The rapid evolution of digital currency trading, fueled by the integration of blockchain technology, has led to both innovation and the emergence of smart Ponzi schemes. A smart Ponzi scheme is a fraudulent investment operation in smart contract that uses funds from new investors to pay returns to earlier investors. Traditional Ponzi scheme detection methods based on deep learning typically rely on fully supervised models, which require large amounts of labeled data. However, such data is often scarce, hindering effective model training. To address this challenge, we propose a novel contrastive learning framework, CASPER (Contrastive Approach for Smart Ponzi detectER with more negative samples), designed to enhance smart Ponzi scheme detection in blockchain transactions. By leveraging contrastive learning techniques, CASPER can learn more effective representations of smart contract source code using unlabeled datasets, significantly reducing both operational costs and system complexity. We evaluate CASPER on the XBlock dataset, where it outperforms the baseline by 2.3% in F1 score when trained with 100% labeled data. More impressively, with only 25% labeled data, CASPER achieves an F1 score nearly 20% higher than the baseline under identical experimental conditions. These results highlight CASPER's potential for effective and cost-efficient detection of smart Ponzi schemes, paving the way for scalable fraud detection solutions in the future.

Temporal knowledge graph (TKG) extrapolation aims to predict future, previously unseen events based on historical facts. However, most existing temporal knowledge graph extrapolation methods either focus on global cyclic regularities or on local adjacent transitions. These methods overlook the multi-granularity nature of temporal signals and often rely on heuristic fusion schemes that are sensitive to noise. To address these limitations, we propose MTRM, a Multi-granularity Trend Retrieval and Modeling framework for TKG extrapolation. Specifically, we first apply semantic clustering to retrieve a compact set of long-term trend clusters from sequences of historical subgraphs, capturing enduring interaction patterns. Then, we introduce a trend-aware attention-enhancing evolution module with an auxiliary contrastive loss to learn fine-grained short-term dynamics by aligning each hidden state with its subsequent subgraph. To integrate information at different granularities, we design a multi-granularity attention layer that adaptively fuses the long-term clusters with the short-term trend states for each query entity. Additionally, an inter-granularity contrastive objective is employed to align these representations and enhance robustness to noisy snapshots. Experiments on four benchmark datasets demonstrate that MTRM outperforms state-of-the-art baselines by up to 5.89% in mean reciprocal rank (MRR), indicating improved robustness on large-scale noisy event streams. Moreover, MTRM provides interpretable insights into how long- and short-term temporal granularities jointly drive future-event prediction.

Current rumor detectors exhibit limitations in fully exploiting responses to the source tweet as essential public opinions, and in explaining and indicating the reliability of the results obtained. Additionally, the joint utilization of both responses and the multimodal source content for detection presents challenges due to the heterogeneous nature of the data points. In this work, to address the first challenge, we initially prompt the Large Language Model (LLM) with both multimodal source content and the corresponding response set to extract contrasting evidence to enable maximal utilization of informative responses. To overcome the second challenge, we introduce an uncertainty-aware evidential evaluator to assess the evidence intensity from the multimodal source content and dual-sided reasoning, from which the final prediction is derived. As we model the second-order probability, we can effectively indicate the model’s uncertainty (i.e., the reliability) of the results. The reasoning from the correct perspective also serves as a natural language-based explanation. To this end, the third challenge is also addressed as we fully leverage the available resources. Extensive experiments validate the effectiveness, uncertainty awareness in predictions, helpful explainability for human judgment, and superior efficiency of our approach compared to contemporary works utilizing LLMs.

Blockchain technology has catalyzed innovation in digital currency trading but has also facilitated the emergence of smart Ponzi schemes, wherein returns to earlier investors are fraudulently paid using capital from newer participants. Deep learning-based detection methods show promise in identifying these schemes. However, they rely heavily on large amounts of labeled data, which are often limited in real-world applications. This paper introduces Semi-Supervised Smart Ponzi Identification (S3PI), a novel framework designed to address data scarcity by integrating contrastive learning for robust contract representation. The proposed semi-supervised approach reduces reliance on extensive labels, thereby lowering operational complexity and cost, while maintaining high detection accuracy. Experiments on the XBlock dataset indicate that S3PI achieves strong performance with only 25% labeled data, surpassing the best baseline by approximately 10%. Further experiments on multiple datasets evaluate the model’s transferability and generalization across different datasets, demonstrating its adaptability to diverse data distributions and effectiveness in varied settings. This result highlights its effectiveness in detecting Ponzi schemes under label-scarce conditions, demonstrating the potential of semi-supervised learning in smart contract fraud detection. These findings suggest that S3PI offers a scalable and efficient solution for fraud detection in blockchain systems.

Temporal knowledge graph reasoning is an essential task that holds immense value in diverse real-world applications. Existing studies mainly focus on leveraging structural and sequential dependencies, excelling in tasks like entity and link prediction. However, they confront a notable interpretability gap in their predictions, a pivotal facet for comprehending model behavior. In this study, we propose an innovative method, LSGAT, which not only exhibits remarkable precision in entity predictions but also enhances interpretability by identifying pivotal historical events influencing event predictions. LSGAT enables concise explanations for prediction outcomes, offering valuable insights into the otherwise enigmatic "black box" reasoning process. Through an exploration of the implications of the most influential events, it facilitates a deeper understanding of the underlying mechanisms governing predictions.

Source localization, as a reverse problem of information dissemination on graphs, is crucial for tracking social rumors, detecting computer viruses, and identifying epidemic spreaders. However, existing methods face challenges due to the inherent uncertainty of graph diffusion, as the same diffused observations may start with diverse sources. Furthermore, general graph diffusion models did not consider important properties of the diffusion process. To address these issues, we propose a denoising diffusion probabilistic model (DDPM)-based source localization framework, DDSL. In this framework, we consider two distinct characteristics of information dissemination, namely source prominence and monotone increasing, and present a source localization-oriented invertible graph neural network (GNN). To capture the propagation uncertainties of sources, we design a DDPM-based source generator to generate effective and diverse sources for enhancing model’s robustness. Our experiments demonstrate the effectiveness of the proposed model in improving source localization performance.

Accurately predicting information diffusion is critical for a vast range of applications. Existing methods generally consider user re-sharing behaviors to be driven by a single intent, and/or assume cascade temporal influence to be unchanged, which might not be consistent with real-world scenarios. To address these issues, we propose a self-supervised disentanglement framework (DisenIDP) for information diffusion prediction. First, we construct intent-aware hypergraphs to capture users' potential intents from different perspectives, and then perform the light hypergraph convolution to adaptively activate disentangled intents. Second, we extract long-term and short-term cascade influence via independent attention-based encoders. Finally, we set a self-supervised disentanglement task to alleviate the information loss and learn better-disentanglement representations. Extensive experiments conducted on two real-world social datasets demonstrate that DisenIDP outperforms state-of-the-art models across several settings.

The exponential growth of digital information has amplified the necessity for effective rumor detection on social media. However, existing approaches often neglect the inherent noise and uncertainty in rumor propagation, leading to obscure learning mechanisms. Moreover, current deep-learning methodologies, despite their top-tier performance, are heavily dependent on supervised learning, which is labor-intensive and inefficient. Their prediction credibility is also questionable. To tackle these issues, we present a new framework, TrustRD, for reliable rumor detection. Our framework incorporates a self-supervised learning module, designed to derive interpretable and informative representations with less reliance on large labeled data sets. A downstream model based on Bayesian networks, which is further refined with adversarial training, enhances performance while providing a quantifiable trustworthiness assessment of results. Our methods' effectiveness is confirmed through experiments on two benchmark datasets.

Detecting rumors on social media has grown in importance as the amount of digital material available online grows quickly. Recent approaches to rumor detection heavily rely on supervised learning, which requires a significant amount of labeled data for training and provides limited interpretability of prediction results. Furthermore, these approaches show a lack of robustness and are vulnerable to overfitting. In this paper, we propose a novel framework Self-Supervised Rumor Detection with Augmented Variational Graphs (SSRD-AVG) from a self-supervised learning (SSL) view, which employs a pre-trained generative model to facilitate data augmentation with enhanced interpretability. Specifically, the generative model first harnesses neighboring information to extract salient features for rumor propagation structures and user engagement. Then, the obtained augmented features are subsequently utilized for self-supervised learning. Finally, we fine-tuned the Graph Neural Network (GNN) with labeled data for rumor detection. Comprehensive experiments demonstrate that our model attains state-of-the-art performance and remains robust in real-world scenarios.

Cross-network alignment aims at identifying users who participate in different social networks, which benefits a variety of downstream social applications such as precise content delivery, fraud detection, and content/user recommender systems. Recent advances in network representations and graph neural networks have spurred various network structure-based methods for capturing underlying node similarities across social networks, thereby addressing the network alignment problem. However, most of the existing solutions rely on embedding methods that compute node similarity in Euclidean space, resulting in severe distortion or semantic loss when representing real-world social networks, which are usually scale free and with hierarchical structures. We address these issues by presenting a novel model: Hyperbolic Graph Embedding for Network Alignment (HGENA), which learns the structural semantics more efficiently by embedding nodes in hyperbolic space instead of Euclidean. HGENA overcomes the scalability issue since it requires far fewer dimensions in Riemannian manifolds and increases the capability of learning hierarchical structures, while enabling smaller distortion for tree-liked networks to facilitate node alignment. We also introduce alternative network mapping functions to compute node similarity across-network based on its distance on the Poincare ball. Experimental evaluations conducted on real world datasets demonstrate that HGENA achieves superior performance on social network alignment, especially for more tree-liked networks.

User identity linkage (UIL) aims to link identical users engaging in multiple social networks. It has received considerable attention in both academia and industry due to its profound implications for multiple applications. Although existing approaches have achieved promising progress in UIL using various graph learning methods, they usually require a large number of labeled anchor nodes which, however, are difficult to obtain in real-world social platforms due to privacy issues. We introduce a novel UIL model NWUIL (Network Wasserstein learning for UIL) to identify anchor users across social networks in a fully unsupervised manner. Instead of point vector embedding of nodes as in previous methods, NWUIL captures node distribution in Wasserstein space with graph neural networks. We also propose to reformulate the UIL task as an optimal network transport problem, and then introduce an unsupervised mapping process based on the network Wasserstein distance for UIL. In this way, our method not only improves the anchor node aligning accuracy but also alleviates the issues caused by insufficient labeled anchor nodes. We conduct extensive experiments using real-world datasets, and demonstrate that NWUIL significantly outperforms existing unsupervised baselines while showing competitive performance as some state-of-the-art supervised approaches.

Bounded Model Checking (BMC) converts a verification problem within a user-specified bound into satisfiability checks of propositional formulas. As the bound deepens, the formulas become larger in size and harder to solve. In this paper, we propose a hybrid approach in which stateless explicit-state exploration (SESE) is integrated into the BMC process to improve the scalability and performance of BMC for the verification of properties expressed in Linear Temporal Logic (LTL). Specifically, SESE is utilized to traverse, under the constraints of Bounded-Context Switching (BCS), the state space of a system design and memorize legal execution paths. These paths are classified according to heuristic state predicates into path clusters, which are then encoded into propositional formulas representing, together with the encoded formula for an LTL property, independent BMC instances. Such BMC instances are solved with SMT solvers running on mutilcores in parallel. Once a counterexample is found for one of the instances, the entire model checking (SESE as well as BMC) terminates. This hybrid checking procedure progresses in an incremental fashion until either a counterexample is found or the user-specified bound is reached. We have implemented this proposed hybrid approach in a tool called Garakabu2 with Yices 2 as its back-end solver. The experimental results show that Garakabu2 outperforms significantly the state-of-the-art BMC methods implemented in SAL for both safety and liveness properties.

We propose a hybrid approach to improving the verification performance of SMT-based bounded model checking for LTL properties. In this approach, stateless explicit-state exploration is utilized to traverse, under the constraints of bounded context switches, the state space of a system design and memorize legal execution paths. These paths are classified according to certain predicates into path clusters, which are then encoded into propositional formulas representing, together with the encoded formula for an LTL property, independent BMC instances. Such BMC instances are solved with SMT solvers running on mutilcores in parallel. Once a counterexample is found for one of the instances, the entire model checking terminates. This hybrid checking procedure progresses in an incremental fashion until either a counterexample is found or the user-specified bound is reached. We have implemented this proposed hybrid approach in a tool called Garakabu2 with CVC4 as its backend solver. The experimental results show that Garakabu2 often outperforms the state-of-the-art pure BMC methods implemented in SAL infinite bounded model checker for both safety and liveness properties.

Hierarchical State Transition Matrix (HSTM) is a table-based modeling language for developing designs of software systems. We have proposed a Satisfiability Modulo Theory (SMT) based Bounded Model Checking (BMC) approach in [1] to provide formal verification supports for conducting rigorous and automatic analysis to improve reliability of HSTM designs. In this paper, we continue that work by developing and evaluating approaches to accelerating BMC of HSTM designs. The approaches center around an unrolled Bounded Reach ability Tree (BRT) of a HSTM design that is built with stateless explicit state exploration. Specifically, reach ability of invalid cells (representing undesired states) of a HSTM design, which occurs within the bound concerned, could be discovered during construction of the BRT, and furthermore, if no such occurrence, the constructed BRT could be utilized to rule out unnecessary subformulas of a BMC instance for verification of LTL properties. We have implemented these approaches in a tool called Garakabu2 with the state-of-the-art SMT solver CVC3 as its back-ended solver. Our preliminary experiments show that verification could be accelerated substantially.