With the increasing interconnectivity of devices, the Internet of Things (IoT) has revolutionized the industry and daily life. However, the proliferation of IoT devices has also increased security risks, necessitating robust protection mechanisms for sensitive data and critical infrastructure. The Advanced Encryption Standard (AES) remains the benchmark for securing IoT systems while balancing low power consumption, minimal area usage, and moderate throughput with high security. This paper offers a comprehensive review of the latest lightweight AES architectural designs, including optimizations to the Substitution Box (S-Box), Sub-Bytes, Shift Rows, Mix Columns, and Add Round Key steps, assessing their impact on gate count, area, maximum frequency, power consumption, and throughput in field programable gate arrays (FPGA) and Application-specific integrated circuit (ASIC) implementations. In addition, this study addresses vulnerabilities in lightweight AES cryptographic hardware to side-channel attacks (SCA), specifically focusing on Differential Fault Analysis (DFA). Furthermore, the analysis explored fault scenarios, rounds, and injection positions to evaluate fault severity. Countermeasures to DFA are reviewed, emphasizing fault detection methods, error detection levels, protection positions, and associated design overheads such as area, frequency, and throughput penalties, with special consideration for resource-constrained IoT devices. This study identifies critical gaps in lightweight AES and security challenges while discussing countermeasures that balance security with design efficiency. Finally, this study provides valuable insights for finding research directions to strengthen the robustness of AES in lightweight IoT environments.

Background/Objectives: Skin cancer is one of the most common diseases in the world, whose early and accurate detection can have a survival rate more than 90% while the chance of mortality is almost 80% in case of late diagnostics. Methods: A modified EfficientNet-B0 is developed based on mobile inverted bottleneck convolution with squeeze and excitation approach. The 3 × 3 convolutional layer is used to capture low-level visual features while the core features are extracted using a sequence of Mobile Inverted Bottleneck Convolution blocks having both 3 × 3 and 5 × 5 kernels. They not only balance fine-grained extraction with broader contextual representation but also increase the network’s learning capacity while maintaining computational cost. The proposed architecture hyperparameters and extracted feature vectors of standard benchmark datasets (HAM10000, ISIC 2019 and MSLD v2.0) of dermoscopic images are optimized with the quantum-behaved particle swarm optimization algorithm (QBPSO). The merit function is formulated by the training loss given in the form of standard classification cross-entropy with label smoothing, mean fitness value (mf^val), average accuracy (mAcc), mean computational time (mCT) and other standard performance indicators. Results: Comprehensive scenario-based simulations were performed using the proposed framework on a publicly available dataset and found an mAcc of 99.62% and 92.5%, mfval of 2.912 × 10⁻¹⁰ and 1.7921 × 10⁻⁸, mCT of 501.431 s and 752.421 s for HAM10000 and ISIC2019 datasets, respectively. The results are compared with state of the art, pre-trained existing models like EfficentNet-B4, RegNetY-320, ResNetXt-101, EfficentNetV2-M, VGG-16, Deep Lab V3 as well as reported techniques based on Mask RCCN, Deep Belief Net, Ensemble CNN, SCDNet and FixMatch-LS techniques having varying accuracies from 85% to 94.8%. The reliability of the proposed architecture and stability of QBPSO is examined through Monte Carlo simulation of 100 independent runs and their statistical soundings. Conclusions: The proposed framework reduces diagnostic errors and assists dermatologists in clinical decisions for an improved patient outcomes despite the challenges like data imbalance and interpretability.

Reading proficiency is a fundamental skill in higher education, yet many undergraduate students in Pakistan struggle with comprehension due to limited support and outdated instructional methods. The integration of Artificial Intelligence (AI) tools into education offers new possibilities for personalized and adaptive learning, but little research has examined their specific impact on reading development in Pakistani undergraduate settings. This study explores the perceptions and experiences of undergraduate students regarding the use of AI tools—such as ChatGPT, Rewordify, and ReadTheory—for enhancing academic reading proficiency. Using a qualitative research approach, semi-structured interviews were conducted with eight students from diverse disciplines who had prior exposure to AI tools. The thematic analysis framework was used to identify key themes across participant responses. Emerging themes include personalized and self-paced learning, vocabulary support and comprehension, increased engagement, and improved academic confidence. However, challenges such as over-reliance on AI-generated content, inconsistent usage habits, and digital literacy gaps were also noted. Findings suggest that AI tools, implemented with appropriate guidance, equitable access, and curriculum alignment, can significantly enhance students’ reading confidence, comprehension, and motivation. The study contributes to the growing body of literature on AI in education by offering context-specific insights for technology integration in literacy instruction. It highlights the need for inclusive, student-centered strategies to promote reading proficiency in resource-constrained academic environments. Future research should explore teacher perspectives, long-term learning outcomes, and the integration of AI tools across diverse linguistic and institutional contexts to further validate and expand on these findings.

Internet of Things (IoT) devices have found applications in numerous areas, including smart homes, health- care monitoring, smart metering, waste management, and industrial systems. These devices frequently collect and transmit data to fog nodes or cloud platforms for further processing. Ensuring the security of such data remains challenging, particularly when resource-constrained devices are involved. This paper presents a lightweight chaos-based image encryption scheme based on a two-stage confusion–diffusion structure. The proposed method employs the three-dimensional Intertwining Logistic Map (ILM) to gener- ate chaotic sequences that control the encryption operations. In the confusion stage, the input RGB image is decomposed into its red, green, and blue components, from which two color components are dynam- ically selected according to the generated chaotic sequences. Diagonal strips of varying positions and directions are then repeatedly exchanged between the selected components to disrupt both the spatial ar- rangement and cross-channel distribution of the image pixels. In the diffusion stage, a chaos-generated mask is combined with the scrambled image through bitwise XOR operations to modify the pixel intensities while maintaining computational simplicity. The proposed scheme is evaluated through a range of statis- tical and security analyses. Experimental results yield an average information entropy of 7.9975, close to the ideal value of 8 for an 8-bit image, while an average encryption time of 0.4269 seconds is obtained under the specified conventional computing environment. These results demonstrate the effectiveness and computational feasibility of the proposed approach and indicate its potential for resource-conscious image security applications. Validation on resource-constrained IoT and embedded platforms remains an impor- tant direction for future investigation.

Many sectors have seen significant improvements in operational efficiency and modern connection due to the increasing usage of industrial internet of things (IIoT) devices, a phenomenon known as IIoT. However, efficient intrusion detection systems (IDS) are essential for IIoT security because this quick growth also makes the network more susceptible to sophisticated cyberattacks. The high dimensionality and dynamic nature of IIoT data frequently cause traditional IDS algorithms to struggle, which results in insufficient threat detection and response. This paper uses a novel technique called Bayesian hyperparameter optimization with decision trees to enhance IDS performance in complex IIoT scenarios. Using the large IIoT-23 dataset, the method achieved an excellent 96.07% accuracy rate and a weighted average F1-score of 96.03%, in addition to demonstrating a significant increase in identifying a range of attack scenarios. In direction to growth the precision and robustness of our intrusion detection algorithms, this invention emphasizes how important it is to carefully modify complex hyperparameters and combine decision trees with ensemble methods like genetic algorithms and Bayesian optimization. The results show how successfully the approach adapts to budding threat landscape, expressively lessens false positives, and efficiently handles the enormous volumes of data that are a part of IIoT networks.