Abstract
The development of Artificial Intelligence (AI) models with an accurate prediction of human facial expression has become a significant challenge for the cases in which masks and sunglasses cover critical facial areas. Given that a substantial portion of human interactions involves non-verbal communication, accurately detecting human emotions such as anger, fear, disgust, happiness, sadness, and surprise would benefit a wide range of applications, from security assessments to psychological treatments. As a workaround, the current study explores the performance of a novel multi-channel arrangement comprised of a Haar-wavelet, Histogram of Oriented Gradients (HOG), and grayscale filters to improve the predictions of deep Convolutional Neural Network (CNN) on occluded results. This study uses the FER-2013 dataset and produces occluded samples by applying a virtual mask that covers almost 55% of facial areas comprising the mouth, lips, and jaw locations. Further investigations, including the impact of each filter, utilizing pre-trained models on occluded samples (transfer learning), and comparison to prior models are also carried out. The proposed approach yields an accuracy rate of 71% for non-occluded and 66% for the occluded samples, which are 6% to 11% higher than the base model. Further transfer learning technique increases the accuracy metrics by 18%, indicating that non-occluded pre-trained models can reveal a broader range of features and their relation, which to some extent compensates for the removed features due to the occlusion. These results suggest the potential capabilities of the proposed technique for similar imaging applications.

Abstract
Recent advancements in Software Defined Networks (SDN), Open Radio Access Network (O-RAN), and 5G technology have significantly expanded the capabilities of wireless networks, extending beyond mere data transmission. This progression has led to the emergence of Virtual Networks (VN) and Network Slicing, enabling industries to enhance their services and applications by establishing virtual networks that utilize shared physical infrastructure. Many works in the literature have considered optimizing the allocation of on-demand slices, assuming the absolute availability of resources and their accurate load. However, accurately allocating future network slices remains challenging due to the error in load prediction, diverse Key Performance Indicators (KPIs), resource price variations, and the potential for over-or under-provisioning. This study presents a two-phase intelligent approach to address these challenges. The framework proactively predicts different slice loads while considering prediction errors in optimizing future slices with varied KPIs in a cost-efficient manner. Specifically, our method utilizes historical load data per service and employs AI-based forecasts for service load prediction. Subsequently, it employs a Deep Reinforcement Learning (DRL) agent on O-RAN's virtual Control Unit (vCU) and virtual Distributed unit (vDU) to correct errors in prediction and optimize the cost of slice allocation based on service KPI requirements, ultimately pre-allocating future network slices at reduced costs. Through experimental validation against various baselines and state-of-the-art solutions, we demonstrate the efficacy of our proposed solution, achieving a notable reduction (37-51%) in the average cost of allocated slices while inquiring about (1.5-7%) of additional resources compared to the state-of-the-art..

Abstract
As we transition from centralized machine learning to distributed learning, new practices can significantly enhance intelligent Internet of Things (IoT) systems. This article introduces the concept of Opportunistic Distributed Learning (ODL), a general framework that enables any node in a network to initiates learning tasks by leveraging local, unused distributed resources collaboratively. ODL, facilitated by edge intelligence, promotes collective responsibility, pervasive and flexible distributed learning, allowing participating nodes to freely move, group, and regroup based on their conditions and benefits. The article discusses key research challenges of ODL in intelligent IoT systems, presents the ODL framework, proposes a reputation-based node selection scheme, and highlights the benefits and future research directions of the ODL system.

Abstract
The future of electric grids is undergoing a remarkable transformation driven by the increasing adoption of emerging technologies, notably Artificial Intelligence (AI) and Blockchain. These innovative technologies are revolutionizing smart grid management by introducing novel approaches that enhance efficiency, reliability, and sustainability, all while securing information across distributed grid components. AI empowers predictive analytics and real-time optimization, while Blockchain ensures secure and transparent transactions, laying the foundation for a more resilient and adaptive electrical grid system. This article introduces a novel Secure, Distributed, and Collaborative Learning (SDCL) framework for the smart grid. The SDCL framework leverages advances in distributed learning and blockchain technologies to provide scalability, secure data exchange, and rapid response capabilities. The proposed architecture not only enables secure data and model exchange among different microgrids but also facilitates the integration of multiple microgrids and distributed network operators. This integration enables the correlation of unforeseen events and enhances the management and control of emerging failures. Our resilient, blockchain-based architecture optimizes information sharing and security levels within the blockchain, accommodating diverse requirements for smart grid services. Finally, we highlight the advantages of the proposed SDCL framework and outline future research directions that warrant further investigation.

Abstract
The advancement of the use of Artificial Intelligence (AI) in the healthcare sector makes it possible to use computational intelligence applications to assist healthcare professionals in the diagnosis process, facilitating and optimizing early detection and allowing for a more accurate diagnosis (He et al., 2019). The application of machine learning methods, and, more recently, deep learning, has shown promising results (Barbosa et al., 2022; da Silva et al., 2021; De Oliveira et al., 2020; Espinola et al., 2021a, b; Gomes et al., 2021, 2023; Santana et al., 2018; Torcate et al., 2022). These approaches allow powerful tools to support diagnostic imaging and signs to be built, through the extraction of image features and the creation of a classification system, for example (Yu et al., 2018). There are several diseases known and classified by man, with different causes and prevalence. Therefore, contributing to the early detection of diseases defined as neglected was the initial motivation for this work.

Abstract
Background: We assessed the pertinence of yearly updating the International Working Group on the Diabetic Foot (IWGDF) risk classification in people with diabetes by quantifying the changes in the risk group and its accuracy in identifying those developing an ulcer (DFU) in a primary care setting. Methods: In our retrospective cohort study, we included all people with diabetes with a foot as-sessment registry between January 2016 and December 2018 in the Baixo Alentejo Local Health Unit. Foot-related data was collected at baseline after one and two years. DFU and/or death until December 2019 were registered. The proportion of people changing their risk status each year was calculated. Accuracy measures of the IWGDF classification to predict DFU occurrence at one, two, and three years were calculated. Results: A total of 2097 people were followed for three years, during which 0.1% died, and 12.4% developed a DFU. After two years, 3.6% of the participants had progressed to a higher-risk group. The IWGDF classification presented specificity values superior to 90% and negative predictive values superior to 99%. Conclusion: Foot risk status can be safely updated every two years instead of yearly. The IWGDF classification can accurately identify those not at risk of DFU.