The Design of a Universal Graph Model

Learning and reasoning over graphs is increasingly done by means of probabilistic models, e.g. exponential random graph models, graph embedding models, and graph neural networks. When graphs are modeling relations between people, however, they will inevitably reflect biases, prejudices, and other forms of inequity and inequality. An important challenge is thus to design accurate graph modeling approaches while guaranteeing fairness according to the specific notion of fairness that the problem requires. Yet, past work on the topic remains scarce, is limited to debiasing specific graph modeling methods, and often aims to ensure fairness in an indirect manner.

Exploring social networks through stochastic multilayer graph modeling

Indices of trophic and competitive relations in a planktonic network of a shallow, temperate lagoon. A graph and structural equation modeling approach

This letter addresses the optimization and complexity reduction of switch-reconfigured antennas. A new optimization technique based on graph models is investigated. This technique is used to minimize the redundancy in a reconfigurable antenna structure and reduce its complexity. A graph modeling rule for switch-reconfigured antennas is proposed, and examples are presented.

In the e-commerce environment, it is very common for consumers to select goods or services based on online reviews from social platforms. However, the behavior of some unscrupulous merchants who hire a “water army” to brush up on reviews of their products has been continuously exposed, which seriously misleads consumers’ purchasing decisions and undermines consumer trust. Until now, it has been a challenging task to accurately detect the “water army”, who could easily alter their behaviors or writing styles. The focus of this paper is on some collusive clues between members of the same social platform to propose a new graph model to detect the “water army”. First is the extraction of six kinds of user collusive relationships from two aspects: user content and user behavior. Further, the use of three aggregation methods on such collusive relationships generates a user collusive relationship factor (CRF), which is then used as the edge weight value in our graph-based water army detection model. In the combination of the graph grouping method and evaluation rules on candidate subgraphs, the graph model effectively detects multiple collusive groups automatically. The experimental results based on the Mafengwo platform show that the CRF generated from the coefficient of variation (CV) method demonstrates the best performance in detecting collusive groups, which provides some practical reference for the detection of “water armies” in an e-commerce environment.

The importance of bringing the relational data to other models and technologies has been widely debated. In special, Graph Database Management Systems (DBMS) have gained attention from industry and academia for their analytic potential. One of its advantages is to incorporate facilities to perform topological analysis, such as link prediction, centrality measures analysis, and recommendations. There are already initiatives to map from a relational database to graph representation. However, they do not take into account the different ways to generate such graphs. This work discusses how graph modeling alternatives from data stored in relational datasets may lead to useful results. The main contribution of this paper is towards managing such alternatives, taking into account that the graph model choice and the topological analysis to be used by the user. Experiments are reported and show interesting results, including modeling heuristics to guide the user on the graph model choice.

For parallelizing an algorithm and for mapping a given program onto a parallel or distributed system there are generally many possibilities. Performance models can help to predict which implementation and which mapping is the best for a given algorithm and for a given computer configuration. Stochastic graph modeling is an appropriate method, since the execution order of tasks, their runtime distribution, and branching probabilities are represented.

Financial fraudsters have increasingly taken advantage of digital financial services. Conventional rule- based and classical machine learning systems are not able to model the complex temporal, behavioral and relational patterns over the heterogeneous transaction data. In this paper, a deep learning–oriented study on real-world financial fraud detection schemes is conducted, integrating sequential models, convolutional structures, graph neural networks (GNNs), and hybrid ensembles to address the issue. Methods considered are temporal models (LSTM, GRU), attention- based transformers, CNN feature extractors for engineered transaction embeddings, GNNs for relational interactions (accounts, devices, merchants) and hybrid-pipeline using deep learners as well as gradient-boosted trees with XAI post- hoc explainers. Our complete end-to-end experimental pipeline includes four stages: robust preprocessing and anonymized features (2) class-imbalanced handling with combined sampling and loss-based strategies (focal-loss, class-weighted); (3) temporal plus graph model with a temporal-GNN encoder-decoder architecture, and (4) explainability using SHAP and counterfactual probe modules. Experiments on benchmark and proprietary-style datasets show that the proposed temporal-GNN hybrid achieves an average AUC-ROC of 0.976, precision at top-1% alerting of 0.42, and recall of 0.88, outperforming AUC by4-6% and recall by8-12% with competitive LSTM and LightGBM baselines. The method also decreases false-positive volume (alerts that need to be manually reviewed) by ~18% compared to gradient-boosted baselines. The research highlights deep learning’s potential to reveal multi- dimensional fraud epitasis alongside graph modelling and robust operational checks and controls.

We describe a graph-analytical framework for supporting data-driven decisions whether activities in registering companies point towards fraudulent behavior. The decision support is enabled by the design of a property graph model which public government data is transformed and loaded into. Fraud patterns are encoded as Cypher queries which return all instances of the patterns present in the data. All insights report on anonymized, real cases.

In this chapter, a proposal is made to design a new attack graph model based on shortest path normalization. The main aim of the proposed method is to provide the security for the dynamic network conditions. As the conventional attack, graph models are limited to only static networks, and they cannot ensure a secure data communication in the network. The proposed attack graph model tackles this problem by the accomplishment of a dynamic stochastic modeling on the network. To ensure the robustness, the proposed model is employed in two different cases; they are halt condition and varying condition. Simulation is conducted through several attributes, and the performance is measured through computational time and link probability.

The first-flush is the initial surge of highly concentrated mass of pollutants in storm runoff, mobilizing mass accumulated on land surface after dry periods. Like initial abstraction (Ia) in hydrology—which is water lost through interception, infiltration, evaporation, and surface depression storage and does not contribute to immediate surface runoff—the first-flush also appears at the watershed outlet but isn't a true sediment loss. It is crucial for designing on-site treatment facilities, allowing efficient isolation, storage, and treatment of stormwater in small catchments. Neglecting the first-flush may underestimate early-stage erosion. Developing sediment yield and graph models (SYMs and SGMs) is challenging due to nonlinearity of rainfall-runoff-sediment transport, unrealistic inputs, parameter sensitivities, oversimplified processes, landscape heterogeneity, human impacts, and data limitations. The SYMs predict total sediment load, while SGMs capture its temporal variation. Integrating the first flush with conventional SGM improves its accuracy and reliability by considering loose surface debris, and prior land use. The proposed improved first-flush driven sediment graph model with soil moisture accounting (IFF-SMA-SGM) couples the first-flush concept with the soil conservation service-curve number (SCS-CN) method (now the natural resource conservation service (NRCS) curve number method), power law, and Nash’s instantaneous unit sediment graph (IUSG) approach. It considers initial or existing soil moisture (V0) and Ia to work out the sediment yield. The model was calibrated and validated using a total of 16 sediment graphs (in an 8:8 ratio) recorded at Sub-watershed 6 (W6), Sub-watershed 7 (W7), and Sub-watershed 14 (W14), in the Goodwin Creek (GC) experimental watersheds in Oxford, Mississippi, USA, and the Mansara watershed in Uttar Pradesh, India. The model closely replicated the observed sediment graphs during both calibration and validation, demonstrating strong agreement in peak sediment load (QPS), total sediment load (QS), and time to peak sediment load (tPS). The model demonstrated high efficiency across most calibration and validation events, reflecting strong alignment between simulated and observed sediment yields. This study also highlights the IFF-SMA-SGM model's strong potential for accurate sediment yield prediction in hydro-meteorologically similar watersheds.

With the flourishing of the open-source software community, the problem of software vulnerabilities is becoming more and more serious. Hence, it is urgent to come up with an effective and efficient code vulnerability detection method. Source code vulnerability detection techniques used in practice today like symbolic execution and fuzz testing suffer from high false positives and low code coverage, respectively. Traditional machine-learning-based solutions fail to cope with the diversity of vulnerabilities. To overcome these drawbacks, a large number of deep-learning-based code vulnerability detection works have emerged, aiming at building powerful neural network models to fully learn code semantics and vulnerability patterns. In this survey, we mainly focus on code vulnerability detection approaches based on deep sequence modeling and graph modeling technologies. Our goal is to investigate how these two methods are applied to facilitate code vulnerability detection. We also go over current prevailing datasets that are used to evaluate detection models. At last, we identify the current challenges in this field and share our views on future work.

Modern graph datasets with structural complexity and uncertainties due to incomplete information or data variability require advanced modeling techniques beyond conventional graph models. This study introduces a memristive crossbar array (CBA)-based probabilistic graph model (C-PGM) utilizing Cu0.3Te0.7/HfO2/Pt memristors, which exhibit probabilistic switching, self-rectifying, and memory characteristics. C-PGM addresses the complexities and uncertainties inherent in structural graph data across various domains, leveraging the probabilistic nature of memristors. C-PGM relies on the device-to-device variation across multiple memristive CBAs, overcoming the limitations of previous approaches that rely on sequential operations, which are slower and have a reliability concern due to repeated switching. This new approach enables the fast processing and massive implementation of probabilistic units at the expense of chip area. In this study, the hardware-based C-PGM feasibly expresses small-scale probabilistic graphs and shows minimal error in aggregate probability calculations. The probability calculation capabilities of C-PGM are applied to steady-state estimation and the PageRank algorithm, which is implemented on a simulated large-scale C-PGM. The C-PGM-based steady-state estimation and PageRank algorithm demonstrate comparable accuracy to conventional methods while significantly reducing computational costs.

Knowledge tracing aims to trace students’ evolving knowledge states by predicting their future performance on concept-related exercises. Recently, some graph-based models have been developed to incorporate the relationships between exercises to improve knowledge tracing, but only a single type of relationship information is generally explored. In this article, we present a novel Dual Graph Ensemble learning method for Knowledge Tracing (DGEKT), which establishes a dual graph structure of students’ learning interactions to capture the heterogeneous exercise–concept associations and interaction transitions by hypergraph modeling and directed graph modeling, respectively. To combine the dual graph models, we introduce the technique of online knowledge distillation. This choice arises from the observation that, while the knowledge tracing model is designed to predict students’ responses to the exercises related to different concepts, it is optimized merely with respect to the prediction accuracy on a single exercise at each step. With online knowledge distillation, the dual graph models are adaptively combined to form a stronger ensemble teacher model, which provides its predictions on all exercises as extra supervision for better modeling ability. In the experiments, we compare DGEKT against eight knowledge tracing baselines on three benchmark datasets, and the results demonstrate that DGEKT achieves state-of-the-art performance.

In this article, we model electric power delivery networks as graphs, and conduct studies of two power transmission grids, i.e., the Nordic and the western states (U.S.) transmission grid. We calculate values of topological (structural) characteristics of the networks and compare their error and attack tolerance (structural vulnerability), i.e., their performance when vertices are removed, with two frequently used theoretical reference networks (the Erdös-Rényi random graph and the Barabási-Albert scale-free network). Further, we perform a structural vulnerability analysis of a fictitious electric power network with simple structure. In this analysis, different strategies to decrease the vulnerability of the system are evaluated. Finally, we present a discussion on the practical applicability of graph modeling.