Illegal related-party transactions (RPT) are federal felonies that pose a severe threat to the stability and integrity of modern financial systems. The increasing frequency of RPTs forms complex and dynamic networks. Existing temporal graph learning methods tend to treat entities as functionally homogeneous, ignoring the diverse and evolving structural roles of nodes. Role-based embedding methods model global structure by bridging same-role nodes, but their reliance on a unified mechanism for aggregation and evolution means they fail to distinguish the underlying logic of distinct interactions governed by structural roles. The limitations motivate us to develop a customized role-based strategy. It can also adapt to evolving RPT dynamics, thereby forming a continuous regulatory process to combat illegal activities. In this paper, we propose an innovative Role Perceptual Augmented Temporal Graph Network (RPATGN) for proactive RPT detection. We analyze the structural roles of nodes and employ a role-based message passing mechanism that adapts its aggregation strategy based on the roles of interacting nodes. We integrate a variational graph recurrent neural network, enhanced by temporal contextual attention, to explicitly model the dynamics of the roles and the overall network evolution. Extensive experiments on real-world financial datasets demonstrate the effectiveness of our approach for RPT detection. It holds practical significance for fostering robust financial systems and promoting healthy, transparent financial markets.

Graph Neural Networks (GNNs) have achieved notable success in a wide range of applications. However, their vulnerability to adversarial attacks, particularly graph injection attacks (GIAs), raises serious concerns for their deployment in security-sensitive domains. Existing GIA methods, despite their demonstrated effectiveness, face several inherent limitations. They typically require training surrogate models to approximate the victim model's behavior, which may lead to performance degradation when the surrogate mismatches the target model. Furthermore, the discrete nature of graph data poses challenges for generating effective adversarial features, often resulting in suboptimal solutions. Most critically, these methods show markedly reduced effectiveness when deployed against defended GNN models, limiting real-world applicability. To address these challenges, we introduce SoftHist, a novel gradient-free reinforcement learning framework for black-box graph injection attacks. Our approach incorporates a softened embedding mechanism to avoid suboptimal feature generation, ensuring stable and stealthy node injection. Moreover, we design a topology-aware edge sampler and a defense-aware policy learner with adaptive history reuse optimized for misclassification maximization. These innovations collectively balance attack effectiveness, stealthiness, and robustness against defensive measures. Extensive experiments on eight benchmark datasets demonstrate SoftHist's significant advantages in key scenarios: (1) On discrete-feature datasets like AMComputer, the misclassification rate is 10.34%~38.09% higher than baseline methods; (2) Against defensive models such as RGCN, it maintains 98.23% success rate, surpassing state-of-the-art methods by 12.36%.

Accurate enterprise power consumption forecasting is not only a core component of optimized green energy management but also a key support for promoting the coordinated development of a sustainable society and the digital economy. The temporal fluctuations in power consumption reflect an enterprise's production activity and operational resilience, while credit assessment combined with Web data reveals a two-way coupling relationship between it and energy use: credit changes influence financing and power consumption strategies, while energy anomalies may become early signals of credit risk. However, existing methods still have shortcomings in modeling the co-evolution of Web data and power data. Most models only focus on static or unidirectional correlations, making it difficult to capture the dynamic feedback between credit risk and power consumption; traditional multi-task learning frameworks often rely on parameter sharing or simple attention mechanisms, lacking consistency constraints across time scales and network structures. To address this, this paper proposes CPDGL, a credit-electricity co-evolution framework based on dynamic graph learning, which simultaneously performs power forecasting and credit risk assessment within a unified multi-task system. Its co-evolution path interaction module explicitly models the feedback loop between credit dynamics and power behavior, learning bidirectional causal relationships through an adaptive influence matrix; the semantic path aggregation module integrates static and dynamic features, strengthening cross-modal expression and global reasoning capabilities. Large-scale experiments conducted in a real-world enterprise environment of one of the world's largest power suppliers demonstrate that CPDGL achieves state-of-the-art performance in both power forecasting and credit assessment tasks. The results validate its broad applicability in multi-source Web data fusion scenarios, significantly improving forecasting accuracy and dispatch efficiency in clean energy management, and showcasing practical value and social impact in smart cities and sustainable development.

Graph Neural Networks (GNNs) have achieved remarkable success in machine learning tasks by learning the features of graph data. However, experiments show that vanilla GNNs fail to achieve good classification performance in the field of graph anomaly detection. To address this issue, we propose and theoretically prove that the high-Class Homophily Variance (CHV) characteristic is the reason behind the suboptimal performance of GNN models in anomaly detection tasks. Statistical analysis shows that in most standard node classification datasets, homophily levels are similar across all classes, so CHV is low. In contrast, graph anomaly detection datasets have high CHV, as benign nodes are highly homophilic while anomalies are not, leading to a clear separation. To mitigate its impact, we propose a novel GNN model named Homophily Edge Augment Graph Neural Network (HEAug). Different from previous work, our method emphasizes generating new edges with low CHV value, using the original edges as an auxiliary. HEAug samples homophily adjacency matrices from scratch using a self-attention mechanism, and leverages nodes that are relevant in the feature space but not directly connected in the original graph. Additionally, we modify the loss function to punish the generation of unnecessary heterophilic edges by the model. Extensive comparison experiments demonstrate that HEAug achieved the best performance across eight benchmark datasets, including anomaly detection, edgeless node classification and adversarial attack. We also defined a heterophily attack to increase the CHV value in other graphs, demonstrating the effectiveness of our theory and model in various scenarios.

With growing demands for data privacy and model robustness, graph unlearning (GU), which erases the influence of specific data on trained GNN models, has gained significant attention. However, existing exact unlearning methods suffer from either low efficiency or poor model performance. While more utility-preserving and efficient, current approximate methods require access to the forget set during unlearning, which makes them inapplicable in immediate deletion scenarios, thereby undermining privacy. Additionally, these approximate methods, which attempt to directly perturb model parameters, still raise significant concerns regarding unlearning power in empirical studies. To fill the gap, we propose Transferable Condensation Graph Unlearning (TCGU), a data-centric solution to graph unlearning. Specifically, we first develop a two-level alignment strategy to pre-condense the original graph into a compact yet utility-preserving dataset for subsequent unlearning tasks. Upon receiving an unlearning request, we fine-tune the pre-condensed data with a low-rank plugin, to directly align its distribution with the remaining graph, thus efficiently revoking the information of deleted data without accessing them. A novel similarity distribution matching approach and a discrimination regularizer are proposed to effectively transfer condensed data and preserve its utility in GNN training, respectively. Finally, we retrain the GNN on the transferred condensed data. Extensive experiments on 7 benchmark datasets demonstrate that TCGU can achieve superior performance in terms of model utility, unlearning efficiency, and unlearning efficacy compared to existing GU methods. To the best of our knowledge, this is the first study to explore graph unlearning with immediate data removal using a data-centric approximate method.

Self-supervised learning (SSL) provides a promising paradigm for hypergraph representation learning without reliance on costly labels. However, existing hypergraph SSL methods predominantly employ contrastive learning with instance level discrimination, encountering two significant challenges: (1) Unreliable negative sampling, where arbitrarily selected negative samples introduce bias by misclassifying similar and dissimilar pairs; and (2) High computational cost, as effective training requires a large number of negative samples. To address these limitations, we propose SE-HSSL, a hypergraph SSL frame work with three sampling-efficient self-supervised objectives. Specifically, two sampling-free objectives based on canonical correlation analysis serve as node- and group-level signals, while a hierarchical membership-level contrastive objective exploits the cascading overlap structure of hypergraphs. Overall, these designs mitigate negative sampling bias and enhance sampling efficiency, leading to better downstream performance and faster training. Beyond these challenges, deep hypergraph models are prone to biased predictions against groups defined by sensitive attributes (e.g., gender and race). To address fairness concerns, we first theoretically show that imbalanced contributions of demographic groups during hypergraph message passing amplify sensitive biases in the training data. Motivated by this, we propose FairHSSL, a fairness-aware variant of SE-HSSL equipped with a two-level debiasing augmentation strategy. To generate a fair hypergraph view, the augmentation integrates two complementary components: feature-level debiasing and structure-level perturbation. Specifically, orthogonal projection is applied at the feature level to decouple node features from sensitive attributes, while rebalance-based perturbation is introduced at the structure level to equalize the contributions of different sensitive groups during message passing. Finally, by aligning the fair augmented and original biased views in SSL, we mitigate the influence of sensitive information. Extensive experiments on 10 real-world hypergraphs demonstrate the superior effectiveness and efficiency of SE-HSSL. Moreover, FairHSSL consistently outperforms state of-the-art (SOTA) baselines in terms of utility-fairness trade-off across all datasets.

Fraud detection is a typical data mining mission in the field of finance. In recent years, due to their capability of mining hidden associations between entities, graph neural networks (GNNs) have been widely applied to detect financial fraudsters. However, GNNs are fragile in their data aggregation process and will be attacked on purpose by fraudsters, therefore, some other pioneers have explored methods to enhance the robustness of GNN-based fraud detection models. But most existing models based on ideal settings, as real-life criminals tend to attack as far as they can reach, struggle to establish a unified effective approach for attacked and unattacked data of different scales in realistic scenarios. Furthermore, mainstream robust defense models indiscriminately modifying and truncating data will lose important information of the major unattacked parts in the original graph, which lowers their overall fraud detection precision. Therefore, in this work, we propose a novel generative fraud detection framework called GSRGNN. In particular, we first design a generative structure to obtain augmented node features. Then we prioritize the nodes with high degrees to create a more stable graph structure on local distributions. Finally, we pass the enhanced features and structure with the input ones in pairs through GNN layers, and create synthetic representations with abundant information and sufficient resistance to perturbations for subsequent fraud detection. In addition, we also design a novel black-box attack algorithm to realistically imitate the perturbations conducted by fraudsters on graph features and structure. Experiments on the world's leading electronic trading platform and public anti-fraud datasets demonstrate the outstanding performance of our proposed method compared with those state-of-the-art models, showing its superiority in precision and robustness on financial fraud detection missions.

Safety analysis is a crucial approach for improving safety in the aircraft system design process. Traditional methods for safety analysis often lag behind the design process and rely heavily on engineers' expertise, limiting their effectiveness for complex aircraft systems. This paper proposes an advanced safety analysis method for airborne systems, integrating formal verification with the Systems Modeling Language (SysML). A detailed SysML-based system model is constructed to align safety analysis with system design seamlessly. Additionally, We introduce a novel coding approach that translates SysML state machine models into symbolic Kripke structures, utilizing Boolean functions to implicitly represent graphical system behavior. To mitigate the state explosion issue common in complex systems, an improved reachability analysis algorithm utilizing partition sorting is proposed. Building on this foundation, symbolic model checking techniques are employed to automate safety analysis, including the verification of safety requirements and the derivation of minimal cut sets for airborne systems. The proposed method is validated with case studies on an electro-mechanical actuator and a flight control system. Compared to existing methods, this approach ensures consistency between safety and system design models, eliminates complex transformations between modeling languages, and improves both the efficiency and accuracy of the analysis.

Transformer-based and MLP-based methods have emerged as leading approaches in time series forecasting (TSF). However, real-world time series often show different patterns at different scales, and future changes are shaped by the interplay of these overlapping scales, requiring high-capacity models. While Transformer-based methods excel in capturing long-range dependencies, they suffer from high computational complexities and tend to overfit. Conversely, MLP-based methods offer computational efficiency and adeptness in modeling temporal dynamics, but they struggle with capturing temporal patterns with complex scales effectively. Based on the observation of multi-scale entanglement effect in time series, we propose a novel MLP-based Adaptive Multi-Scale Decomposition (AMD) framework for TSF. Our framework decomposes time series into distinct temporal patterns at multiple scales, leveraging the Multi-Scale Decomposable Mixing (MDM) block to dissect and aggregate these patterns. Complemented by the Dual Dependency Interaction (DDI) block and the Adaptive Multi-predictor Synthesis (AMS) block, our approach effectively models both temporal and channel dependencies and utilizes autocorrelation to refine multi-scale data integration. Comprehensive experiments demonstrate our AMD framework not only overcomes the limitations of existing methods but also consistently achieves state-of-the-art performance across various datasets.

Graph Neural Networks (GNNs) demonstrate superior performance in various graph learning tasks, yet their wider real-world application is hindered by the computational overhead when applied to large-scale graphs. To address the issue, the Graph Lottery Hypothesis (GLT) has been proposed, advocating the identification of subgraphs and subnetworks, i.e., winning tickets, without compromising performance. The effectiveness of current GLT methods largely stems from the use of iterative magnitude pruning (IMP), which offers greater stability and better performance than one-shot pruning. However, identifying GLTs is highly computationally expensive, due to the iterative pruning and retraining required by IMP. In this paper, we reevaluate the correlation between one-shot pruning and IMP: while one-shot tickets are suboptimal compared to IMP, they offer a fast track to tickets with a stronger performance. We introduce a one-shot pruning and denoising framework to validate the efficacy of the fast track. Compared to current IMP-based GLT methods, our framework achieves a double-win situation of graph lottery tickets with higher sparsity and faster speeds. Through extensive experiments across 4 backbones and 6 datasets, our method demonstrates a 1.32%-45.62% improvement in weight sparsity and a 7.49%-22.71% increase in graph sparsity, along with a 1.7-44× speedup over IMP-based methods and 95.3%-98.6% MAC savings.

Yanwei Yue, Guibin Zhang, Boyang Liu, Guancheng Wan, Kun Wang, Dawei Cheng, Yiyan Qi. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Nowadays, as more users turn to WeChat Pay and other e-commerce platforms for transactions, an increasing number of fraudsters are being attracted to these platforms to conduct fraudulent activities, thereby stealing money. To address this issue, Graph Neural Networks (GNNs) have been widely adopted and have shown great success. However, with the rise of various transaction methods, users are increasingly engaging in multiple transaction networks, which creates a new scenario that requires models to detect fraud across these diverse networks. Unfortunately, current GNN-based fraud detection strategies often exhibit suboptimal performance and high time complexity in this evolving scenario, as they typically can handle only one transaction network at a time. Recently, advancements in graph prompt learning have demonstrated great success in managing various types of graph data and improving the generalization capabilities of the model, showing great promise for addressing this new fraud detection scenario. Nevertheless, the practical application of graph prompt learning in real-world fraud detection is still constrained, as they may exhibit bias when dealing with multiplex transaction networks and may fail to model the intrinsic relationships between nodes and their neighbors, which is crucial for effective fraud detection. To address these two challenges, we propose GPCF, an efficient graph pre-training and prompt learning framework. GPCF first incorporates a meta-learning-based strategy within neighbor-enhanced contrastive learning to pre-train the GNN model across diverse transaction networks. Then it aligns fraud detection tasks with the well-pre-trained model by simply fine-tuning the prompts. Extensive experiments demonstrate that GPCF achieves state-of-the-art results on open-access fraud and transaction datasets, as well as on real-world fraud datasets from WeChat Pay, one of the largest e-commerce platforms globally, showing the effectiveness of GPCF in practical applications.

Time series forecasting has witnessed significant advancements through deep learning techniques. However, most existing methods struggle in non-stationary environments, where data distributions evolve over time due to concept drift. To address the challenge of non-stationarity in time series, various stabilization techniques have been proposed to mitigate temporal variations. Nonetheless, these methods operate at the instance level, assuming a homogeneous distribution across all time steps within an instance and relying on fixed statistical normalization. This limits their ability to effectively capture fine-grained distributional shifts.

This paper proposes a novel approach to predicting stock returns based on Relational Temporal Graph Neural Networks (RTGNN). This method provides insights into the dynamic and asymmetric relationships that exist both among different stocks and within individual stocks over time, addressing the limitations of existing methods in modeling asymmetric relationships, as well as the neglect of market dynamics. The proposed method employs a temporal attention mechanism to process continuous financial data, with the objective of adaptively constructing asymmetric graphs of the dynamics of the stock relationship based on historical stock data and changes in momentum of each stock. The experimental results, which were performed on the NASDAQ 100 and CSI 300 datasets, demonstrate that the proposed model outperforms traditional methods in terms of investment returns and information coefficients. In the 2023 investment backtesting experiment, the proposed methodology exhibited superior performance, achieving over 35% excess return in both datasets. Our sources are released at https://github.com/tj-lg100/RTGNN.

Data generation is a fundamental research problem in data management due to its diverse use cases, ranging from testing database engines to data-specific applications. However, real-world entities often involve complex interactions that cannot be effectively modeled by traditional tabular data. Therefore, graph data generation has attracted increasing attention recently. Although various graph generators have been proposed in the literature, there are three limitations: i) They cannot capture the co-evolution pattern of graph structure and node attributes. ii) Few of them consider edge direction, leading to substantial information loss. iii) Current state-of-the-art dynamic graph generators are based on the temporal random walk, making the simulation process time-consuming. To fill the research gap, we introduce VRDAG, a novel variational recurrent framework for efficient dynamic attributed graph generation. Specifically, we design a bidirectional message-passing mechanism to encode both directed structural knowledge and attribute information of a snapshot. Then, the temporal dependency in the graph sequence is captured by a recurrence state updater, generating embeddings that can preserve the evolution pattern of early graphs. Based on the hidden node embeddings, a conditional variational Bayesian method is developed to sample latent random variables at the neighboring timestep for new snapshot generation. The proposed generation paradigm avoids the time-consuming path sampling and merging process in existing random walk-based methods, significantly reducing the synthesis time. Finally, comprehensive experiments on real-world datasets are conducted to demonstrate the effectiveness and efficiency of the proposed model.

As artificial intelligence technology continues to advance, artificial intelligence-generated content (AIGC) has begun to evolve towards generating complex and structured data, particularly graph data. As an important topic in many fields such as database, data mining, and machine learning, graph generation holds significant value for simulating complex relationships between entities and has shown vast potential for applications in fields such as molecular generation, drug design, and material discovery. In this context, AIGC technology for graph generation has received widespread attention. This tutorial outlines the latest developments in AIGC for graph generation. We categorize existing methods into two main types according to their objectives and motivations: similarity-based generation and function-driven generation. We first provide an overview of AIGC models for graph generation. Then, we conduct a thorough review of the existing works. Finally, we explore the current trends and future directions, discussing potential ways to integrate database and machine learning techniques for graph generation.

Graph simulation has recently received a surge of attention in graph processing and analytics. In real-life applications, e.g. social science, biology, and chemistry, many graphs are composed of a series of evolving graphs (i.e., temporal graphs). While most of the existing graph generators focus on static graphs, the temporal information of the graphs is ignored. In this paper, we focus on simulating temporal graphs, which aim to reproduce the structural and temporal properties of the observed real-life temporal graphs. In this paper, we first give an overview of the existing temporal graph generators, including recently emerged learning-based approaches. Most of these learning-based methods suffer from one of the limitations: low efficiency in training or slow generating, especially for temporal random walk-based methods. Therefore, we propose an efficient learning-based approach to generate graph snapshots, namely temporal graph autoencoder (TGAE). Specifically, we propose an attention-based graph encoder to encode temporal and structural characteristics on sampled ego-graphs. And we proposed an ego-graph decoder that can achieve a good trade-off between simulation quality and efficiency in temporal graph generation. Finally, the experimental evaluation is conducted among our proposed TGAE and representative temporal graph generators on real-life temporal graphs and synthesized graphs. It is reported that our proposed approach outperforms the state-of-the-art temporal graph generators by means of simulation quality and efficiency.

Graph data augmentation (GDA) has shown significant promise in enhancing the performance, generalization, and robustness of graph neural networks (GNNs). However, contemporary methodologies are often limited to static graphs, whose applicability on dynamic graphs—more prevalent in real-world applications—remains unexamined. In this paper, we empirically highlight the challenges faced by static GDA methods when applied to dynamic graphs, particularly their inability to maintain temporal consistency. In light of this limitation, we propose a dedicated augmentation framework for dynamic graphs, termed $\texttt{DyAug}$, which adaptively augments the evolving graph structure with temporal consistency awareness. Specifically, we introduce the paradigm of graph rationalization for dynamic GNNs, progressively distinguishing between causal subgraphs (\textit{rationale}) and the non-causal complement (\textit{environment}) across snapshots. We develop three types of environment replacement, including, spatial, temporal, and spatial-temporal, to facilitate data augmentation in the latent representation space, thereby improving the performance, generalization, and robustness of dynamic GNNs. Extensive experiments on six benchmarks and three GNN backbones demonstrate that $\texttt{DyAug}$ can \textbf{(I)} improve the performance of dynamic GNNs by $0.89\\%\sim3.13\\%\uparrow$; \textbf{(II)} effectively counter targeted and non-targeted adversarial attacks with $6.2\\%\sim12.2\\%\\uparrow$ performance boost; \textbf{(III)} make stable predictions under temporal distribution shifts.

Recent advancements in large language model (LLM)-powered agents have shown that collective intelligence can significantly outperform individual capabilities, largely attributed to the meticulously designed inter-agent communication topologies. Though impressive in performance, existing multi-agent pipelines inherently introduce substantial token overhead, as well as increased economic costs, which pose challenges for their large-scale deployments. In response to this challenge, we propose an economical, simple, and robust multi-agent communication framework, termed $\texttt{AgentPrune}$, which can seamlessly integrate into mainstream multi-agent systems and prunes redundant or even malicious communication messages. Technically, $\texttt{AgentPrune}$ is the first to identify and formally define the $\textit{Communication Redundancy}$ issue present in current LLM-based multi-agent pipelines, and efficiently performs one-shot pruning on the spatial-temporal message-passing graph, yielding a token-economic and high-performing communication topology. Extensive experiments across six benchmarks demonstrate that $\texttt{AgentPrune}$ $\textbf{(I)}$ achieves comparable results as state-of-the-art topologies at merely $\\$5.6$ cost compared to their $\\$43.7$, $\textbf{(II)}$ integrates seamlessly into existing multi-agent frameworks with $28.1\\%\sim72.8\\%\downarrow$ token reduction, and $\textbf{(III)}$ successfully defend against two types of agent-based adversarial attacks with $3.5\\%\sim10.8\\%\uparrow$ performance boost. The source code is available at \url{https://github.com/yanweiyue/AgentPrune}.

Time series forecasting methods generally fall into two main categories: Channel Independent (CI) and Channel Dependent (CD) strategies. While CI overlooks important covariate relationships, CD captures all dependencies without distinction, introducing noise and reducing generalization. Recent advances in Channel Clustering (CC) aim to refine dependency modeling by grouping channels with similar characteristics and applying tailored modeling techniques. However, coarse-grained clustering struggles to capture complex, time-varying interactions effectively. To address these challenges, we propose TimeFilter, a GNN-based framework for adaptive and fine-grained dependency modeling. After constructing the graph from the input sequence, TimeFilter refines the learned spatial-temporal dependencies by filtering out irrelevant correlations while preserving the most critical ones in a patch-specific manner. Extensive experiments on 13 real-world datasets from diverse application domains demonstrate the state-of-the-art performance of TimeFilter. The code is available at https://github.com/TROUBADOUR000/TimeFilter.

Recent advancements in large language model (LLM)-based agents have demonstrated that collective intelligence can significantly surpass the capabilities of individual agents, primarily due to well-crafted inter-agent communication topologies. Despite the diverse and high-performing designs available, practitioners often face confusion when selecting the most effective pipeline for their specific task: Which topology is the best choice for my task, avoiding unnecessary communication token overhead while ensuring high-quality solution? In response to this dilemma, we introduce G-Designer, an adaptive, efficient, and robust solution for multi-agent deployment, which dynamically designs task-aware, customized communication topologies. Specifically, G-Designer models the multi-agent system as a multi-agent network, leveraging a variational graph auto-encoder to encode both the nodes (agents) and a task-specific virtual node, and decodes a task-adaptive and high-performing communication topology. Extensive experiments on six benchmarks showcase that G-Designer is: (1) high-performing, achieving superior results on MMLU with accuracy at $84.50\%$ and on HumanEval with pass@1 at $89.90\%$; \textbf{(2) task-adaptive}, architecting communication protocols tailored to task difficulty, reducing token consumption by up to $95.33\%$ on HumanEval; and \textbf{(3) adversarially robust}, defending against agent adversarial attacks with merely $0.3\%$ accuracy drop.

The outbreak of pandemic has a huge impact on production and consumption in the business world, especially for the retail sector. As a crucial component of decision-support technology in the retail industry, sales forecasting is significant for production planning and optimizing the supply of essential goods during the pandemic. However, due to the irregular fluctuation pattern caused by uncertainty and the complex temporal correlation between multiple covariates and sales, there is still no effective approach for sales forecasting in this extreme event. To fill this gap, we propose a Pandemic-Compatible Attentive Network (PCAN) for retail sales forecasting. Specifically, to capture the irregular fluctuation patterns from the sales series, we design a fluctuation attention mechanism based on association discrepancy in the time series. Then, a parallel attention module is developed to learn the complex relationship between target sales and various dynamic influence factors in a decoupled manner. Finally, we introduce a novel rectification decoding strategy to indicate fluctuation points in prediction. By evaluating PCAN on four real-world retail food datasets from the SF Express international supply chain system, the results show that our method achieves superior performance over the existing state-of-the-art baselines. The model has been deployed in the supply chain system as a fundamental component to serve a world-leading food retailer.

Portfolio optimization encounters persistent challenges in adapting to dynamic markets due to static assumptions and high-dimensional decision spaces. Although reinforcement learning (RL) has emerged as a potential solution, conventional reward engineering often fails to capture complex market dynamics. Recent advances in deep RL and graph neural networks have attempted to enhance market microstructure modeling. However, these methods still struggle with the systematic integration of financial knowledge. To address the above issues, we propose a novel heuristic-guided inverse reinforcement learning framework for portfolio optimization. Specifically, our framework provides an interpretable expert strategy generation mechanism that takes into account sector diversification and correlation constraints. Then, a multi-objective reward optimization method is adopted to adaptively strike a balance between returns and risks. Furthermore, it also utilizes heterogeneous graph policy learning with hierarchical attention mechanisms to explicitly model inter-stock relationships. Finally, we conduct extensive experiments on real-world financial market data to demonstrate that our framework outperforms several state-of-the-art deep learning and RL baselines in terms of risk-adjusted returns. We provide case studies to showcase the ability of our framework to balance return maximization and risk containment. Our code is publicly available at https://github.com/ChloeWenyiZhang/SmartFolio/.

Graph Neural Networks (GNNs) have achieved significant success in various real-world applications, including social networks, finance systems, and traffic management. Recent researches highlight their vulnerability to backdoor attacks in node classification, where GNNs trained on a poisoned graph misclassify a test node only when specific triggers are attached. These studies typically focus on single attack categories and use adaptive trigger generators to create node-specific triggers. However, adaptive trigger generators typically have a simple structure, limited parameters, and lack category-aware graph knowledge, which makes them struggle to handle backdoor attacks across multiple categories as the number of target categories increases. We address this gap by proposing a novel approach for Effective and Unnoticeable Multi-Category (EUMC) graph backdoor attacks, leveraging subgraph from the attacked graph as category-aware triggers to precisely control the target category. To ensure the effectiveness of our method, we construct a Multi-Category Subgraph Triggers Pool (MC-STP) using the subgraphs of the attacked graph as triggers. We then exploit the attachment probability shifts of each subgraph trigger as category-aware priors for target category determination. Moreover, we develop a ``select then attach'' strategy that connects suitable category-aware trigger to attacked nodes for unnoticeability. Extensive experiments across different real-world datasets confirm the efficacy of our method in conducting multi-category graph backdoor attacks on various GNN models and defense strategies.

Spoofing is a deceptive trading strategy where fraudsters place a large number of fake orders to manipulate market prices, severely distorting market fairness and threatening market stability.With the advancement of fraudulent tactics, spoofing patterns span across various levels of interaction, involving not only the local structure of individual spoofing transactions but also spoofing groups and global patterns.Relying solely on local context makes it challenging to capture multi-granularity risk signals, especially for organized and covert spoofing.Additionally, existing methods fail to consider the differences and relative importance between features of varying granularity, leading to feature distortion and noise.Therefore, we propose a multi-granularity augmented graph learning method that differentially captures fraud signals at local, group, and global levels.It utilizes multi-hop differential aggregation and communityaugmented strategy to capture information from local to global perspectives, adaptively distinguishing the contributions of different granularity.To avoid excessive fusion of multi-granularity information, we combine contrastive loss and cross-entropy loss for joint optimization, preserving key features while enhancing the method's robustness and accuracy.Extensive experiments on real-world datasets demonstrate the effectiveness of our proposed approach in spoofing detection, providing a robust solution for regulatory agencies.Our work will help financial institutions enhance their regulatory capabilities, protect investors' interests, and promote the healthy development of financial markets.

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.

Nowadays, Graph Fraud Detection (GFD) in financial scenarios has become an urgent research topic to protect online payment security. However, as organized crime groups are becoming more professional in real-world scenarios, fraudsters are employing more sophisticated camouflage strategies. Specifically, fraudsters disguise themselves by mimicking the behavioral data collected by platforms, ensuring that their key characteristics are consistent with those of benign users to a high degree, which we call Adaptive Camouflage. Consequently, this narrows the differences in behavioral traits between them and benign users within the platform's database, thereby making current GFD models lose efficiency. To address this problem, we propose a relation diffusion-based graph augmentation model Grad. In detail, Grad leverages a supervised graph contrastive learning module to enhance the fraud-benign difference and employs a guided relation diffusion generator to generate auxiliary homophilic relations from scratch. Based on these, weak fraudulent signals would be enhanced during the aggregation process, thus being obvious enough to be captured. Extensive experiments have been conducted on two real-world datasets provided by WeChat Pay, one of the largest online payment platforms with billions of users, and three public datasets. The results show that our proposed model Grad outperforms SOTA methods in both various scenarios, achieving at most 11.10% and 43.95% increases in AUC and AP, respectively.

Electric power resources are essential for the efficient and orderly development of society. Accurate power load forecasting is a key driver for the low-carbon upgrade of power systems. Traditional forecasting methods often struggle to capture long-term dependencies. Additionally, extracting complex nonlinear features from data remains a significant challenge, making it challenging to meet the accuracy demands of modern power systems. Besides, current deep learning-based forecasting methods cannot simulate multi-granularity power load data. To address these challenges, this paper presents a Generative Pre-trained Transformer model, GPT4PLTS, designed for power data simulation and fine-grained power load forecasting. The model leverages the Transformer architecture, incorporating the first six layers of the GPT decoder structure. It utilizes a multi-head attention mechanism to extract temporal features and includes a time alignment layer to maintain the sequence of time-series data, addressing both short-term and long-term dependencies. Extensive experiments are conducted on load observations from 2000 enterprises. The results demonstrate that GPT4PLTS achieves high accuracy in data simulation and forecasts across different time granularities, particularly excelling in short and medium-term predictions. Future research could focus on optimizing the model structure to enhance the model’s generalization ability.

Financial transaction systems have become the critical backbone of modern society, and the sharp increase in fraudulent transactions has become an unavoidable significant topic. Their presence poses a severe threat to financial markets, impacting the health of the economic and social welfare systems of various countries. However, most existing fraud detection methods are limited to detecting individual fraudulent entities within static transaction networks, which are neither suitable for continuously changing dynamic transaction networks nor capable of detecting the increasingly prevalent organized fraud crimes. This paper introduces a novel approach, Parallel Graph Learning with Temporal Stamp Encoding (PGLTSE). On the one hand, it designs a history information module to perform temporal dimension feature learning to adapt to the continuous changes in transaction information in Continuous-Time Dynamic Graphs (CTDG). On the other hand, it designs a gang-aware risk propagation algorithm to infer the risk of organized fraudulent activities in the global transaction relation graph. By simultaneously conducting parallel graph representation learning in both homogeneous global transaction relation graphs and heterogeneous local entity interaction graphs, it aggregates local interaction and global association information for end-to-end training. Extensive experiments on diverse real-world datasets substantiate the superior performance of PGLTSE over existing methods, demonstrating its practical efficacy in detecting complex and evolving fraudulent behaviors in financial networks.

Graph generation plays an essential role in understanding the formation of complex network structures across various fields, such as biological and social networks. Recent studies have shifted towards employing deep learning methods to grasp the topology of graphs. Yet, most current graph generators fail to adequately capture the community structure, which stands out as a critical and distinctive aspect of graphs. Additionally, these generators are generally limited to smaller graphs because of their inefficiencies and scaling challenges. This paper introduces the Community-Preserving Graph Adversarial Network (CPGAN), designed to effectively simulate graphs. CPGAN leverages graph convolution networks within its encoder and maintains shared parameters during generation to encapsulate community structure data and ensure permutation invariance. We also present the Scalable Community-Preserving Graph Attention Network (SCPGAN), aimed at enhancing the scalability of our model. SCPGAN considerably cuts down on inference and training durations, as well as GPU memory usage, through the use of an ego-graph sampling approach and a short-pipeline autoencoder framework. Tests conducted on six real-world graph datasets reveal that CPGAN manages a beneficial balance between efficiency and simulation quality when compared to leading-edge baselines. Moreover, SCPGAN marks substantial strides in model efficiency and scalability, successfully increasing the size of generated graphs to the 10 million node level while maintaining competitive quality, on par with other advanced learning models.

Depression is an affective disorder that poses a serious threat to both mental and physical health. Utilizing fuzzy -based neural network models for the identification and screening of depression can facilitate early intervention and treatment. Intuitionistic fuzzy stochastic configuration networks (IFSCNs) utilize a cost-sensitive learning framework, which enhances generalization performance for solving binary classification problems. However, IFSCNs ignore the impact of the relative neighborhood density of imbalanced samples with outliers. To learn more discriminant information from class imbalance depression recognition task, in this paper, we propose a novel deep ensemble stochastic configuration network via graph intuitionistic fuzzy, termed as DeSCN-GIF. Specifically, we first use graph-based intuitionistic fuzzy method to determine the membership and non-membership functions through the relative neighborhood density of imbalanced samples; moreover, an incremental self-ensemble deep stochastic configuration framework is presented to learn multi-level discriminative features, in which graph intuitionistic fuzzy cost-sensitive least squares loss function and weighted supervision mechanisms are applied to determine the parameters of DeSCN-GIF. Experimental results on Chinese syllable-based imbalanced depression voice datasets show that DeSCN-GIF has better binary classification performance compared to other learning models such as IFSCN, DSCN, SCN, GE-IFRVFL-CIL, IFRVFL, EDRVFL, DRVFL, RVFL, and DIFL-TSVM.

Fraud detection, a classical data mining problem in finance applications, has risen in significance amid the intensifying confrontation between fraudsters and anti-fraud forces. Recently, an increasing number of criminals are constantly expanding the scope of fraud activities to covet the property of innocent victims. However, most existing approaches require abundant historical records to mine fraud patterns from financial transaction behaviors, thereby leading to significant challenges to protect minority groups, who are less involved in the modern financial market but also under the threat of fraudsters nowadays. Therefore, in this paper, we propose a novel community-enhanced multi-relation graph neural network-based model, named CMR-GNN, to address the important defects of existing fraud detection models in the tail effect situation. In particular, we first construct multiple types of relation graphs from historical transactions and then devise a clustering-based neural network module to capture diverse patterns from transaction communities. To mitigate information lacking tailed nodes, we proposed tailed-groups learning modules to aggregate features from similarly clustered subgraphs by graph convolution networks. Extensive experiments on both the real-world and public datasets demonstrate that our method not only surpasses the state-of-the-art baselines but also could effectively harness information within transaction communities while mitigating the impact of tail effects.

Credit card fraud is a severe issue that causes significant losses for both cardholders and issuing banks. Existing methods utilize machine learning-based classifiers to identify fraudulent transactions from labeled transaction records. However, labeled data are often scarce compared to the billions of real transactions due to the high cost of annotation, which means that previous methods do not fully utilize the rich features of unlabeled data. What’s more, contemporary methods succumb to a fallacy of unawareness of the local risk structure and the inability to capture certain risk patterns. Therefore, we propose the Risk-aware Gated Temporal Attention Network (RGTAN) for fraud detection in this work. Specifically, we first build a temporal transaction graph based on the transaction records, which consists of temporal transactions (nodes) and their interactions (edges). Then we leverage a Gated Temporal Graph Attention (GTGA) Mechanism to propagate messages among the nodes and learn adaptive representations of transactions. We also model the fraud patterns through risk propagation, taking advantage of the relations among transactions. More importantly, we devise a neighbor risk-aware representation learning layer to enhance our method’s perception of multi-hop risk structures. We conduct extensive experiments on a real-world credit card transaction dataset and two public fraud detection datasets. The results show that our proposed method, RGTAN, outperforms other state-of-the-art methods on three fraud detection datasets. The risk-aware semi-supervised experiments also demonstrate the excellent performance of our model with only a small fraction of manually labeled data. Moreover, RGTAN has been deployed in a world-leading credit card issuer for credit card fraud detection, and the case study results show the effectiveness of our method in uncovering real-world fraud patterns.

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.

As a prominent application of deep neural networks in financial literature, bank credit ratings play a pivotal role in safeguarding global economic stability and preventing crises. In the contemporary financial system, interconnectivity among banks has reached unprecedented levels. However, many existing credit risk models continue to assess each bank independently, resulting in inevitable suboptimal performance. Thus, developing advanced neural networks to model intricate temporal dynamics and interconnected relationships in the banking system is essential for an effective credit rating and risk assessment learning system. To this end, we propose a novel hierarchical fusion transformer for interbank credit rating and risk assessment (HFTCRNet), which includes the long-term temporal transformer (LT3) module, short-term cross-graph transformer (STCGT) module, attentive risk contagion transformer (ARCT) module, and hierarchical fusion transformer (HFT) module to capture the long-term growth trajectories of banks, the short-term interbank network variance, the potential propagation of risks within interbank network, and integrate these information hierarchically. We further develop an interbank credit rating dataset, encompassing quarterly financial data, interbank lending networks, and key indicators such as credit ratings and systemic risk (SRISK) for 4548 banks from 2016Q1 to 2023Q1. Notably, we also adapt the minimum density algorithm to stabilize the interbank loan network over time, aiding in the analysis of long-term and short-term network effects. Our learning system uses semi-supervised learning to handle labels of varying sparsity, integrating credit ratings and SRISK for a comprehensive assessment of individual bank creditworthiness and systemic interbank risk. Extensive experimental results on our interbank dataset show that HFTCRNet not only outperforms all the baselines in terms of credit rating accuracy but also can evaluate the systemic risk within the interbank network. Code will be available at: https://github.com/AI4Risk/HFTCRNet.

Accurately assessing and forecasting bank credit ratings at an early stage is vitally important for a healthy financial environment and sustainable economic development. However, the evaluation process faces challenges due to individual attacks on the rating model. Some participants may provide manipulated information in an attempt to undermine the rating model and secure higher scores, further complicating the evaluation process. Therefore, we propose a novel approach called the preferential selective-aware graph neural network (PSAGNN) model to simultaneously defend against feature and structural nontarget poisoning attacks on Interbank credit ratings. In particular, the model establishes a phased optimization approach combined with biased perturbation and explores the Interbank preferences and scale-free nature of networks, to adaptively prioritize the poisoning training data and simulate a clean graph. Finally, we apply a weighted penalty on the opposition function to optimize the model so that the model can distinguish between attackers. Extensive experiments on our newly collected Interbank quarter dataset and case studies demonstrate the superior performance of our proposed approach in preventing credit rating attacks compared to state-of-the-art baselines.

Medical insurance fraud has always been a crucial challenge in the field of healthcare industry. Existing fraud detection models mostly focus on offline learning scenes. However, fraud patterns are constantly evolving, making it difficult for models trained on past data to detect newly emerging fraud patterns, posing a severe challenge in medical fraud detection. Moreover, current incremental learning models are mostly designed to address catastrophic forgetting, but often exhibit suboptimal performance in fraud detection. To address this challenge, this paper proposes an innovative online learning method for medical insurance fraud detection, named POCL. This method combines contrastive learning pre-training with online updating strategies. In the pre-training stage, we leverage contrastive learning pre-training to learn on historical data, enabling deep feature learning and obtaining rich risk representations. In the online learning stage, we adopt a Temporal Memory Aware Synapses online updating strategy, allowing the model to perform incremental learning and optimization based on continuously emerging new data. This ensures timely adaptation to fraud patterns and reduces forgetting of past knowledge. Our model undergoes extensive experiments and evaluations on real-world insurance fraud datasets. The results demonstrate our model has significant advantages in accuracy compared to the state-of-the-art baseline methods, while also exhibiting lower running time and space consumption. Our sources are released at https://github.com/finint/POCL.

High-frequency algorithmic trading has consistently attracted attention in both academic and industrial fields, which is formally modeled as a near real-time sequential decision problem. DRL methods are treated as a promising direction compared with the traditional approaches, as they have shown great potential in chasing maximum accumulative return. However, the financial data gathered from volatile market change rapidly, which makes it dramatically difficult to grasp crucial factors for effective decision-making. Existing works mainly focus on capturing temporal relations while ignoring deriving essential factors across features. Therefore, we propose a DRL-based cross-contextual sequential optimization (CCSO) method for algorithmic trading. In particular, we employ a convolution module in the first stage to derive latent factors via inter-sequence aggregation and apply a well-designed self-attention module in the second stage to capture market dynamics by aggregating temporal intra-sequence details. With the two-stage extractor as encoder and a RNN-based decision-maker as decoder, an Encoder-Decoder module is established as the policy network to conduct potent feature analysis and suggest action plans. Then, we design a dynamic programming based learning method to address the challenge of complex network updates in reinforcement learning, leading to considerable enhancement in learning stability and efficiency. To the best of our knowledge, this is the first work that solves the sequential optimization problem by joint representation of trading data across time and features in the DRL framework. Extensive experiments demonstrate the superior performance of our method compared to other state-of-the-art algorithmic trading approaches in various widely-used metrics.

Portfolio optimization is a significant concern in finance. Existing research on portfolio optimization fails to adequately learn from the long and short-term relationships among equities, which inevitably leads to suboptimal performance. In this paper, we propose a Dynamic Graph-based Deep Reinforcement Learning (DGDRL) for optimal portfolio decisions. We achieve this goal by devising two mechanisms for naturally modeling the financial market. Firstly, we utilize the static and dynamic graphs to represent the long and short-term relations, which are then naturally represented by the proposed multi-channel graph attention neural network. Secondly, compared with the traditional two-phase approach, forecasting equity's trend and then weighting them by combinatorial optimization, we naturally optimize the portfolio decisions, which could directly guide the model to converge to optimal rewards. Through extensive experiments on three real-world datasets, we have demonstrated that our method significantly outperforms state-of-the-art benchmark methods in portfolio management. Furthermore, the evaluation of the industrial trading system has shown the applicability of our model to real-world financial markets.

Stock price prediction is a challenging problem in the field of finance and receives widespread attention. In recent years, with the rapid development of technologies such as deep learning and graph neural networks, more research methods have begun to focus on exploring the interrelationships between stocks. However, existing methods mostly focus on the short-term dynamic relationships of stocks and directly integrating relationship information with temporal information. They often overlook the complex nonlinear dynamic characteristics and potential higher-order interaction relationships among stocks in the stock market. Therefore, we propose a stock price trend prediction model named LSR-IGRU in this paper, which is based on long short-term stock relationships and an improved GRU input. Firstly, we construct a long short-term relationship matrix between stocks, where secondary industry information is employed for the first time to capture long-term relationships of stocks, and overnight price information is utilized to establish short-term relationships. Next, we improve the inputs of the GRU model at each step, enabling the model to more effectively integrate temporal information and long short-term relationship information, thereby significantly improving the accuracy of predicting stock trend changes. Finally, through extensive experiments on multiple datasets from stock markets in China and the United States, we validate the superiority of the proposed LSR-IGRU model over the current state-of-the-art baseline models. We also apply the proposed model to the algorithmic trading system of a financial company, achieving significantly higher cumulative portfolio returns compared to other baseline methods. Our sources are released at https://github.com/ZP1481616577/Baselines\_LSR-IGRU.

In this paper, we present a novel hierarchical approach to adversarial attacks targeting Graph Neural Networks (GNNs), tailored to overcome the complexities inherent in large-scale poisoning attacks. Traditional global attack strategies often fail to yield effective results on extensive graph structures. Our innovative method implements a divide-and-conquer tactic, clustering nodes based on their embeddings and forming coarse-grained graphs from these clusters. We initiate perturbations at this coarse level, gradually honing them in more detailed, finer-grained graphs, while keeping non-essential nodes grouped. By employing meta-gradients derived from these refined graphs, we pinpoint critical edges for perturbation, thereby vastly simplifying the process and reducing the intricacy involved in manipulating large-scale graphs. This hierarchical strategy not only enhances the efficacy of the attacks but also maintains operational efficiency across expansive network structures.

Graph Neural Networks (GNNs) have emerged as the leading deep learning models for graph-based representation learning. However, the training and inference of GNNs on large graphs remain resource-intensive, impeding their utility in real-world scenarios and curtailing their applicability in deeper and more sophisticated GNN architectures. To address this issue, the Graph Lottery Ticket (GLT) hypothesis assumes that GNN with random initialization harbors a pair of core subgraph and sparse subnetwork, which can yield comparable performance and higher efficiency to that of the original dense network and complete graph. Despite that GLT offers a new paradigm for GNN training and inference, existing GLT algorithms heavily rely on trial-and-error pruning rate tuning and scheduling, and adhere to an irreversible pruning paradigm that lacks elasticity. Worse still, current methods suffer scalability issues when applied to deep GNNs, as they maintain the same topology structure across all layers. These challenges hinder the integration of GLT into deeper and larger-scale GNN contexts. To bridge this critical gap, this paper introduces an $\textbf{A}$daptive, $\textbf{D}$ynamic, and $\textbf{A}$utomated framework for identifying $\textbf{G}$raph $\textbf{L}$ottery $\textbf{T}$ickets ($\textbf{AdaGLT}$). Our proposed method derives its key advantages and addresses the above limitations through the following three aspects: 1) tailoring layer-adaptive sparse structures for various datasets and GNNs, thus endowing it with the capability to facilitate deeper GNNs; 2) integrating the pruning and training processes, thereby achieving a dynamic workflow encompassing both pruning and restoration; 3) automatically capturing graph lottery tickets across diverse sparsity levels, obviating the necessity for extensive pruning parameter tuning. More importantly, we rigorously provide theoretical proofs to guarantee $\textbf{AdaGLT}$ to mitigate over-smoothing issues and obtain improved sparse structures in deep GNN scenarios. Extensive experiments demonstrate that $\textbf{AdaGLT}$ outperforms state-of-the-art competitors across multiple graph datasets of various scales and types, particularly in scenarios involving deep GNNs.

In this study, we present $\text{aL\small{LM}4T\small{S}}$, an innovative framework that adapts Large Language Models (LLMs) for time-series representation learning. Central to our approach is that we reconceive time-series forecasting as a self-supervised, multi-patch prediction task, which, compared to traditional mask-and-reconstruction methods, captures temporal dynamics in patch representations more effectively. Our strategy encompasses two-stage training: (i). a causal continual pre-training phase on various time-series datasets, anchored on next patch prediction, effectively syncing LLM capabilities with the intricacies of time-series data; (ii). fine-tuning for multi-patch prediction in the targeted time-series context. A distinctive element of our framework is the patch-wise decoding layer, which departs from previous methods reliant on sequence-level decoding. Such a design directly transposes individual patches into temporal sequences, thereby significantly bolstering the model’s proficiency in mastering temporal patch-based representations. $\text{aL\small{LM}4T\small{S}}$ demonstrates superior performance in several downstream tasks, proving its effectiveness in deriving temporal representations with enhanced transferability and marking a pivotal advancement in the adaptation of LLMs for time-series analysis.

Graph Neural Networks (GNNs) excel in various graph learning tasks but face computational challenges when applied to large-scale graphs. A promising solution is to remove non-essential edges to reduce the computational overheads in GNN. Previous literature generally falls into two categories: topology-guided and semantic-guided. The former maintains certain graph topological properties yet often underperforms on GNNs. % due to low integration with neural network training. The latter performs well at lower sparsity on GNNs but faces performance collapse at higher sparsity levels. With this in mind, we propose a new research line and concept termed **Graph Sparse Training** **(GST)**, which dynamically manipulates sparsity at the data level. Specifically, GST initially constructs a topology & semantic anchor at a low training cost, followed by performing dynamic sparse training to align the sparse graph with the anchor. We introduce the **Equilibria Sparsification Principle** to guide this process, balancing the preservation of both topological and semantic information. Ultimately, GST produces a sparse graph with maximum topological integrity and no performance degradation. Extensive experiments on 6 datasets and 5 backbones showcase that GST **(I)** identifies subgraphs at higher graph sparsity levels ($1.67\%\sim15.85\%$$\uparrow$) than state-of-the-art sparsification methods, **(II)** preserves more key spectral properties, **(III)** achieves $1.27-3.42\times$ speedup in GNN inference and **(IV)** successfully helps graph adversarial defense and graph lottery tickets.

After the Transformer model demonstrated excellent performance in natural language processing (NLP) tasks and computer vision tasks, people have started to explore the use of Transformer models in the field of time series prediction. Because of the significant role of the stock market in the global economy, stock market prediction is of paramount importance for investors. Stock indices forecasting is one of the fields of stock market forecasting and researchers have also set their sights on Transformer. However, with limited semantic information available in time series data and the unique characteristics of the self-attention mechanism, the Transformer model has not gained widespread adoption in stock indices forecasting. In this paper, we propose a history window serialization based Transformer model (HWSformer) specifically designed for predicting stock price indices. Our innovation is to introduce the historical window serialization layer to solve the problem of limited semantic richness in time series data, which affects the validity of self-attention. Additionally, in order to capture the original distribution accurately and retain the valuable non-stationary information, we incorporate the Reversible Instance Normalization (RevIN) method. We conducted experiments on 12 stock price index datasets collected from multiple countries and demonstrated that HWSformer outperforms traditional Transformer models by approximately 20% and varying degrees of improvement compared to other recent variants of Transformers.

Training high-quality deep models necessitates vast amounts of data, resulting in overwhelming computational and memory demands. Recently, data pruning, distillation, and coreset selection have been developed to streamline data volume by \textit{retaining}, \textit{synthesizing}, or \textit{selecting} a small yet informative subset from the full set. Among these methods, data pruning incurs the least additional training cost and offers the most practical acceleration benefits. However, it is the most vulnerable, often suffering significant performance degradation with imbalanced or biased data schema, thus raising concerns about its accuracy and reliability in on-device deployment. Therefore, there is a looming need for a new data pruning paradigm that maintains the efficiency of previous practices while ensuring balance and robustness. Unlike the fields of computer vision and natural language processing, where mature solutions have been developed to address these issues, graph neural networks (GNNs) continue to struggle with increasingly large-scale, imbalanced, and noisy datasets, lacking a unified dataset pruning solution. To achieve this, we introduce a novel dynamic soft-pruning method, \ourmethod, designed to update the training ``basket'' during the process using trainable prototypes. \ourmethod first constructs a well-modeled graph embedding hypersphere and then samples \textit{representative, balanced, and unbiased subsets} from this embedding space, which achieves the goal we called {\fontfamily{lmtt}\selectfont \textbf{Graph Training Debugging}}. Extensive experiments on four datasets across three GNN backbones, demonstrate that \ourmethod (I) achieves or surpasses the performance of the full dataset with $30\%\sim50\%$ fewer training samples, (II) attains up to a $2.81\times$ lossless training speedup, and (III) outperforms state-of-the-art pruning methods in imbalanced training and noisy training scenarios by $0.3\%\sim4.3\%$ and $3.6\%\sim7.8\%$, respectively.

Vulnerable node detection in uncertain graphs is a typical graph mining problem that seeks to identify nodes at a high risk of breakdown under the joint effect from both the self and contagion risk probability. This is an NP-hard problem that is crucial for risk management in many real-world applications such as networked loans and smart grids. Monte Carlo (MC) simulation and its optimized algorithms are commonly used to approximate the breakdown probability, but these methods require a large number of samples to ensure accuracy, which is computationally expensive for large-scale networks. Although recent studies employ Graph Neural Networks (GNNs) to model the contagion process and accelerate the inference, many of these methods suffer from the over-smoothing problem, leading to suboptimal performance under the long-distance risk contagion process. To this end, we propose a novel risk-adaptive deep reinforcement learning-based framework (AdaRisk) for vulnerable nodes detection in uncertain graphs. In particular, we design the Markov Decision Process (MDP) of the vulnerability estimation process in which our agent would approach the risk adaptively based on contagion probability accumulated in prior iterations. To encode state embeddings that incorporate multi-hop contagion information, the agent utilizes a long-distance adaptable policy network to process the input graph and output actions as the vulnerable probability of nodes. We conducted extensive experiments on four benchmark networks and three real-world financial networks to evaluate our proposed framework's performance. Our results demonstrate that AdaRisk outperforms state-of-the-art baselines in terms of detection performance, and also offers significant running time reductions compared to MC simulation.

The networked-loan is major financing support for Micro, Small and Medium-sized Enterprises (MSMEs) in some developing countries. But external shocks may weaken the financial networks' robustness; an accidental default may spread across the network and collapse the whole network. Thus, predicting the critical firms in networked-loans to stem contagion risk and prevent potential systemic financial crises is of crucial significance to the long-term health of inclusive finance and sustainable economic development. Existing approaches in the banking industry dismiss the contagion risk across loan networks and need extensive knowledge with sophisticated financial expertise. Regarding the issues, we propose a novel approach to predict critical firms for stemming contagion risk in the bank industry with deep reinforcement learning integrated with high-order graph message-passing networks. We demonstrate that our approach outperforms the state-of-the-art baselines significantly on the dataset from a large commercial bank. Moreover, we also conducted empirical studies on the real-world loan dataset for risk mitigation. The proposed approach enables financial regulators and risk managers to better track and understands contagion and systemic risk in networked-loans. The superior performance also represents a paradigm shift in addressing the modern challenges in financing support of MSMEs and sustainable economic development.

Credit card fraud incurs a considerable cost for both cardholders and issuing banks. Contemporary methods apply machine learning-based classifiers to detect fraudulent behavior from labeled transaction records. But labeled data are usually a small proportion of billions of real transactions due to expensive labeling costs, which implies that they do not well exploit many natural features from unlabeled data. Therefore, we propose a semi-supervised graph neural network for fraud detection. Specifically, we leverage transaction records to construct a temporal transaction graph, which is composed of temporal transactions (nodes) and interactions (edges) among them. Then we pass messages among the nodes through a Gated Temporal Attention Network (GTAN) to learn the transaction representation. We further model the fraud patterns through risk propagation among transactions. The extensive experiments are conducted on a real-world transaction dataset and two publicly available fraud detection datasets. The result shows that our proposed method, namely GTAN, outperforms other state-of-the-art baselines on three fraud detection datasets. Semi-supervised experiments demonstrate the excellent fraud detection performance of our model with only a tiny proportion of labeled data.

Sales forecasting during the launch of new products has always been a challenging task, due to the lack of historical sales data. The dynamic market environment and consumer preferences also increase the uncertainty of predictions. Large chains face even greater difficulties due to their extensive presence across various regions. Traditional time-series forecasting methods usually rely on statistical models and empirical judgments, which are difficult to handle large, variable data and often fail to achieve satisfactory performance for new products. In this paper, we propose a Multi-granularity AdversaRial Learning framework (MARL) to leverage knowledge from old products and improve the quality of invariant representations for more accurate sales predictions. To evaluate our proposed method, we conducted extensive experiments on both a real-world dataset from a prominent international Café chain and a public dataset. The results demonstrated that our method is more effective than the existing state-of-the-art baselines for new product sales forecasting.

Taolin Zhang, Ruyao Xu, Chengyu Wang, Zhongjie Duan, Cen Chen, Minghui Qiu, Dawei Cheng, Xiaofeng He, Weining Qian. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.

Accurately credit rating on Interbank assets is essential for a healthy financial environment and substantial economic development. But individual participants tend to provide manipulated information in order to attack the rating model to produce a higher score, which may conduct serious adverse effects on the economic system, such as the 2008 global financial crisis. To this end, in this paper, we propose a novel selective-aware graph neural network model (SA-GNN) for defense the Interbank credit rating attacks. In particular, we first simulate the rating information manipulating process by structural and feature poisoning attacks. Then we build a selective-aware defense graph neural model to adaptively prioritize the poisoning training data with Bernoulli distribution similarities. Finally, we optimize the model with weighed penalization on the objection function so that the model could differentiate the attackers. Extensive experiments on our collected real-world Interbank dataset, with over 20 thousand banks and their relations, demonstrate the superior performance of our proposed method in preventing credit rating attacks compared with the state-of-the-art baselines.

Medical insurance plays a vital role in modern society, yet organized healthcare fraud causes billions of dollars in annual losses, severely harming the sustainability of the social welfare system. Existing works mostly focus on detecting individual fraud entities or claims, ignoring hidden conspiracy patterns. Hence, they face severe challenges in tackling organized fraud. In this paper, we proposed RDPGL, a novel Risk Diffusion-based Parallel Graph Learning approach, to fighting against medical insurance criminal gangs. In particular, we first leverage a heterogeneous graph attention network to encode the local context from the beneficiary-provider graph. Then, we devise a community-aware risk diffusion model to infer the global context of organized fraud behaviors with the claim-claim relation graph. The local and global representations are parallel concatenated together and trained simultaneously in an end-to-end manner. Our approach is extensively evaluated on a real-world medical insurance dataset. The experimental results demonstrate the superiority of our proposed approach, which could detect more organized fraud claims with relatively high precision compared with state-of-the-art baselines.

High-frequency trading is an extraordinarily intricate financial task, which is normally treated as a near real-time sequential decision problem. Compared with the traditional two-phase approach, forecasting equity's trend and then weighting them by combinatorial optimization, deep reinforcement learning (DRL) methods have shown advances in reward chasing with optimal policies. However, existing DRL-based methods either leverage portfolio optimization on low-frequency scenarios or only support a very limited number of assets with discrete action space, facing significant computing efficiency challenges. Therefore, we propose an efficient DRL-based policy optimization (DRPO) method for high-frequency trading. In particular, we model the portfolio management task with Markov Decision Process by directly inferring the equity weights in the action space guided by maximum accumulated returns. To reduce agents' interaction complexity without reducing interpretation, we detach the environment into the "static'' market states and "dynamic'' portfolio weight states. Then, we design an efficient reward expectation calculation algorithm via probabilistic dynamic programming, which enables our agents directly collect feedback away from trajectory sampling-based morass. To the best of our knowledge, this is the first work that solves the high-frequency portfolio optimization problem by devising an efficient continuous space policy optimization algorithm in the DRL framework. Through extensive experiments on the real-world data from Dow Jones, Coinbase and SSE exchanges, we show that our proposed DRPO significantly outperforms state-of-the-art benchmark methods. The results demonstrate the practical applicability and effectiveness of the proposed method.

Named entity recognition (NER) is a fundamental task for information extraction applications. NER is challenging because of semantic ambiguities in academic literature, especially for non-Latin languages. Besides word semantic information, recognizing Chinese named entities needs to consider word boundary information, as words contained in Chinese texts are not separated with spaces. Leveraging word boundary information could help to determine entity boundaries and thus improve entity recognition performance. In this article, we propose to combine word boundary information and semantic information for named entity recognition based on multi-task adversarial learning. Specifically, we learn commonly shared boundary information of entities from multiple kinds of tasks, including Chinese word segmentation (CWS), part-of-speech (POS) tagging, and entity recognition, with adversarial learning. We learn task-specific semantic information of words from these tasks and combine the learned boundary information with the semantic information to improve entity recognition with multi-task learning. We then propose a compression method based on improved clustering to accelerate the proposed model. We conduct extensive experiments on four public benchmark datasets and two private datasets, compared with state-of-the-art baseline models, and the experimental results demonstrate that our model achieves considerable performance improvements on various evaluation datasets.

Uncertain graphs have been widely used to model complex linked data in many applications, such as guaranteed-loan networks and power grids. In these networks, a node usually has a certain chance of default due to self-factors or the influence from upstream nodes. For regulatory authorities, it is critical to efficiently identify the vulnerable nodes, i.e., nodes with high default risks, such that they could pay more attention to these nodes for the purpose of risk management. In this paper, we propose and investigate the top-k vulnerable nodes detection problem in uncertain graphs. We formally define the model and prove it hardness. A sampling-based approach is first proposed. Rigorous theoretical analysis is conducted to bound the quality of returned results. Novel optimization techniques and a bottom-k sketch based approach are further developed to scale for large networks. We demonstrate the performance of proposed techniques on 3 real financial networks and 5 benchmark networks. Moreover, to further verify the advantages of our model, we integrate the proposed techniques with our loan risk control system, which is deployed in the collaborated bank. Particularly, we show that our proposed model can better estimate the default risks of enterprises compared to the state-of-the-art techniques.

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.

Under debt obligation promises, small- and medium-sized enterprises (SMEs) can guarantee each other to enhance their financial security to get loans from commercial banks. When the economy rises, the banks may reduce the threshold to some extent, which may introduce default risk during the economy down period, especially when many SMEs bind together and form complex networks. The risk may diffuse across the guarantee network and may result in a financial crisis. Macroprudential oversight of the guarantee network to eliminate any potential systematics financial risk is the central task of the regulatory commission and the commercial banks. Based on our observation, the delinquent probability of an SME depends not only on self-financial status but also highly related to its temporal behaviors and structural position in networks. The classic approach for loan assessment criteria face challenges in extracting temporal and structural patterns from dynamic networks. To address these issues, we propose a temporal delinquent event prediction (TDEP) framework that preserves temporal network structures and credit behavior sequences in an end-to-end model. In particular, we first employ a graph attention layer to learn the representation of nodes in temporal guarantee networks. We then design a recursive and self-attention mechanism to integrate both credit behavior and network structure information. The learned attentional weights are leveraged to uncover high-risk guarantee patterns that effectively accelerate the risk assessment process. Afterward, we conduct extensive experiments in a real-world guaranteed-loan data set to evaluate its performance. The results show the effectiveness of our proposed approach compared with the state-of-the-art baselines. Finally, we integrate the proposed model in a real-world loan risk management system. We present the implementation details of each subcomponent of the system and report out the performance after online deployment.

Algorithmic trading is a crucial yet challenging task in the financial domain, where trading decisions are made sequentially from milliseconds to days based on the historical price movements and trading frequency. To model such a sequential decision making process in the dynamic financial markets, Deep Reinforcement Learning (DRL) based methods have been applied and demonstrated their success in finding trading strategies that achieve profitable returns. However, the financial markets are complex imperfect information games with high-level of noise and uncertainties which usually make the exploration policy of DRL less effective. In this paper, we propose an end-to-end DRL method that explores solutions on the whole graph via a probabilistic dynamic programming algorithm. Specifically, we separate the state into environment state and position state, and model the position state transition as a directed acyclic graph. To obtain reliable gradients for model training, we adopt a probabilistic dynamic programming algorithm to explore solutions over the whole graph instead of sampling a path. By avoiding the sampling procedure, we propose an efficient training algorithm and overcome the efficiency problem in most existing DRL methods. Furthermore, our method is compatible with most recurrent neural network architecture, which makes our method easy to implement and very effective in practice. Extensive experiments have been conducted on two real-world stock datasets. Experimental results demonstrate that our method can generate stable trading strategies for both high-frequency and low-frequency trading, significantly outperforming the baseline DRL methods on annualized return and Sharpe ratio.

The price movement prediction of stock market has been a classical yet challenging problem, with the attention of both economists and computer scientists. In recent years, graph neural network has significantly improved the prediction performance by employing deep learning on company relations. However, existing relation graphs are usually constructed by handcraft human labeling or nature language processing, which are suffering from heavy resource requirement and low accuracy. Besides, they cannot effectively response to the dynamic changes in relation graphs. Therefore, in this paper, we propose a temporal and heterogeneous graph neural network-based (THGNN) approach to learn the dynamic relations among price movements in financial time series. In particular, we first generate the company relation graph for each trading day according to their historic price. Then we leverage a transformer encoder to encode the price movement information into temporal representations. Afterward, we propose a heterogeneous graph attention network to jointly optimize the embeddings of the financial time series data by transformer encoder and infer the probability of target movements. Finally, we conduct extensive experiments on the stock market in the United States and China. The results demonstrate the effectiveness and superior performance of our proposed methods compared with state-of-the-art baselines. Moreover, we also deploy the proposed THGNN in a real-world quantitative algorithm trading system, the accumulated portfolio return obtained by our method significantly outperforms other baselines.

Uncertain graphs have been widely used to model complex linked data in many applications, such as guaranteed-loan networks and power grids. In these networks, a node usually has a certain chance of default due to self-factors or the influence from upstream nodes. For regulatory authorities, it is critical to efficiently and accurately identify the vulnerable nodes, i.e., nodes with high default risk, such that people could pay more attention to these nodes for the purpose of risk management. In this paper, we propose and investigate the top-k vulnerable nodes detection problem in uncertain graphs. Due to the hardness of the problem, sampling-based methods are proposed with tight theoretical guarantee. We demonstrate the performance of proposed techniques on 3 real financial networks and 5 benchmark networks.

Graph generation is beneficial to comprehend the creation of meaningful structures of networks in a broad spec-trum of applications such as social networks and biological net-works. Recent studies tend to leverage deep learning techniques to learn the topology structures in graphs. However, we notice that the community structure, which is one of the most unique and prominent features of the graph, cannot be well captured by the existing graph generators. Moreover, the existing advanced deep learning-based graph generators are not efficient and scalable, which can only handle small graphs. In this paper, we propose a novel community-preserving generative adversarial network (CPGAN) for effective and efficient (scalable) graph simulation. We employ graph convolution networks in the encoder and share parameters in the generation process to transmit information about community structures and preserve the permutation-invariance in CPGAN. We conducted extensive experiments on benchmark datasets, including six sets of real-life graphs. The results demonstrate that CPGAN can achieve a good trade-off between efficiency (scalability) and graph simulation quality for real-life graph simulation compared with state-of-the-art baselines.

Named entity recognition (NER) isa preliminary task in natural language processing (NLP). Recognizing Chinese named entities from unstructured texts is challenging due to the lack of word boundaries. Even if performing Chinese Word Segmentation (CWS) could help to determine word boundaries, it is still difficult to determine which words should be clustered together for entity identification, since entities are often composed of multiple-segmented words. As dependency relationships between segmented words could help to determine entity boundaries, it is crucial to employ information related to syntactic dependency relationships to improve NER performance. In this paper, we propose a novel NER model to learn information about syntactic dependency graphs with graph neural networks, and merge learned information into the classic Bidirectional Long Short-Term Memory (BiLSTM) - Conditional Random Field (CRF) NER scheme. In addition, we extract various kinds of task-specific hidden information from multiple CWS and part-of-speech (POS) tagging tasks, to further improve the NER model. We finally leverage multiple self-attention components to integrate multiple kinds of extracted information for named entity identification. Experimental results on three public benchmark datasets show that our model outperforms the state-of-the-art baselines in most scenarios.

Card fraud is an important issue and incurs a considerable cost for both cardholders and issuing banks. Contemporary methods apply machine learning-based approaches to detect fraudulent behavior from transaction records. But manually generating features needs domain knowledge and may lay behind the modus operandi of fraud, which means we need to automatically focus on the most relevant fraudulent behavior patterns in the online detection system. Therefore, in this work, we propose a spatial-temporal attention-based graph network (STAGN) for credit card fraud detection. In particular, we learn the temporal and location-based transaction graph features by a graph neural network first. Afterwards, we employ the spatial-temporal attention on top of learned tensor representations, which are then fed into a 3D convolution network. The attentional weights are jointly learned in an end-to-end manner with 3D convolution and detection networks. After that, we conduct extensive experiments on the real-word card transaction dataset. The result shows that STAGN performs better than other state-of-the-art baselines in both AUC and precision-recall curves. Moreover, we conduct empirical studies with domain experts on the proposed method for fraud detection and knowledge discovery; the result demonstrates its superiority in detecting suspicious transactions, mining spatial and temporal fraud hotspots, and uncover fraud patterns. The effectiveness of the proposed method in other user behavior-based tasks is also demonstrated. Finally, in order to tackle the challenges of big data, we integrate our proposed STAGN into the fraud detection system as the predictive model and present the implementation detail of each module in the system.

Financial documents often contain rich domain information, such as named entities, which could be used to indicate the documents' classification categories. Existing classification methods either ignore such contained financial domain information, achieving less optimal performances, or train document representations in supervised ways, with expensive data labeling costs. In this paper, we propose to leverage domain information to improve classification performance for financial documents, via a graph representation learning model, namely G-MoCo, based on unsupervised graph momentum contrast. With G-MoCo, we could extract latent features from massive unlabeled raw data, and then further use the learned representations for document classification. Compared with the state-of-the-art baselines, representations learned by our method could improve performances by significant margins on a financial document dataset and three non-financial public graph datasets.

Credit card fraud is an important issue and incurs a considerable cost for both cardholders and issuing institutions. Contemporary methods apply machine learning-based approaches to detect fraudulent behavior from transaction records. But manually generating features needs domain knowledge and may lay behind the modus operandi of fraud, which means we need to automatically focus on the most relevant patterns in fraudulent behavior. Therefore, in this work, we propose a spatial-temporal attention-based neural network (STAN) for fraud detection. In particular, transaction records are modeled by attention and 3D convolution mechanisms by integrating the corresponding information, including spatial and temporal behaviors. Attentional weights are jointly learned in an end-to-end manner with 3D convolution and detection networks. Afterward, we conduct extensive experiments on real-word fraud transaction dataset, the result shows that STAN performs better than other state-of-the-art baselines in both AUC and precision-recall curves. Moreover, we conduct empirical studies with domain experts on the proposed method for fraud post-analysis; the result demonstrates the effectiveness of our proposed method in both detecting suspicious transactions and mining fraud patterns.

Aesthetic image cropping is a practical but challenging task which aims at finding the best crops with the highest aesthetic quality in an image. Recently, many deep learning methods have been proposed to address this problem, but they did not reveal the intrinsic mechanism of aesthetic evaluation. In this paper, we propose an interpretable image cropping model to unveil the mystery. For each image, we use a fully convolutional network to produce an aesthetic score map, which is shared among all candidate crops during crop-level aesthetic evaluation. Then, we require the aesthetic score map to be both composition-aware and saliency-aware. In particular, the same region is assigned with different aesthetic scores based on its relative positions in different crops. Moreover, a visually salient region is supposed to have more sensitive aesthetic scores so that our network can learn to place salient objects at more proper positions. Such an aesthetic score map can be used to localize aesthetically important regions in an image, which sheds light on the composition rules learned by our model. We show the competitive performance of our model in the image cropping task on several benchmark datasets, and also demonstrate its generality in real-world applications.

Static image action recognition, which aims to recognize action based on a single image, usually relies on expensive human labeling effort such as adequate labeled action images and large-scale labeled image dataset. In contrast, abundant unlabeled videos can be economically obtained. Therefore, several works have explored using unlabeled videos to facilitate image action recognition, which can be categorized into the following two groups: (a) enhance visual representations of action images with a designed proxy task on unlabeled videos, which falls into the scope of self-supervised learning; (b) generate auxiliary representations for action images with the generator learned from unlabeled videos. In this paper, we integrate the above two strategies in a unified framework, which consists of Visual Representation Enhancement (VRE) module and Motion Representation Augmentation (MRA) module. Specifically, the VRE module includes a proxy task which imposes pseudo motion label constraint and temporal coherence constraint on unlabeled videos, while the MRA module could predict the motion information of a static action image by exploiting unlabeled videos. We demonstrate the superiority of our framework based on four benchmark human action datasets with limited labeled data.

Many institutions are devoted to providing investment advising services to stock investors to help them make sound investment decisions. Industry analysts at these institutions need to analyze huge amounts of financial news documents, and yield investment advising reports to the service subscribers. Automatic document classification is required to organize collected financial news documents into pre-defined fine-grained categories, before the document analysis tasks. It is challenging to implement accurate fine-grained classification over massive financial documents, because documents from close fine-grained categories are highly semantically similar, while existing classification methods may fail to differentiate the subtle differences for documents from close fine-grained categories. In this paper, we implement a document classification framework, named GraphSEAT, to classify financial documents for a leading financial information service provider in China. Specifically, we build a heterogeneous graph to model the global structure of our targeting financial documents, where documents and financial named entities are deemed as nodes, and a document is connected to a contained named entity with an edge, and we then train a graph convolutional network (GCN) with attention mechanisms, to learn an embedding representation containing domain information for a document. We also extract semantic information from a document's word sequence with a neural sequence encoder, and finally form an overall embedding representation for a document and make the prediction, via fusing the two learned representations of the document with attention mechanisms. We perform extensive experiments on our real-world financial news dataset and three public datasets, to evaluate the performance of the document classification framework, and the experimental results demonstrate that GraphSEAT outperforms all compared eight baseline models, especially on our dataset.

Learning from web data has attracted lots of research interest in recent years. However, crawled web images usually have two types of noises, label noise and background noise, which induce extra difficulties in utilizing them effectively. Most existing methods either rely on human supervision or ignore the background noise. In this paper, we propose a novel method, which is capable of handling these two types of noises together, without the supervision of clean images in the training stage. Particularly, we formulate our method under the framework of multi-instance learning by grouping ROIs (i.e., images and their region proposals) from the same category into bags. ROIs in each bag are assigned with different weights based on the representative/discriminative scores of their nearest clusters, in which the clusters and their scores are obtained via our designed memory module. Our memory module could be naturally integrated with the classification module, leading to an end-to-end trainable system. Extensive experiments on four benchmark datasets demonstrate the effectiveness of our method.

In the context of urbanization and the rapid growth of energy demand, understanding the spatial and temporal dynamics of urban energy use is crucial for identifying energy-saving potentials. In this demo, we present a visual analysis tool, VAP, that allows users to explore the dynamics of urban energy use at different spatial and temporal scales. In contrast to traditional statistical and machine learning methods, the visual analysis based tool focuses on analytical thinking, user interactions and answering business questions by examining different visual analysis views. In the demonstration, conference attendees will interact with VAP and learn its capabilities in discovering typical consumption patterns and spatio-temporal shift patterns from a real-world case study of electricity.

The guaranteed loan is a debt obligation promise that if one corporation gets trapped in risks, its guarantors will back the loan. When more and more companies involve, they subsequently form complex networks. Detecting and predicting risk guarantee in these networked-loans is important for the loan issuer. Therefore, in this paper, we propose a dynamic graph-based attention neural network for risk guarantee relationship prediction (DGANN). In particular, each guarantee is represented as an edge in dynamic loan networks, while companies are denoted as nodes. We present an attention-based graph neural network to encode the edges that preserve the financial status as well as network structures. The experimental result shows that DGANN could significantly improve the risk prediction accuracy in both the precision and recall compared with state-of-the-art baselines. We also conduct empirical studies to uncover the risk guarantee patterns from the learned attentional network features. The result provides an alternative way for loan risk management, which may inspire more work in the future.

The share prices of listed companies in the stock trading market are prone to be influenced by various events. Performing event detection could help people to timely identify investment risks and opportunities accompanying these events. The financial events inherently present hierarchical structures, which could be represented as tree-structured schemes in real-life applications, and detecting events could be modeled as a hierarchical multi-label text classification problem, where an event is designated to a tree node with a sequence of hierarchical event category labels. Conventional hierarchical multi-label text classification methods usually ignore the hierarchical relationships existing in the event classification scheme, and treat the hierarchical labels associated with an event as uniform labels, where correct or wrong label predictions are assigned with equal rewards or penalties. In this paper, we propose a neural hierarchical multi-label text classification method, namely F-HMTC, for a financial application scenario with massive event category labels. F-HMTC learns the latent features based on bidirectional encoder representations from transformers, and directly maps them to hierarchical labels with a delicate hierarchy-based loss layer. We conduct extensive experiments on a private financial dataset with elaborately-annotated labels, and F-HMTC consistently outperforms state-of-art baselines by substantial margins. We will release both the source codes and dataset on the first author's repository.

The small and medium-sized enterprises (SMEs) are allowed to guarantee each other and form complex loan networks to receive loans from banks during the economic expansion stage. However, external shocks may weaken the robustness, and an accidental default may spread across the network and lead to large-scale defaults, even systemic crisis. Thus, predicting and rating the default contagion chains in the guarantee network in order to reduce or prevent potential systemic financial risk, attracts a grave concern from the Regulatory Authority and the banks. Existing credit risk models in the banking industry utilize machine learning methods to generate a credit score for each customer. Such approaches dismiss the contagion risk from guarantee chains and need extensive feature engineering with deep domain expertise. To this end, we propose a novel approach to rate the risk of contagion chains in the bank industry with the deep neural network. We employed the temporal inter-chain attention network on graph-structured loan behavior data to compute risk scores for the contagion chains. We show that our approach is significantly better than the state-of-the-art baselines on the dataset from a major financial institution in Asia. Besides, we conducted empirical studies on the real-world loan dataset for risk assessment. The proposed approach enabled loan managers to monitor risks in a boarder view and avoid significant financial losses for the financial institution.

Event representative learning aims to embed news events into continuous space vectors for capturing syntactic and semantic information from text corpus, which is benefit to event-driven quantitative investments. However, the financial market reaction of events is also influenced by the lead-lag effect, which is driven by internal relationships. Therefore, in this paper, we present a knowledge graph-based event embedding framework for quantitative investments. In particular, we first extract structured events from raw texts, and construct the knowledge graph with the mentioned entities and relations simultaneously. Then, we leverage a joint model to merge the knowledge graph information into the objective function of an event embedding learning model. The learned representations are fed as inputs of downstream quantitative trading methods. Extensive experiments on real-world dataset demonstrate the effectiveness of the event embeddings learned from financial news and knowledge graphs. We also deploy the framework for quantitative algorithm trading. The accumulated portfolio return contributed by our method significantly outperforms other baselines.

Commercial banks normally require Small and Medium Enterprises (SMEs) to provide their warranties when applying for a loan. If the borrower defaults, the guarantor is obligated to repay its loan. Such a guarantee system is designed to reduce delinquent risks, but may introduce a new dimension risk if more and more SMEs involve and subsequently form complex temporal networks. Monitoring the financial status of SMEs in these networks, and preventing or reducing systematic loan risk, is an area of great concern for both the regulatory commission and the banks. To allow possible actions to be taken in advance, this paper studies the problem of predicting repayment delinquency in the networked-guarantee loans. We propose a dynamic default prediction framework (DDPF), which preserves temporal network structures and loan behavior sequences in an end-to-end model. In particular, we design a gated recursive and attention mechanism to integrate both the loan behavior and network information. Then, we uncover risky warrant patterns by the learned weights, which effectively accelerate risk evaluation process. Finally, we conduct extensive experiments in a real-world loan risk control system to evaluate its performance, the results demonstrate the effectiveness of our proposed approach compared with state-of-the-art baselines.

Assessing and predicting the default risk of networked-guarantee loans is critical for the commercial banks and financial regulatory authorities. The guarantee relationships between the loan companies are usually modeled as directed networks. Learning the informative low-dimensional representation of the networks is important for the default risk prediction of loan companies, even for the assessment of systematic financial risk level. In this paper, we propose a high-order graph attention representation method (HGAR) to learn the embedding of guarantee networks. Because this financial network is different from other complex networks, such as social, language, or citation networks, we set the binary roles of vertices and define high-order adjacent measures based on financial domain characteristics. We design objective functions in addition to a graph attention layer to capture the importance of nodes. We implement a productive learning strategy and prove that the complexity is near-linear with the number of edges, which could scale to large datasets. Extensive experiments demonstrate the superiority of our model over state-of-the-art method. We also evaluate the model in a real-world loan risk control system, and the results validate the effectiveness of our proposed approaches.

Groups of enterprises can guarantee each other and form complex networks in order to try to obtain loans from banks. Monitoring the financial status of a network, and preventing or reducing systematic risk in case of a crisis, is an area of great concern for the regulatory commission and for the banks. We set the ultimate goal of developing a visual analytic approach and tool for risk dissolving and decision-making. We have consolidated four main analysis tasks conducted by financial experts: i) Multi-faceted Default Risk Visualization, whereby a hybrid representation is devised to predict the default risk and an interface developed to visualize key indicators; ii) Risk Guarantee Patterns Discovery. We follow the Shneiderman mantra guidance for designing interactive visualization applications, whereby an interactive risk guarantee community detection and a motif detection based risk guarantee pattern discovery approach are described; iii) Network Evolution and Retrospective, whereby animation is used to help users to understand the guarantee dynamic; iv) Risk Communication Analysis. The temporal diffusion path analysis can be useful for the government and banks to monitor the spread of the default status. It also provides insight for taking precautionary measures to prevent and dissolve systematic financial risk. We implement the system with case studies using real-world bank loan data. Two financial experts are consulted to endorse the developed tool. To the best of our knowledge, this is the first visual analytics tool developed to explore networked-guarantee loan risks in a systematic manner.

Studies show that motor imagery based Brain-Computer Interface (BCI) systems can be utilized therapeutically in stroke rehabilitation. Efficient decoding of subjects' motor intentions is essential in BCI-based rehabilitation systems to manipulate a neural prosthesis or other devices for motor relearning. However, due to cortical reorganization, the desynchro-nization potential evoked by the motor imagery of patients with brain lesions is quite different from that evoked by the motor imagery of normal subjects. These differences can be attributed to active cortex regions, frequency bands and amplitude. In this paper, we use a deep learning method to explore the EEG patterns of key channels and the frequency band for stroke patients. The EEG data is bandpass filtered into multiple sub-bands split by a sliding window strategy. Under these sub-bands, diverse spatial-spectral features are extracted and fed into a deep neural network for classification, uncovering the spectral patterns (bandpass filters) and spatial patterns (spatial weights). Experimental results from five stroke patients show that our method has higher classification accuracy than several state-of-the-art approaches. By tracking gradual changes in EEG patterns during rehabilitation, we try to uncover the neurophysiological plasticity mechanism in the impaired cortexes of stroke patients.

Portfolio risk control is vital to financial institutions: investors seek to build equities with the highest return but with minimum risk. However, a general phenomenon is significant comovement among many financial signals, such as stocks and futures. One investment strategy is to choose less correlated assets. Classic approaches quantifying such relationships in real financial markets make it difficult to exclude factors such as market trends and autocorrelation. In this paper, we propose a signal process perspective for quantitative measurement. A machine learning based algorithm is designed to model returns, taking account of market sensitivity, autocorrelation, and relationships with other stocks. We then extend the model training algorithm using regularized least square and gradient descent to estimate parameters. A penalty factor is designed in the optimization function to address extreme large negative returns. After denoising common factors, the learned pure relationship parameters are applied to construct a relationship matrix. Finally, we use this matrix to build portfolios by constrained optimization. Empirical experiments on two stock datasets show that the proposed method outperforms several state-of-the-art methods in terms of mean average precision and cumulative returns.

Networked-guarantee loans may cause the systemic risk related concern of the government and banks in China. The prediction of default of enterprise loans is a typical extremely imbalanced prediction problem, and the networked-guarantee make this problem more difficult to solve. Since the guaranteed loan is a debt obligation promise, if one enterprise in the guarantee network falls into a financial crisis, the debt risk may spread like a virus across the guarantee network, even lead to a systemic financial crisis. In this paper, we propose an imbalanced network risk diffusion model to forecast the enterprise default risk in a short future. Positive weighted k-nearest neighbors (p-wkNN) algorithm is developed for the stand-alone case - when there is no default contagious; then a data-driven default diffusion model is integrated to further improve the prediction accuracy. We perform the empirical study on a real-world three-years loan record from a major commercial bank. The results show that our proposed method outperforms conventional credit risk methods in terms of AUC. In summary, our quantitative risk evaluation model shows promising prediction performance on real-world data, which could be useful to both regulators and stakeholders.

Nowadays statistical machine learning is widely adopted in various domains such as data mining, image recognition and automated driving. However, software quality assurance for machine learning is still in its infancy. While recent efforts have been put into improving the quality of training data and trained models, this paper focuses on code-level bugs in the implementations of machine learning algorithms. In this explorative study we simulated program bugs by mutating Weka implementations of several classification algorithms. We observed that 8%-40% of the logically non-equivalent executable mutants were statistically indistinguishable from their golden versions. Moreover, other 15%-36% of the mutants were stubborn, as they performed not significantly worse than a reference classifier on at least one natural data set. We also experimented with several approaches to killing those stubborn mutants. Preliminary results indicate that bugs in machine learning code may have negative impacts on statistical properties such as robustness and learning curves, but they could be very difficult to detect, due to the lack of effective oracles.