Search Results (1,089)

3D point cloud object tracking plays a critical role in fields such as autonomous driving and robotics, making the security of these models essential. Adversarial attacks are a key approach for studying the robustness and security of tracking models. However, research on the generalization of adversarial attacks for 3D point-cloud-tracking models is limited, and the frequency-domain information of the point cloud’s geometric structure is often overlooked. This frequency information is closely related to the generalization of 3D point-cloud-tracking models. To address these limitations, this paper proposes a novel adversarial method for 3D point cloud object tracking, utilizing frequency-domain attacks based on the importance of frequency bands. The attack operates in the frequency domain, targeting the low-frequency components of the point cloud within the search area. To make the attack more targeted, the paper introduces a frequency band importance saliency map, which reflects the significance of sub-frequency bands for tracking and uses this importance as attack weights to enhance the attack’s effectiveness. The proposed attack method was evaluated on mainstream 3D point-cloud-tracking models, and the adversarial examples generated from white-box attacks were transferred to other black-box tracking models. Experiments show that the proposed attack method reduces both the average success rate and precision of tracking, proving the effectiveness of the proposed adversarial attack. Furthermore, when the white-box adversarial samples were transferred to the black-box model, the tracking metrics also decreased, verifying the transferability of the attack method. Full article

The increasing prevalence of Arabic text in digital media offers significant potential for sentiment analysis. However, challenges such as linguistic complexity and limited resources make Arabic sentiment analysis (ASA) particularly difficult. In addition, explainable artificial intelligence (XAI) has become crucial for improving the transparency and trustworthiness of artificial intelligence (AI) models. This paper addresses the integration of XAI techniques in ASA through a scoping review of developments. This study critically identifies trends in model usage, examines explainability methods, and explores how these techniques enhance the explainability of model decisions. This review is crucial for consolidating fragmented efforts, identifying key methodological trends, and guiding future research in this emerging area. Online databases (IEEE Xplore, ACM Digital Library, Scopus, Web of Science, ScienceDirect, and Google Scholar) were searched to identify papers published between 1 January 2016 and 31 March 2025. The last search across all databases was conducted on 1 April 2025. From these, 19 peer-reviewed journal articles and conference papers focusing on ASA with explicit use of XAI techniques were selected for inclusion. This time frame was chosen to capture the most recent decade of research, reflecting advances in deep learning and the transformer-based and explainable AI methods. The findings indicate that transformer-based models and deep learning approaches dominate in ASA, achieving high accuracy, and that local interpretable model-agnostic explanations (LIME) is the most widely used explainability tool. However, challenges such as dialectal variation, small or imbalanced datasets, and the black box nature of advanced models persist. To address these challenges future research directions should include the creation of richer Arabic sentiment datasets, the development of hybrid explainability models, and the enhancement of adversarial robustness. Full article

In today’s digital era, cyberattacks are rapidly evolving, rendering traditional security mechanisms increasingly inadequate. The adoption of AI-based Network Intrusion Detection Systems (NIDS) has emerged as a promising solution, due to their ability to detect and respond to malicious activity using machine learning techniques. However, these systems remain vulnerable to adversarial threats, particularly data poisoning attacks, in which attackers manipulate training data to degrade model performance. In this work, we examine tree classifiers, Random Forest and Gradient Boosting, to model black box poisoning attacks. We introduce FortiNIDS, a robust framework that employs a surrogate neural network to generate adversarial perturbations that can transfer between models, leveraging the transferability of adversarial examples. In addition, we investigate defense strategies designed to improve the resilience of NIDS in smart city Internet of Things (IoT) settings. Specifically, we evaluate adversarial training and the Reject on Negative Impact (RONI) technique using the widely adopted CICDDoS2019 dataset. Our findings highlight the effectiveness of targeted defenses in improving detection accuracy and maintaining system reliability under adversarial conditions, thereby contributing to the security and privacy of smart city networks. Full article

The Shizhuyuan polymetallic deposit in Hunan Province, China, is a world-class ore field rich in tungsten (W), tin (Sn), molybdenum (Mo), and bismuth (Bi), now facing resource depletion due to prolonged exploitation. This study addresses the limitations of traditional geological prediction methods in complex terrain by integrating multi-source datasets—including γ-ray spectrometry, high-precision magnetometry, induced polarization (IP), and soil radon measurements—across 5049 samples. Unsupervised factor analysis was employed to extract five key ore-indicating factors, explaining 82.78% of data variance. Based on these geological features, predictive models including Support Vector Machine (SVM), Random Forest (RF), and Extreme Gradient Boosting (XGBoost) were constructed and compared. SHAP values were employed to quantify the contribution of each geological feature to the prediction outcomes, thereby transforming the machine learning “black-box models” into an interpretable geological decision-making basis. The results demonstrate that machine learning, particularly when integrated with multi-source data, provides a powerful and interpretable approach for deep mineral prospectivity mapping in concealed terrains. Full article

Background/Objectives: The integration of artificial intelligence (AI) into obstetric care poses significant potential to enhance clinical decision-making and optimize maternal and neonatal outcomes. Traditional prediction methods in maternal-foetal medicine often rely on subjective clinical judgment and limited statistical models, which may not fully capture complex patient data. By integrating computational innovation with mechanistic biology and rigorous clinical validation, AI can finally fulfil the promise of precision obstetrics by transforming pregnancy complications into a preventable, personalised continuum of care. This study aims to map the current landscape of AI applications across the continuous spectrum of maternal–foetal health, identify the types of models used, and compare clinical targets and performance, potential pitfalls, and strategies to translate innovation into clinical impact. Methods: A literature search of peer-reviewed studies that employ AI for prediction, diagnosis, or decision support in Obstetrics was conducted. AI algorithms were categorised by application area: foetal monitoring, prediction of preterm birth, prediction of pregnancy complications, and/or labour and delivery. Results: AI-driven models consistently demonstrate superior performance to traditional approaches. Nevertheless, their widespread clinical adoption is hindered by limited dataset diversity, “black-box” algorithms, and inconsistent reporting standards. Conclusions: AI holds transformative potential to improve maternal and neonatal outcomes through earlier diagnosis, personalised risk assessment, and automated monitoring. To fulfil this promise, the field must prioritize the creation of large, diverse, open-access datasets, mandate transparent, explainable model architectures, and establish robust ethical and regulatory frameworks. By addressing these challenges, AI can become an integral, equitable, and trustworthy component of Obstetric care worldwide. Full article

Accurate prediction of soil–structure interface shear strength (τ_max) is critical for reliable geotechnical design. This study combines experimental testing with interpretable machine learning to overcome the limitations of traditional empirical models and black-box approaches. Ninety large-displacement ring shear tests were performed on five sands and three interface materials (steel, PVC, and stone) under normal stresses of 25–100 kPa. The results showed that particle morphology, quantified by the regularity index (RI), and surface roughness (R_t) are dominant factors. Irregular grains and rougher interfaces mobilised higher τ_max through enhanced interlocking, while smoother particles reduced this benefit. Harder surfaces resisted asperity crushing and maintained higher shear strength, whereas softer materials such as PVC showed localised deformation and lower resistance. These experimental findings formed the basis for a hybrid symbolic regression framework integrating Genetic Programming (GP) with Shapley Additive Explanations (SHAP), Fourier feature augmentation, and physics-informed constraints. Compared with multiple linear regression and other hybrid GP variants, the Physics-Informed Neural Fourier GP (PIN-FGP) model achieved the best performance (R² = 0.9866, RMSE = 2.0 kPa). The outcome is a set of five interpretable and physics-consistent formulas linking measurable soil and interface properties to τ_max. The study provides both new experimental insights and transparent predictive tools, supporting safer and more defensible geotechnical design and analysis. Full article

Accurate quantification of minor mineral phases is important in Powder X-Ray Diffraction (PXRD) and Rietveld phase quantification. The precise limit of quantification for the various phases is rarely considered but rather approximated to 0.2–2 wt% by applying a global minimum weight percentage threshold. This approximation often leads to false positive or false negative phase quantity, jeopardizing the trustworthiness of the analytic method in general. In this work (1) we propose a dynamic and adaptable false positive filtering method for Rietveld Quantitative X-ray diffraction (QXRD) based on a phase-specific signal-to-noise ratio referred to as “Phase-SNR”; (2) we introduce the method baptized “Phase Guard” which is implemented in the software HighScore Plus. Phase Guard is based on peaks counting statistics and it automatically adapts to different mineral scattering powers, different mineral crystallinity, instrumental configuration and measurement time. Its applicability and benefits are demonstrated with several examples in cement and mining applications. The adoption of Phase Guard is especially beneficial for industrial black-box solutions, where all “probable” phases are included in the model, even when they are absent from the sample. Phase Guard eliminates false positives, it reduces the likelihood of false negatives, and it is an essential tool to answer the question “what is the limit of quantification for Rietveld analysis?” Full article

At present, artificial intelligence (AI) has shown significant potential in digestive endoscopy image analysis, serving as a powerful auxiliary tool for the accurate diagnosis and treatment of gastrointestinal diseases. However, mainstream models represented by deep learning are often characterized as complex “black boxes,” with decision-making processes that are difficult for humans to interpret. The lack of interpretability undermines physicians’ trust in model results and hinders the broader use of models in clinical practice. To address this core challenge, Explainable AI (XAI) has emerged to enhance the transparency of decision-making, thereby establishing a foundation of trust for human–machine collaboration. The review systematically reviews 34 articles (7 articles in esophagogastroduodenoscopy, 13 articles in colonoscopy, 9 articles in endoscopic ultrasonography, and 5 articles in wireless capsule endoscopy), focusing on the research progress and applications of XAI in the field of digestive endoscopic image analysis, with particular emphasis on the visual explanation-based methods. We first clarify the definition and mainstream classification of XAI, then introduce the principles and characteristics of key XAI methods based on visual explanation. Subsequently, we review the applications of these methods in digestive endoscopy image analysis. Lastly, we explore the obstacles presently faced in this domain and the future directions. This study provides a theoretical basis for constructing a trustworthy and transparent AI-assisted digestive endoscopy diagnosis and treatment system and promotes the implementation and application of XAI in clinical practice. Full article

Data-driven modeling methods have been preliminarily applied in the development of tight-gas reservoirs, demonstrating unique advantages in post-fracturing productivity prediction. However, most of the established predictive models are “black-box” models, which provide productivity predictions based on a set of input parameters without revealing the internal prediction mechanisms. This lack of transparency reduces the credibility and practical utility of such models. To address the challenges of poor performance and low trustworthiness of “black-box” machine learning models, this study explores a data-driven approach to “black-box” predictive modeling by integrating ensemble learning with interpretability methods. The results indicate the following: The post-fracturing productivity prediction model for tight-gas reservoirs developed in this study, based on ensemble learning, achieves a goodness of fit of 0.923, representing a 26.09% improvement compared to the best-performing individual machine learning model. The stacking ensemble model predicts post-fracturing productivity for horizontal wells more accurately and effectively mitigates the prediction biases of individual machine learning models. An interpretability method for the “black-box” ensemble learning-based productivity prediction model was established, revealing the ranked importance of factors influencing post-fracturing productivity: reservoir properties, controllable operational parameters, and rock mechanics. This ranking aligns with the results of orthogonal experiments from mechanism-driven numerical models, providing mutual validation and enhancing the credibility of the ensemble learning-based productivity prediction model. In conclusion, this study integrates mechanistic numerical models and data-driven models to explore the influence of various factors on post-fracturing productivity. The cross-validation of results from both approaches underscores the reliability of the findings, offering theoretical and methodological support for the design of fracturing schemes and the iterative advancement of fracturing technologies in tight-gas reservoirs. Full article

The formation of strategies and tactics of destructive impact (DI) at the stages of complex computer attacks (CCAs) largely depends on the content of intelligence data obtained by the intruder about the attacked elements of distributed information systems (DISs). This study analyzes scientific papers, methodologies and standards in the field of assessing the indicators of awareness of the intruder about the objects of DI and symmetrical indicators of intelligence security of the elements of the DIS. It was revealed that the aspects of changing the quantitative and qualitative characteristics of intelligence data (ID) at the stages of CCA, as well as their impact on the possibilities of using certain types of simple computer attacks (SKAs), are poorly studied and insufficiently systematized. This paper uses technologies for modeling the process of an intruder obtaining ID based on the application of the methodology of black, grey and white boxes and the theory of fuzzy sets. This allowed us to identify the relationship between certain arrays of ID and the possibilities of applying certain types of SCA end-structure arrays of ID according to the levels of identifying objects of DI, and to create a scale of intruder awareness symmetrical to the scale of intelligence protection of the elements of the DIS. Experiments were conducted to verify the practical applicability of the developed models and techniques, showing positive results that make it possible to identify vulnerable objects, tactics and techniques of the intruder in advance. The result of this study is the development of an intruder awareness scale, which includes five levels of his knowledge about the attacked system, estimated by numerical intervals and characterized by linguistic terms. Each awareness level corresponds to one CCA stage: primary ID collection, penetration and legalization, privilege escalation, distribution and DI. Awareness levels have corresponding typical ID lists that can be potentially available after conducting the corresponding type of SCA. Typical ID lists are classified according to the following DI levels: network, hardware, system, application and user level. For each awareness level, the method of obtaining the ID by the intruder is specified. These research results represent a scientific contribution. The practical contribution is the application of the developed scale for information security (IS) incident management. It allows for a proactive assessment of DIS security against CCAs—modeling the real DIS structure and various CCA scenarios. During an incident, upon detection of a certain CCA stage, it allows for identifying data on DIS elements potentially known by the intruder and eliminating further development of the incident. The results of this study can also be used for training IS specialists in network security, risk assessment and IS incident management. Full article

As 5G and beyond networks grow in heterogeneity, complexity, and scale, traditional Intrusion Detection Systems (IDS) struggle to maintain accurate and precise detection mechanisms. A promising alternative approach to this problem has involved the use of Deep Learning (DL) techniques; however, DL-based IDS suffer from issues relating to interpretation, performance variability, and high computational overheads. These issues limit their practical deployment in real-world applications. In this study, CiNeT is introduced as a novel DL-based IDS employing Convolutional Neural Networks (CNN) within a bijective encoding–decoding framework between network traffic features (such as IPv6, IPv4, Timestamp, MAC addresses, and network data) and their RGB representations. This transformation facilitates our DL IDS in detecting spatial patterns without sacrificing fidelity. The bijective pipeline enables complete traceability from detection decisions to their corresponding network traffic features, enabling a significant initiative towards solving the ‘black-box’ problem inherent in Deep Learning models, thus facilitating digital forensics. Finally, the DL IDS has been evaluated on three datasets, UNSW NB-15, InSDN, and ToN_IoT, with analysis conducted on accuracy, GPU usage, memory utilisation, training, testing, and validation time. To summarise, this study presents a new CNN-based IDS with an end-to-end pipeline between network traffic data and their RGB representation, which offers high performance and enhanced interpretability through revisable transformation. Full article

Accurately identifying the load status of a ball mill is critical for optimizing grinding efficiency and ensuring operational stability. However, the one-dimensional vibration signals collected from ball mills exhibit strong non-stationarity and a high degree of entanglement between multi-scale local transient features and long-range temporal evolution patterns. To address this, rather than relying on a purely black-box approach, this paper introduces a novel Deep Multi-scale Spatial–Temporal Feature Decoupling Network (DMSTFD-Net) guided by a clear feature decoupling philosophy to enhance model interpretability. The core of DMSTFD-Net lies in its hierarchical collaborative feature refinement mechanism. It first utilizes a one-dimensional residual network (ResNet) to adaptively capture and preliminarily decouple multi-scale spatial characteristics from the raw signal. Subsequently, the extracted high-level feature sequences are fed into a bidirectional gated recurrent unit (Bi-GRU) to decouple high-order temporal dynamic patterns. Experiments on a multi-condition dataset demonstrate that the proposed network achieves a state-of-the-art accuracy of 97.65%. Furthermore, dedicated cross-condition experiments and t-SNE visualizations validate the framework’s effectiveness. The results confirm that DMSTFD-Net provides a powerful, robust, and more interpretable solution for ball mill load identification. Full article

The widespread adoption of deep learning (DL) models in neuroscience research has introduced a fundamental epistemological paradox: while these models demonstrate remarkable performance in pattern recognition and prediction tasks, their inherent opacity contradicts neuroscience’s foundational goal of understanding biological mechanisms. This review article examines the growing trend of using DL models to interpret neural dynamics and extract insights about brain function, arguing that the black box nature of these models fundamentally undermines their utility for mechanistic understanding. We explore the distinction between computational performance and scientific explanation, analyze the limitations of current interpretability techniques, and discuss the implications for neuroscience research methodology. We propose that the field must critically evaluate whether DL models can genuinely contribute to our understanding of neural processes or whether they merely provide sophisticated curve-fitting tools that obscure rather than illuminate the underlying biology. Full article

Background: Facing global wheat production pressures such as environmental degradation and reduced cultivated land, breeding innovation is urgent to boost yields. Genomic selection (GS) is a useful wheat breeding technology to make the breeding process more efficient, increasing the genetic gain per unit time and cost. Precise genomic estimated breeding value (GEBV) via genome-wide markers is usually hampered by high-dimensional genomic data. Methods: To address this, we propose GRE, a framework combining genome-wide association study (GWAS)’s biological significance and random forest (RF)’s prediction efficiency for an explainable machine learning GS model. First, GRE identifies significant SNPs affecting wheat yield traits by comparison of the constructed 24 SNP subsets (intersection/union) selected by leveraging GWAS and RF, to analyze the marker scale’s impact. Furthermore, GRE compares six GS algorithms (GBLUP and five machine learning models), evaluating performance via prediction accuracy (Pearson correlation coefficient, PCC) and error. Additionally, GRE leverages Shapley additive explanations (SHAP) explainable techniques to overcome traditional GS models’ “black box” limitation, enabling cross-scale quantitative analysis and revealing how significant SNPs affect yield traits. Results: Results show that XGBoost and ElasticNet perform best in the union (383 SNPs) of GWAS and RF’s TOP 200 SNPs, with high accuracy (PCC > 0.864) and stability (standard deviation, SD < 0.005), and the significant SNPs identified by XGBoost are precisely explained by their main and interaction effects on wheat yield by SHAP. Conclusions: This study provides tool support for intelligent breeding chip design, important trait gene mining, and GS technology field transformation, aiding global agricultural sustainable productivity. Full article

Groundwater level (GWL) prediction in arid regions faces two fundamental challenges in conventional numerical modeling: (i) irreducible parameter uncertainty, which systematically reduces predictive accuracy; (ii) oversimplification of nonlinear process interactions, which leads to error propagation. Although machine learning (ML) methods demonstrate strong nonlinear mapping capabilities, their standalone applications often encounter prediction bias and face the accuracy–generalization trade-off. This study proposes a hybrid TCN–Transformer–LSTM (TTL) model designed to address three key challenges in groundwater prediction: high-frequency fluctuations, medium-range dependencies, and long-term memory effects. The TTL framework integrates TCN layers for short-term features, Transformer blocks to model cross-temporal dependencies, and LSTM to preserve long-term memory, with residual connections facilitating hierarchical feature fusion. The results indicate that (1) at the monthly scale, TTL reduced RMSE by 20.7% (p < 0.01) and increased R² by 0.15 compared with the Groundwater Modeling System (GMS); (2) during abrupt hydrological events, TTL achieved superior performance (R² = 0.96–0.98, MAE < 0.6 m); (3) PCA revealed site-specific responses, corroborating the adaptability and interpretability of TTL; (4) Grad-CAM analysis demonstrated that the model captures physically interpretable attention mechanisms—particularly evapotranspiration and rainfall—thereby providing clear cause–effect explanations and enhancing transparency beyond black-box models. This transferable framework supports groundwater forecasting, risk warning, and practical deployment in arid regions, thereby contributing to sustainable water resource management. Full article