graph neural networks

Simplicial Graph-Based Detection and Localization of Cyber Attacks Against Large-Scale Smart Grids

Ensuring the resilience of large-scale power systems against cyber attacks is critical to maintaining the stability of modern cyber-physical energy systems. Existing machine learning (ML)-based detection frameworks predominantly focus on pairwise node interactions (i.e., edges) and often overlook capturing higher-order topological structures (i.e., simplicial complexes) that emerge within power networks. This paper introduces a novel simplicial recurrent graph neural network (SRGNN) for attack detection and localization in large-scale smart grid infrastructure. Unlike conventional graph models, SRGNN incorporates higher-order simplicial interactions to capture group-wise dependencies among buses and transmission lines, allowing the model to better reflect the multi-scale dynamics of grid operation and control. We investigate the robustness of our model against benchmark and proposed complex simplicial-based attack node selection strategies. Our extensive experiments on large-scale transmission power networks (<inline-formula> <tex-math notation="LaTeX">$2{,}869$ </tex-math></inline-formula>-bus, <inline-formula> <tex-math notation="LaTeX">$9{,}241$ </tex-math></inline-formula>-bus, and <inline-formula> <tex-math notation="LaTeX">$70{,}000$ </tex-math></inline-formula>-bus systems) demonstrate that the proposed SRGNN model outperforms ML-based benchmark models, achieving superior detection and localization performance by 9–39% and 8–35%, respectively, in detection rate against complex attacks. These results underscore the importance of modeling higher-order topological structures for robust and scalable security in power systems.

HVAC subsystem-aligned explainable graph neural network mixture of experts (GNN-MoE) for multi-horizon work order prediction

Dynamic response analysis method of parametric structures based on modal reduction basis predicted by physics-informed graph neural networks

Full-Reference Point Cloud Quality Assessment Using Graph Neural Network-Based Regression

Manifold-Aware General Coded Computing for Straggler-Resilient Distributed Computing

Existing coded-computing designs do not explicitly exploit the intrinsic structure of the input data. In communication systems, statistical structure and redundancy are often removed through source coding (or compression) before channel coding is applied. This principle, however, does not transfer directly to coded computation. In many computational tasks, particularly in machine learning, the structure of the data is precisely what the computation seeks to exploit to infer outputs or learn meaningful patterns. Consequently, coded-computing schemes should preserve and leverage this structure in their code design, rather than ignoring or eliminating it through source coding. This observation motivates a different perspective on code construction. In many channel-coding schemes, such as Reed-Solomon codes, coded symbols are generated by evaluating a low-dimensional algebraic representation at selected points. In contrast, many high-dimensional datasets naturally concentrate near low-dimensional manifolds. In this paper, we exploit this intrinsic geometry by designing coded samples that follow the natural manifold of the data, rather than imposing an artificial low-dimensional structure unrelated to the data distribution. Inspired by graph-based manifold learning, we propose a manifold-aware encoding strategy for general coded computing (GCC). Experiments on neural network inference and high-dimensional polynomial evaluation demonstrate that the proposed strategy consistently and significantly reduces the mean squared recovery error under straggling compared with standard GCC.

Why Multi-Layer Message Passing Works: Completeness Theory for Graph Neural Network Interatomic Potentials

We prove that the Hypergraph Neural Network, an invariant architecture with 3-body message passing, is a universal approximator for potential energy surfaces. Our main contribution is a multi-layer completeness theory. We show that $L$ layers of message passing on sparse, cutoff-based graphs achieve the same representational power as having access to the full $L$-hop neighborhood, provided the configurations are generic, satisfy an overlap condition and a connectivity condition. This provides the first rigorous justification for the common practice of using multi-layer message passing with a per-layer cutoff smaller than the physical interaction range, the setting used by virtually all practical graph neural network based machine-learned interatomic potentials. As immediate consequences, we show that both DPA3 and CHGNet architectures inherit universal approximation.

A Systematic Literature Review on Software Testing Prediction Models

Software testing prediction models play a critical role in improving software quality, reducing faults, and optimizing testing. Although past studies show that these models have improved over time, little focus has been given to them. Many studies rely on limited datasets that do not fully capture the complexity of real software systems, which limits how well the models can generalize. There is also insufficient evidence on how these models use existing datasets in practical, real-life settings, since most evaluations are conducted under controlled or experimental conditions. In addition, key aspects such as interpretability, generalizability, and practical usability are still not adequately addressed, which reduces trust and slows down adoption in practice. This study presents a systematic literature review of recent empirical research on predictive approaches in software testing, focusing on machine learning, deep learning, and hybrid techniques. A structured methodology was used, including clearly defined inclusion and exclusion criteria, systematic searches across major academic databases, quality assessment, and data extraction from 22 selected studies. The analysis considered model types, datasets, feature methods, evaluation metrics, and methodological approaches. The findings show a shift from traditional statistical models to advanced artificial intelligence techniques, including graph neural networks, contrastive learning, and deep fuzzy clustering. In addition, optimization and data balancing techniques improve predictive performance, while explainable artificial intelligence enhances model interpretability. However, challenges such as limited cross-project generalization and insufficient industrial validation still exist. 

Energy Consumption Graph Inference for Financing Risk Prediction in Industrial Enterprises

Energy cost financing is increasingly used by industrial enterprises to manage electricity, gas, steam, and fuel expenses during periods of high production demand or cash-flow pressure. However, abnormal energy consumption, declining production intensity, delayed utility payments, and unstable downstream orders may signal future repayment risk before conventional financial indicators change. This study proposes an energy consumption graph inference model for financing risk prediction in industrial enterprises. The model constructs a heterogeneous graph linking enterprises, utility providers, production facilities, electricity meters, gas accounts, invoices, financing contracts, downstream customers, and overdue payment records. A graph neural network is used to encode enterprise production-energy dependency, while a knowledge reasoning module identifies risk chains involving shrinking energy use, abnormal peak-valley consumption, repeated utility arrears, customer-order decline, and concentrated financing exposure. Experiments are conducted on an industrial energy finance dataset containing 41,600 enterprises, 126,000 utility accounts, 2.84 million monthly energy bills, 760,000 customer-supplier relations, 68,000 financing contracts, and 8,420 confirmed repayment-risk cases over 48 months. Compared with a financial-ratio baseline, the proposed model shortens median warning time before repayment deterioration from 84 days to 29 days. Graph inference identifies 5,930 energy-production risk chains and 2,170 enterprises with simultaneous order decline and utility arrears. Risk-path aggregation reduces 13,800 raw enterprise alerts to 3,640 investigation cases. Full monthly portfolio assessment is completed in 15.2 minutes, with a median inference latency of 41 ms per enterprise node. These results indicate that energy consumption graph inference can improve industrial financing risk prediction by connecting production activity, utility payment behavior, and enterprise relationship networks.

Default Contagion Detection in Industrial Enterprise Networks Using Multi-Relational Graph Neural Inference

The results demonstrate that multi-relational graph neural inference can capture default transmission mechanisms that are difficult to observe through firm-level indicators alone.

A compact physics-informed graph neural network for real-time wind-pressure prediction of tensioned membrane roofs across multiple wind directions

Designing a Good Virtual Node: Addressable and Cardinality-Preserving Global Memory for Message Passing Architectures

Virtual nodes give message-passing neural networks a simple global communication route, but the standard node--VN--node pipeline compresses the graph into one homogeneous state and broadcasts it identically to every node. Building on the Two-Radius analysis of Mishayev et al., we ask how auxiliary virtual memory can relieve this finite-capacity bottleneck without self-attention. We identify two requirements. First, the global memory should be factorized into independently writable and readable states: this can be achieved using addressable cross-attention slots. Second, addressability alone does not preserve multiplicity, because softmax attention is invariant to uniform replication. Inserting each slot query as a private key/value anchor recovers the discarded normalization mass and yields, on bounded color domains, an injective multiset representation able to implement a 1-WL refinement. Experiments on multiplicity-aware Two-Radius, motif counting, and constrained link-set prediction support this addressable and cardinality-preserving virtual memory at (O(nMd)) arithmetic cost.

Leveraging Cyber-Physical Security Solutions Blended With Machine Learning for Advanced IoT Botnet Detection

The proliferation of Internet of Things (IoT) ecosystems has significantly increased the attack surface of cyber-physical systems, leading to the emergence of large-scale botnets that exploit device vulnerabilities for distributed and persistent attacks. This survey comprehensively reviews state-of-the-art techniques that integrate cyber-physical security solutions with machine learning (ML) for advanced detection and mitigation of IoT botnets. It categorizes existing methods into networkcentric, host-based, and hybrid cyber-physical detection frameworks, emphasizing their detection granularity, scalability, and computational feasibility in resource-constrained environments. The paper systematically analyzes supervised, unsupervised, and deep learning approaches, ranging from Random Forests and Support Vector Machines to Autoencoders, LSTMs, and Graph Neural Networks, highlighting their adaptability to evolving botnet behaviours and zero-day threats. Furthermore, the survey explores the integration of federated learning, edge computing, and software-defined networking (SDN) to enable distributed, privacy-preserving, and realtime detection architectures. Key challenges, including data imbalance, adversarial resilience, explainability, and cross-domain generalization, are critically discussed. Finally, this work outlines a taxonomy of cyber-physical and ML-based IoT botnet detection models and identifies future research directions toward autonomous, adaptive, and explainable cyber-physical defense systems.