Small and medium-sized enterprises in the automotive supply base face mounting pressure to adopt artificial intelligence, yet they are least able to absorb the consequences of misscoped adoption. Assessing this risk means reasoning under vagueness: Data maturity, workforce skill, financial capacity, and above all safety and regulatory readiness are graded, interdependent judgments that crisp, compensatory instruments discard. This paper proposes the fuzzy knowledge inference framework, a two-stage rough–fuzzy framework. Eight automotive-specific risk factors are represented as linguistic variables over an information system; rough-set analysis establishes their nonredundancy; and a hierarchical fuzzy inference process—anchored, compliance-weighted, and noncompensatory—yields an interpretable risk index. Using no primary human data, it is well calibrated at its anchors, monotone in favorability, and rank-convergent with two independent instruments, while escalating the profiles a compensatory model would understate. Its contribution to the literature on AI-adoption risk assessment for automotive SMEs is the noncompensatory treatment of safety and regulatory readiness.
Traditional software defect prediction (SDP) models mainly rely on software metrics or handcrafted structural features derived from dependency graphs. While network embedding techniques have recently been introduced, existing approaches often fail to fully capture higher-order and multi-hop dependencies among software components, limiting their ability to represent complex structural relationships within software systems. This study proposes a call graph-based embedding framework for software defect prediction that systematically captures structural dependencies among classes and integrates them with conventional software metrics. The goal is to improve predictive performance while preserving meaningful structural information that is often overlooked in traditional metric-based and graph-feature approaches. Ten real-world software projects were analyzed. For each project, call graphs were constructed and processed using six network embedding algorithms to generate 32-and 128-dimensional representations. A total of 550 models per project were developed using 25 dataset configurations and 22 classifiers. Performance was evaluated using Accuracy and AUC, with statistical significance assessed through non-parametric tests and effect size analysis. A cost-effectiveness analysis was conducted to determine practical deployment thresholds. The combination of call graph embeddings and software metrics consistently outperformed standalone approaches. The best-performing model (SVM with RBF kernel and GraRep embeddings) achieved a mean AUC of 70.73%, improving AUC by 5.2% over metric-only models and 12.8% over embedding-only models. Cost analysis indicates effectiveness for projects with lower defect densities. Call graph-based embeddings enhance SDP performance while providing actionable guidance for cost-effective adoption.
Software Fault Prediction (SFP) is the process of predicting fault-prone software constructs during the initial phases of software development. Deep Neural Networks (DNN) have been hugely successful in the field of computer vision, audio, etc., where the input data is correlated spatially and temporally. In contrast, SFP operates on tabular data, rows of software metrics that lack the inherent spatial structure exploited by convolutional architectures in image domains. The authors have tried to remodel the most successful Convolutional Neural Network (CNN) for tabular data. A novel framework is proposed employing a tree-based feature selection technique, LSelect, to find the most significant features and an adaptive 1-dimensional Convolutional Neural Network (ACNN) for the classification task, which selects an optimal learning rate automatically. ACNN converts the tabular data (1-D) into 2-D using adaptive pooling layers, thereby forming an image from 1-D data. The framework classification results (Area under Curve) are compared with nine state-of-the-art algorithms, such as XGBoost, LightGBM, etc., and performance is validated using the Bayesian Signed Rank Test. It is found that the proposed framework performs comparably with the state-of-the-art methods with reduced model complexity. Also, the LSelect feature selection technique improves average model performance by 1.3%.
Micro-expressions are brief, involuntary facial movements that reveal concealed emotions. While Convolutional Neural Networks (CNNs) are effective for recognizing such expressions, they are often computationally intensive and memory heavy. To address this, a Feature-Pruned and Weighted Convolutional Neural Network (FPW-CNN) is proposed to compress and accelerate performance. This model prunes filters and selects only the most relevant responses, reducing memory usage and computational load. Instead of stacking outputs, convolutional layers are flattened. Experiments on five spontaneous micro-expression datasets—SMIC, CASME, CASME II, CAS(ME)2, and SAMM—show that FPW-CNN achieves accuracies of 99.75%, 99.93%, 99.95%, 99.94%, and 99.90%, respectively. The model also reduces FLOPs by over 41% and maintains over 95% accuracy at 128×128 resolution, demonstrating robustness. These results confirm that FPW-CNN outperforms recent CNN and transformer models, making it suitable for real-time micro-expression recognition on resource-constrained devices.
The Internet of Things (IoT) with its pervasive connectivity imposes substantial challenges regarding privacy, security, and trust. Traditional security measures, such as cryptography and access control, often prove inadequate in the dynamic and resource-constrained IoT environment. The decentralized nature of IoT networks, coupled with the heterogeneity of devices, demands innovative security solutions. Trust is essential in human interaction and plays a crucial role in reducing challenges and security risks within IoT networks. This article focuses on trust-based methods for securing IoT systems, particularly within the context of Wireless Sensor Networks (WSNs). Trust-based approaches offer advantages such as lower computational complexity, leading to improved power management compared to traditional authentication and encryption methods. This article examines the latest trust models, as well as machine learning and deep learning algorithms specifically designed and implemented to enhance the security and reliability of trust-based WSN-assisted IoT applications.
