the dataset itself without any prior hypothesis. We also impose auxiliary word embeddings to improve the quality of topics. The effectiveness of our proposed model is comprehensively verified on three different types of datasets which are hyperlinked networks of web pages, social networks of friends, and coauthor networks of publications. Experimental results show that our approach achieves significant improvements compared with the current state of the art.

The soaring of social media services has greatly propelled the prevalence of document networks. Rather than a set of plain texts, documents are nodes in graphs. An observable link connects the documents at its two ends, thus it implicitly reflects the semantic association between the document pair. Previous work assumes that only similar documents tend to be connected, which neglects the rich connective patterns in the topological structure. In this paper, we introduce a latent correlation factor to categorize the links into several categories, and each category corresponds to a unique kind of association. By fitting the data, the relational information (e.g., homophily and heterophily) can be comprehensively captured. By resorting to Canonical Correlation Analysis (CCA), we maximize the correlation between all pairs of linked documents. We propose a pure generative model and derive efficient learning algorithms based on the variational EM methods. Experiments on three different datasets demonstrate that the proposed model is competitive and usually better than the state-of-the-art baselines on both topic modeling and link prediction.

Document network is a kind of intriguing dataset which provides both topical (texts) and topological (links) information. Most previous work assumes that documents closely linked with each other share common topics. However, the associations among documents are usually complex, which are not limited to the homophily (i.e., tendency to link to similar others). Actually, the heterophily (i.e., tendency to link to different others) is another pervasive phenomenon in social networks. In this paper, we introduce a new tool, called copula, to separately model the documents and links, so that different copula functions can be applied to capture different correlation patterns. In statistics, a copula is a powerful framework for explicitly modeling the dependence of random variables by separating the marginals and their correlations. Though widely used in Economics, copulas have not been paid enough attention to by researchers in machine learning field. Besides, to further capture the potential associations among the unconnected documents, we apply the tree-averaged copula instead of a single copula function. This improvement makes our model achieve better expressive power, and also more elegant in algebra. We derive efficient EM algorithms to estimate the model parameters, and evaluate the performance of our model on three different datasets. Experimental results show that our approach achieves significant improvements on both topic and link modeling compared with the current state of the art.

It is a valuable study for Location-based Social Network (LBSN) make a more accurate Points-of-Interest (POI) recommendation since that can improve users' experiences. There have been many methods of POIs recommendation that consider context, personal preference pattern, and/or matrix factorization. However, the continuous contexts have not been thoroughly considered in these methods. This paper first proposes a locations splitting method which can handle both continuous and discrete contexts. Moreover, we present a Context-aware Probabilistic Matrix Factorization method (CPMF) that factorizes a frequency matrix of contexts and locations to obtain the user-location checkin probabilities. We design a Personal Preference Confidence (PPC) to extract a set of reliable POIs with confidence values for every user. Finally, we propose a hybrid recommender which fuses CPMF with PPC to recommend top-n POIs. Experiments on a large-scale real-world checkins dataset demonstrate that our recommendation method obtains a well performance and effect.

Document network is a kind of intriguing dataset which can provide both topical (textual content) and topological (relational link) information. A key point in viably modeling such datasets is to discover proper denominators beneath the two different types of data, text and link. Most previous work introduces the assumption that documents closely linked with each other share common latent topics. However, the heterophily (i.e., tendency to link to different others) of nodes is neglected, which is pervasive in social networks. In this paper, we simultaneously incorporate community detection and topic modeling in a unified framework, and appeal to Canonical Correlation Analysis (CCA) to capture the latent semantic correlations between the two heterogeneous latent factors, community and topic. Despite of the homophily (i.e., tendency to link to similar others) or heterophily, CCA can properly capture the inherent correlations which fit the dataset itself without any prior hypothesis. Logistic normal prior is also employed in modeling network to better capture the community correlations. We derive efficient inference and learning algorithms based on variational EM methods. The effectiveness of our proposed model is comprehensively verified on three different types of datasets which are namely hyperlinked networks of web pages, social networks of friends and coauthor networks of publications. Experimental results show that our approach achieves significant improvements on both topic modeling and community detection compared with the current state of the art. Meanwhile, our model is impressive in discovering correlations between extracted topics and communities.