Search Results (225)

To thoroughly investigate the distribution of heavy rain disaster risks in the Beijing–Tianjin–Hebei region, this paper analyzes the spatiotemporal evolution characteristics of heavy rain disaster-inducing factors. Based on disaster system theory, we constructed a heavy rain disaster risk assessment framework from four dimensions. We improved the application of model-agnostic meta-learning (MAML) in hyperparameter optimization for the random forest (RF) algorithm, thereby developing the MAML-RF heavy rain disaster risk assessment model. This model was compared with the SCV-RF model, which is based on random search and cross-validation (SCV), to determine which model had higher accuracy. Then we introduced the SHAP (Shapley additive explanations) interpretability algorithm to quantify the impact of each risk factor. The results indicate that (1) the annual characteristics of heavy rain days and rainfall amounts show a significant upward trend over the past 17 years; (2) the MAML-RF model improved the accuracy and precision of heavy rain disaster risk simulation by 4.44% and 3.71%, respectively, and reduced training time by 27.95% compared to the SCV-RF model; and (3) the SHAP interpretability algorithm results show that the top five influential factors are the number of heavy rain days, rainfall amount, slope, drainage pipe density, and impervious surface ratio. Full article

While deep learning-based image matching methods excel at extracting high-level semantic features from remote sensing data, their performance degrades significantly under cross-daylight conditions and wide-baseline geometric distortions, particularly in multi-source street-view scenarios. This paper presents a novel illumination-invariant framework that synergistically integrates geometric topology and semantic consistency to achieve robust multi-view matching for cross-daylight urban perception. We first design a self-supervised learning paradigm to extract illumination-agnostic features by jointly optimizing local descriptors and global geometric structures across multi-view images. To address extreme perspective variations, a homography-aware transformation module is introduced to stabilize feature representation under large viewpoint changes. Leveraging a graph neural network with hierarchical attention mechanisms, our method dynamically aggregates contextual information from both local keypoints and semantic topology graphs, enabling precise matching in occluded regions and repetitive-textured urban scenes. A dual-branch learning strategy further refines similarity metrics through supervised patch alignment and unsupervised spatial consistency constraints derived from Delaunay triangulation. Finally, a topology-guided multi-plane expansion mechanism propagates initial matches by exploiting the inherent structural regularity of street scenes, effectively suppressing mismatches while expanding coverage. Extensive experiments demonstrate that our framework outperforms state-of-the-art methods, achieving a 6.4% improvement in matching accuracy and a 30.5% reduction in mismatches under cross-daylight conditions. These advancements establish a new benchmark for reliable multi-source image retrieval and localization in dynamic urban environments, with direct applications in autonomous driving systems and large-scale 3D city reconstruction. Full article

The construction industry frequently encounters safety hazards, with falls related to undetected openings being a major cause of fatalities. Identifying unstructured openings using computer vision is challenging due to their unpredictable nature and the difficulty of acquiring large labeled datasets in dynamic construction environments. Conventional deep learning methods require substantial data, limiting their applicability. Few-shot learning (FSL) offers a promising alternative by enabling models to learn from limited examples. This study investigates the effectiveness of an FSL approach, specifically model-agnostic meta-learning (MAML), enhanced with domain-specific attributes, for identifying unstructured openings with minimal labeled data. We developed and evaluated an attribute-enhanced MAML framework under various few-shot conditions (k-way, n-shot) and compared its performance against conventional supervised fi-ne-tuning. The results demonstrate that the proposed FSL model achieved high classification accuracy (over 90.5%) and recall (over 85.5%) using only five support shots per class. Notably, the FSL approach significantly outperformed supervised fine-tuning methods under the same limited data conditions, exhibiting substantially higher recall crucial for safety monitoring. These findings validate that FSL, augmented with relevant attributes, provides a data-efficient and effective solution for monitoring unpredictable hazards like unstructured openings, reducing the reliance on extensive data annotation. This research contributes valuable insights for developing adaptive and robust AI-powered safety monitoring systems in the construction domain. Full article

Traditional construction safety management, reliant on manual inspections and heuristic judgments, increasingly fails to address the dynamic, multi-dimensional risks of modern projects, perpetuating fragmented safety governance and reactive hazard mitigation. This study proposes an integrated building information modeling (BIM)–AI platform to unify safety supervision across the project lifecycle, synthesizing spatial-temporal data from BIM with AI-driven probabilistic models and IoT-enabled real-time monitoring for sustainable construction safety management. Employing a Design Science Research methodology, the platform’s phase-agnostic architecture bridges technical–organizational divides, while the Multilayer Neural Risk Coupling Assessment framework quantifies interdependencies among structural, environmental, and human risk factors. Prototype testing in real-world projects demonstrates improved risk detection accuracy, reduced reliance on manual processes, and enhanced cross-departmental collaboration. The system transitions safety regimes from compliance-based protocols to proactive, data-empowered governance. This approach offers scalability across diverse projects. The BIM-AI intelligent fusion platform proposed in this study builds an intelligent construction paradigm with synergistic development of safety governance and sustainability through whole lifecycle risk coupling analysis and real-time dynamic monitoring, which realizes a proactive safety supervision system while significantly reducing construction waste and accident prevention mechanisms. Full article

The advancement of pervasive systems has made distributed real-world data across multiple devices increasingly valuable for training machine learning models. Traditional centralized learning approaches face limitations such as data security concerns and computational constraints. Federated learning (FL) provides privacy benefits but is hindered by challenges like data heterogeneity (Non-IID distributions) and noise heterogeneity (mislabeling and inconsistencies in local datasets), which degrade model performance. This paper proposes a model-agnostic, quality-driven approach, called DQFed, for training machine learning models across distributed and diverse client datasets while preserving data privacy. The DQFed framework demonstrates improvements in accuracy and reliability over existing FL frameworks. By effectively addressing class imbalance and noise heterogeneity, DQFed offers a robust and versatile solution for federated learning applications in diverse fields. Full article

The Controller Area Network (CAN) facilitates efficient communication among vehicle components. While it ensures fast and reliable data transmission, its lightweight design makes it susceptible to data manipulation in the absence of security layers. To address these vulnerabilities, machine learning (ML)-based intrusion detection systems (IDS) have been developed and shown to be effective in identifying anomalous CAN traffic. However, these models often function as black boxes, offering limited transparency into their decision-making processes, which hinders trust in safety-critical environments. To overcome these limitations, this paper proposes a novel method that combines Local Interpretable Model-agnostic Explanations (LIME) with a vision-language model (VLM) to generate detailed textual interpretations of an ML-based CAN IDS. This integration mitigates the challenges of visual-only explanations in traditional XAI and enhances the intuitiveness of IDS outputs. By leveraging the multimodal reasoning capabilities of VLMs, the proposed method bridges the gap between visual and textual interpretability. The method supports both global and local explanations by analyzing feature importance with LIME and translating results into human-readable narratives via VLM. Experiments using a publicly available CAN intrusion detection dataset demonstrate that the proposed method provides coherent, text-based explanations, thereby improving interpretability and end-user trust. Full article

Despite the growing adoption of Property Graph Databases, like Neo4j, interacting with them remains difficult for non-technical users due to the reliance on formal query languages. Natural Language Interfaces (NLIs) address this by translating natural language (NL) into Cypher. However, existing solutions are typically limited to high-resource languages; are difficult to adapt to evolving domains with limited annotated data; and often depend on Machine Learning (ML) approaches, including Large Language Models (LLMs), that demand substantial computational resources and advanced expertise for training and maintenance. We address these limitations by introducing a novel dependency-based, training-free, schema-agnostic Natural Language Interface (NLI) that converts NL queries into Cypher for querying Property Graphs. Our system employs a modular pipeline-integrating entity and relationship extraction, Named Entity Recognition (NER), semantic mapping, triple creation via syntactic dependencies, and validation against an automatically extracted Schema Graph. The distinctive feature of this approach is the reduction in candidate entity pairs using syntactic analysis and schema validation, eliminating the need for candidate query generation and ranking. The schema-agnostic design enables adaptation across domains and languages. Our system supports single- and multi-hop queries, conjunctions, comparisons, aggregations, and complex questions through an explainable process. Evaluations on real-world queries demonstrate reliable translation results. Full article

Fault diagnosis in large-scale systems presents significant challenges due to the complexity and high dimensionality of data, as well as the scarcity of labeled fault data, which are hard to obtain during the practical operation process. This paper proposes a novel approach, called Multi-Variable Meta-Transformer (MVMT), to tackle these challenges. In order to deal with the multi-variable time series data, we modify the Transformer model, which is the currently most popular model on feature extraction of time series. To enable the Transformer model to simultaneously receive continuous and state inputs, we introduced feature layers before the encoder to better integrate the characteristics of both continuous and state variables. Then, we adopt the modified model as the base model for meta-learning—more specifically, the Model-Agnostic Meta-Learning (MAML) strategy. The proposed method leverages the power of Transformers for handling multi-variable time series data and employs meta-learning to enable few-shot learning capabilities. The case studies conducted on the Tennessee Eastman Process database and a Power-Supply System database demonstrate the exceptional performance of fault diagnosis in few-shot scenarios, whether based on continuous-only data or a combination of continuous and state variables. Full article

In this work, the position control of a cable-driven soft robot is proposed through the approximation of its kinematic model. This approximation is derived from artificial learning rules via neural networks and experimentally observed data. To improve the learning process, a combination of active sampling and Model Agnostic Meta Learning is carried out to improve the data based model to be used in the control stage through the inverse velocity kinematics derived from the data based modeling along with a self differentiation procedure to come up with the pseudo inverse of the robot Jacobian. The proposal is verified in a designed and constructed cable-driven soft robot with three actuators and position measurement through a vision system with three-dimensional motion. Some preliminary assessments (tension and repeatability) were performed to validate the robot movement generation, and, finally, a 3D reference trajectory was tracked using the proposed approach, achieving competitive tracking errors. Full article

Background: Investment decisions in stocks are one of the most complex tasks due to the uncertainty of which stocks will increase or decrease in their values. A diversified portfolio statistically reduces the risk; however, stock choice still substantially influences the profitability. Methods: This work proposes a methodology to automate investment decision recommendations with clear explanations. It utilizes generative AI, guided by prompt engineering, to interpret price predictions derived from neural networks. The methodology also includes the Artificial Intelligence Trust, Risk, and Security Management (AI TRiSM) model to provide robust security recommendations for the system. The proposed system provides long-term investment recommendations based on the financial fundamentals of companies, such as the price-to-earnings ratio (PER) and the net margin of profits over the total revenue. The proposed explainable artificial intelligence (XAI) system uses DeepSeek for describing recommendations and suggested companies, as well as several charts based on Shapley additive explanation (SHAP) values and local-interpretable model-agnostic explanations (LIMEs) for showing feature importance. Results: In the experiments, we compared the profitability of the proposed portfolios, ranging from 8 to 28 stock values, with the maximum expected price increases for 4 years in the NASDAQ-100 and S&P-500, where both bull and bear markets were, respectively, considered before and after the custom duties increases in international trade by the USA in April 2025. The proposed system achieved an average profitability of 56.62% while considering 120 different portfolio recommendations. Conclusions: A t-Student test confirmed that the difference in profitability compared to the index was statistically significant. A user study revealed that the participants agreed that the portfolio explanations were useful for trusting the system, with an average score of 6.14 in a 7-point Likert scale. Full article

The escalating global incidence of malware presents critical cybersecurity threats to manufacturing, automation, and industrial process control systems. Given the fast-developing web applications and IoT devices in use by industry operations, securing a transparent and effective malware detection mechanism has become imperative to operational resilience and data integrity. Classical methods of malware detection are conventionally opaque “black boxes” with limited transparency, thus eroding trust and hindering deployment in security-sensitive contexts. In this respect, this research proposes MAL-XSEL—a malware detection framework using an explainable stacking ensemble learning approach for performing high-accuracy classification and interpretable decision-making. MAL-XSEL explicates the model predictions through Shapley additive explanations (SHAP) and local interpretable model-agnostic explanations (LIME), which enable security analysts to validate how the detection logic works and prioritize the features contributing to the most critical threats. Evaluated on two benchmark datasets, MAL-XSEL outperformed conventional machine learning models, achieving top accuracies of 99.62% (ClaMP dataset) and 99.16% (MalwareDataSet). Notably, it surpassed state-of-the-art algorithms such as LightGBM (99.52%), random forest (99.33%), and decision trees (98.89%) across both datasets while maintaining computational efficiency. A unique interaction of ensemble learning and XAI is employed for detection, not only with improved accuracy but also with interpretable insight into the behavior of malware, thereby allowing trust to be substantiated in an automated system. By closing the divide between performance and interpretability, MAL-XSEL enables cybersecurity practitioners to deploy transparent and auditable defenses against an ever-growing resource of threats. This work demonstrates how there can be no compromise on explainability in security-critical applications and, as such, establishes a roadmap for future research on industrial malware analysis tools. Full article

The increasing prevalence of IoT technology in smart homes has significantly enhanced convenience but also introduced new security and safety challenges. Traditional security solutions, reliant on sequences of IoT-generated event data (e.g., notifications of device status changes and sensor readings), are vulnerable to cyberattacks, such as message forgery and interception and delaying attacks, and fail to monitor non-smart devices. Moreover, fragmented smart home ecosystems require vendor cooperation or system modifications for comprehensive monitoring, limiting the practicality of the existing approaches. To address these issues, we propose IoTBystander, a non-intrusive dual-channel smart home security monitoring framework that utilizes two ubiquitous platform-agnostic signals, i.e., audio and network, to monitor user and device activities. We introduce a novel dual-channel aggregation mechanism that integrates insights from both channels and cross-verifies the integrity of monitoring results. This approach expands the monitoring scope to include non-smart devices and provides richer context for anomaly detection, failure diagnosis, and configuration debugging. Empirical evaluations on a real-world testbed with nine smart and eleven non-smart devices demonstrate the high accuracy of IoTBystander in event recognition: 92.86% for recognizing events of smart devices, 95.09% for non-smart devices, and 94.27% for all devices. A case study on five anomaly scenarios further shows significant improvements in anomaly detection performance by combining the strengths of both channels. Full article

Taking the plant as a total uncertainty in a black box with measurable inputs and attainable outputs, this paper presents a constructive control design of agnostic nonlinear dynamic systems with discontinuous input (such as hard nonlinearities in the forms of dead zones, friction, and backlashes). This study expands the model-free sliding mode control (MFSMC), based on the Lyapunov differential inequality, to a total model-free robust control (TMFRC) for this class of piecewise systems, which does not use extra adaptive online data fitting modelling to deal with plant uncertainties and input discontinuities. The associated properties are analysed to justify the constraints and provide assurance for system stability analysis. Numerical examples in control of a non-affine nonlinear plant with three hard nonlinear inputs—a dead zone, Coulomb and viscous friction, and backlash—are used to test the feasibility of the TMFRC. Furthermore, real experimental tests on a permanent magnet synchronous motor (PMSM) are also given to showcase the control’s applicability and offer guidance for implementation. Full article

In resource-constrained mobile systems, efficiently handling incrementally added tasks under dynamically evolving requirements is a critical challenge. To address this, we propose aggregate pruning (AP), a framework that combines pruning with filter aggregation to optimize deep neural networks for continuous incremental multi-task learning (MTL). The approach reduces redundancy by dynamically pruning and aggregating similar filters across tasks, ensuring efficient use of computational resources while maintaining high task-specific performance. The aggregation strategy enables effective filter sharing across tasks, significantly reducing model complexity. Additionally, an adaptive mechanism is incorporated into AP to adjust filter sharing based on task similarity, further enhancing efficiency. Experiments on different backbone networks, including LeNet, VGG, ResNet, and so on, show that AP achieves substantial parameter reduction and computational savings with minimal accuracy loss, outperforming existing pruning methods and even surpassing non-pruning MTL techniques. The architecture-agnostic design of AP also enables potential extensions to complex architectures like graph neural networks (GNNs), offering a promising solution for incremental multi-task GNNs. Full article

Flooding can damage pavement infrastructure significantly, causing both immediate and long-term structural and functional issues. This research investigates how flooding events affect pavement deterioration, specifically focusing on measuring pavement roughness by the International Roughness Index (IRI). To quantify these effects, we utilized 20 years of pavement condition data from TxDOT’s PMIS database, which is integrated with flood event data, including duration and spatial extent. Statistical analyses were performed to compare IRI values before and after flooding and to calculate the deterioration rates influenced by flood exposure. Moreover, we applied explainable artificial intelligence (XAI) techniques, such as Shapley Additive Explanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME), to assess the impact of flooding on pavement performance. The results demonstrate that flood-affected pavements experience a more rapid increase in roughness compared to non-flooded sections. These findings emphasize the need for proactive flood mitigation strategies, including improved drainage systems, flood-resistant materials, and preventative maintenance, to enhance pavement resilience in vulnerable regions. Full article