An enormous number of the Internet of Things (IoT) applications and their networks have significantly impacted people’s lives in diverse situations. With the increasing adoption of these applications in various sectors, ensuring reliability and security has become a critical concern. Moreover, the network that interconnected IoT devices uses advanced communications norms and technologies to capture and transmit data. Still, these networks are subject to various types of attacks that will lead to the loss of user data. Concurrently, the field of anomaly detection for the Internet of Things (IoT) is experiencing rapid expansion. This expansion requires a thorough analysis of application trends and existing gaps. Furthermore, it is critical in detecting interesting phenomena such as device damage and unknown events. However, this task is tough due to the unpredictable nature of anomalies and the complexity of the environment. This paper offers a technique that uses an autoencoder neural network to identify anomalous network communications in IoT networks. More specifically, we propose and implement a model that uses DAE (deep autoencoder) to detect and classify the network data, with an ANOVA F-Test for the feature selection. The proposed model is validated using the NSL-KDD dataset. Compared to some IoT-based anomaly detection models, the experimental results reveal that the suggested model is more efficient at enhancing the accuracy of detecting malicious data. The simulation results show that it works better, with an overall accuracy rate of 85% and 92% successively for the binary and multi-class classifications.

Wireless body area networks (WBANs) are a pivotal solution for continuous health monitoring, but their energy constraints pose a significant challenge for long-term operation. This paper provides a comprehensive review of state-of-the-art energy-efficient mechanisms, critically evaluating solutions across various network layers. We focus on three key approaches: energy-aware MAC protocols that reduce idle listening and optimize duty cycling; energy-efficient routing protocols that enhance data transmission and network longevity; and emerging energy harvesting techniques that offer a path toward energy-autonomous WBANs. Furthermore, the paper provides a detailed analysis of the inherent trade-offs between energy efficiency and other critical performance metrics, such as latency, reliability, and security. It also explores the transformative potential of emerging technologies, such as AI and blockchain, for dynamic energy management and secure data handling. By synthesizing these findings, this work contributes to the development of sustainable WBAN solutions and outlines clear directions for future research.

This research investigates the multifaceted determinants of academic performance in a large-scale study of 5,092 undergraduate students across diverse Moroccan universities. We delved into demographic, academic, psychological, societal, and lifestyle factors, providing a comprehensive analysis of potential influencers on academic success. These dimensions were identified as significant predictors of academic performance, as gauged by the students' Grade Point Average (GPA). The findings reveal the intricate and interconnected nature of these factors and their cumulative influence on a student's academic trajectory. Our study emphasizes the necessity for a holistic approach that integrates not only traditional academic strategies, but also psychological support, balanced lifestyle practices, and effective societal engagement. Our research underscores the rich interplay among these variables and offers a profound understanding of their collective impact on academic performance. The derived insights could be instrumental for educators, policymakers, and researchers striving to create strategies that amplify student success. While acknowledging the inherent limitations of the study, we believe the results serve as a pivotal guidepost in the ongoing quest to optimize student academic outcomes and overall well-being.

Deep learning has emerged as an effective tool for automatic waste classification, supporting cleaner cities and more sustainable recycling systems. Because environmental protection is central to the United Nations Sustainable Development Goals (SDGs), improving the sorting and processing of everyday waste is a practical step toward this broader objective. In many real-world settings, however, waste is still sorted manually, which is slow, labor-intensive, and prone to human error. Although convolutional neural networks (CNNs) can automate this task with high accuracy, many state-of-the-art models remain too large and computationally demanding for low-cost edge devices intended for deployment in homes, schools, and small recycling facilities. In this work, we investigate lightweight waste-classification models suitable for TinyML deployment while preserving competitive accuracy. We first benchmark multiple CNN architectures to establish a strong baseline, then apply complementary compression strategies including quantization, pruning, singular value decomposition (SVD) low-rank approximation, and knowledge distillation. In addition, we evaluate an RL-guided multi-teacher selection benchmark that adaptively chooses one teacher per minibatch during distillation to improve student training stability, achieving up to 85% accuracy with only 0.496 M parameters (FP32 ≈ 1.89 MB; INT8 ≈ 0.47 MB). Across all experiments, the best accuracy–size trade-off is obtained by combining knowledge distillation with post-training quantization, reducing the model footprint from approximately 16 MB to 281 KB while maintaining 82% accuracy. The resulting model is feasible for deployment on mobile applications and resource-constrained embedded devices based on model size and TensorFlow Lite Micro compatibility.