Net Zero Energy Buildings (NZEBs) are an important milestone in the realm of sustainable development, seeking a point of balance between the demand for energy and the production of renewable energy. The integration of Artificial Intelligence (AI) and Machine Learning (ML) presents a vast opportunity for the improvement of the performance and efficiency levels of NZEBs. The paper provides an insight into contemporary developments in the realm of AI and ML, as they are applicable to the context of NZEBs and other aspects such as management of energy, the integration of renewable energy and energy storage, predictions and faults, behavior analysis of occupants, and a building’s thermal comfort and simulations for building designs. Additionally, the paper identifies the symbiotic relationship unfolding between the technologies of AI and other technologies such as the Internet of Things (IoT), Digital Twin, among other technologies associated with it. The discussions evidence the relevance of facts associated with the management of data, security, and ethical aspects associated with the information they convey. Also, the paper, through the description of certain key developments in the technologies associated with implementation challenges, may identify the central role presented by the importance of the technologies of AI and ML within the context of the design of the NZEBs, based on the research pathway recommendations.

The Internet of Underwater Things (IoUT) supports marine sensing, environmental monitoring, subsea inspection, and autonomous underwater operations. However, IoUT communication is constrained by limited bandwidth, long propagation delay, time-varying underwater channels, intermittent connectivity, and strict energy budgets. Semantic Communication (SC) offers a promising alternative by transmitting task-relevant meaning rather than raw data, thereby improving communication efficiency in resource-constrained underwater networks. This paper presents a critical and feasibility-aware survey of SC for IoUT, focusing on opportunities, challenges, limitations, and future research directions. We first review the fundamentals of SC-enabled IoUT systems, including semantic representations, layered architectures, semantic channel modeling, and task-oriented evaluation metrics. We then examine learning-driven approaches based on machine learning (ML), knowledge graphs (KGs), vision-language models (VLMs), generative models, and federated learning (FL), with emphasis on their feasibility under underwater edge constraints. Representative applications, including environmental monitoring, marine ecology, subsea infrastructure inspection, disaster response, and autonomous underwater vehicle (AUV) coordination, are analyzed from an SC perspective. Finally, we identify key research directions involving standardized semantic models, reproducible testbeds, compute–communication trade-offs, trustworthy reconstruction, hybrid underwater links, energy-aware edge intelligence, semantic security, digital twins (DTs), and cross-domain interoperability. This survey provides a structured foundation for developing reliable, efficient, and meaning-driven IoUT communication systems.

Campus innovation can ensure significant advancements in the computer vision industry by implementing a smart attendance system (SAS). The Internet of Things (IoT) is being utilized in conjunction with the deep learning technique for facial identification, using convolutional neural networks (CNNs), to automatically detect faces and track attendance with high accuracy and precision. In order to develop a real-time program that deals with the rote activities of controlling the attendance system in a facility, this research work has focused on detecting a single image. The procedure entails identifying faces using security camera footage captured at various points across the campus, as well as from other system-related information technologies. Using the proposed hierarchal multi-task cascaded convolutional networks (MTCNNs) on small datasets and, in particular, deep learning recognition functions, the experimental results demonstrate that the Yale Database is one of the best datasets for tackling practice tasks in face recognition.

Social interaction is a fundamental part of daily life, but for individuals with autism, it can be difficult and stressful. This paper introduces EchoMind, a virtual reality (VR) training system that helps users build social communication skills through immersive, voice-based role-play. EchoMind combines speech recognition, generative AI, and real-time dialogue with a virtual agent named Allen, who is designed to be patient, supportive, and emotionally supportive. The system provides structured prompts developed by educators, covering scenarios such as school conversations and workplace interactions. By enabling natural, guided conversations in a safe environment, EchoMind aims to enhance user confidence, alleviate anxiety, and foster long-term social development.

This research brings one of the significant amendments in Multi-model Predictive Control (MMPC) for controlling the attitude of an underactuated quadrotor unmanned aerial vehicle (UAV). It integrates the high-tech performance of non-linear model predictive control (NMPC) along with the lightweight computational efficiency of Linear Model Predictive Control (LMPC). Although NMPC produces high performance, its high computational cost makes it unfeasible for real-time applications, such as an attitude controller. LMPC is a good way of achieving this and gives reasonable performance in practice, however, it does not work well with nonlinear dynamics and has stability and precision issues. The MMPC methodology in this work addresses these issues using linear quadrotor attitude dynamics models, particularized by principal component analysis (PCA)-based model reduction resulting in fewer required LMPCs. To reduce the “chattering effect” commonly seen in multi-model approaches, we present an adaptive gain scheduling approach that improves the smoothness of control switching which helps improve stability and reduce actuator wear. This proposed MMPC achieves NMPC-like performance at computational costs similar to LMPC, outperforming other control techniques such as incremental nonlinear dynamic inversion, sliding mode control, and traditional LMPC and NMPC.

In the world, Alzheimer’s disease (AD) is the utmost public reason for dementia. AD causes memory loss and disturbing mental function impairment in aging people. The loss of memory and disturbing mental function brings a significant load on patients as well as on society. So far, there is no actual treatment that can cure AD; however, early diagnosis can slow down this disease. Deep learning has shown substantial success in diagnosing AZ disease. However, challenges remain due to limited data, improper model selection, and extraction of irrelevant features. In this work, we proposed a fully automated framework based on the fusion of a vision transformer and a novel inverted residual bottleneck with self-attention (IRBwSA) for AD diagnosis. In the first step, data augmentation was performed to balance the selected dataset. After that, the vision model is designed and modified according to the dataset. Similarly, a new inverted bottleneck self-attention model is developed. The designed models are trained on the augmented dataset, and extracted features are fused using a novel search-based approach. Moreover, the designed models are interpreted using an explainable artificial intelligence technique named LIME. The fused features are finally classified using a shallow wide neural network and other classifiers. The experimental process was conducted on an augmented MRI dataset, and 96.1% accuracy and 96.05% precision rate were obtained. Comparison with a few recent techniques shows the proposed framework’s better performance.

Abstract A serious, all-encompassing, and deadly cancer that affects every part of the body is skin cancer. The most prevalent causes of skin lesions are UV radiation, which can damage human skin, and moles. If skin cancer is discovered early, it may be adequately treated. In order to diagnose skin lesions with less effort, dermatologists are increasingly turning to machine learning (ML) techniques and computer-aided diagnostic (CAD) systems. This paper proposes a computerized method for multiclass lesion classification using a fusion of optimal deep-learning model features. The dataset used in this work, ISIC2018, is imbalanced; therefore, augmentation is performed based on a few mathematical operations. After that, two pre-trained deep learning models (DarkNet-19 and MobileNet-V2) have been fine-tuned and trained on the selected dataset. After training, features are extracted from the average pool layer and optimized using a hybrid firefly optimization technique. The selected features are fused in two ways: (i) original serial approach and (ii) proposed threshold approach. Machine learning classifiers are used to classify the chosen features at the end. Using the ISIC2018 dataset, the experimental procedure produced an accuracy of 89.0%. Whereas, 87.34, 87.57, and 87.45 are sensitivity, precision, and F1 score respectively. At the end, comparison is also conducted with recent techniques, and it shows the proposed method shows improved accuracy along with other performance measures.

The rapid spread of monkeypox (mpox) across several nations has made the current outbreak a serious public health concern. Early detection and diagnosis are essential for mpox to be effectively treated and managed. In this paper, we propose a novel deep residual self-attention architecture for Mpox classification from dermoscopic images. The proposed model is based on the number of residual blocks and attention modules added in two fashions. Two residual inverted bottleneck blocks are added in the first phase, and weights are computed spatially. In the second phase, squeeze the refined residual blocks through a self-attention module and fuse their information into a single-weight matrix. The resultant weight matrix is finally passed to the fully connected layer. During the training phase, hyperparameters are optimized using Bayesian Optimization (BO) to facilitate smooth learning of the model. In the testing phase, features are extracted from the depth concatenation layer and passed to the shallow neural network classifier for the final classification. The proposed model is evaluated on two publicly available datasets and obtained $\mathbf{9 5. 7 0 \%}$ and $\mathbf{8 5. 9 0 \%}$ accuracy. A comparison with existing techniques shows that the proposed model outperforms the SOTA techniques.

Mobile Edge Computing environments, become very important topic in the field of networking and telecommunication by providing low latency service with their structure and resources. Distributed Denial of Service attacks are particularly vulnerable which can severely Freezing the services, affect the infrastructure, and slowing down network performance. These threats are often fail to addressed through traditional security frameworks. This research describes different categories of solutions aimed at improving resilience in MEC networks while considering their limits and constraints. The research assesses the performance of thirteen models, including machine learning, deep learning, unsupervised, hybrid, and transformer-based approaches for DDoS attack detection. In the realm of ML, Random Forest takes the crown with Accuracy (99.92%) to Boosting and Decision Tree with (99.85%) and (99.86%) respectively. CNN-LSTM in the hybrid deep learning also performed well with 99.69% accuracy, while BiLSTM trailed at 98.74%. K-Means and DBSCAN also proved to be helpful with 94.21% and 95.95% accuracy, K-Means being the least accurate among the unsupervised methods. These models can help reduce latency in Mobile Edge Computing by detecting DDoS attacks quickly and directly at the edge of the network.

Due to the interconnected nature of IoT devices, the systems are more prone to vulnerabilities and security attacks. An intruder may attack the devices, hosts, servers, and gateways in a network. Several security solutions are available to protect IoT networks such as intelligent Intrusion Detection System (IDS) which works on network traffic to train themselves to recognize any intrusion using Machine and Deep Learning (ML/DL) techniques. However, the DL techniques are most suited to train themselves on huge data size of IoT network and requires considerable training time. This study presents a DL technique for malicious attack detection in IoT networks with reduced training time. The proposed model has a GRU layer which is followed by two dense layers. The proposed model is evaluated for multi-class classification using different optimizers such as ADAM, and ADAMAX with batch sizes 32, 64, and 128. The model has achieved up to 99% precision, recall, F1-score, and accuracy with the reduction in training time using MQTT-IoT-IDS2020 dataset.

Remote sensing (RS) images are evolving daily for their applications in surveillance, planned urbanization, law enforcement, climate change detection, agriculture, and monitoring catastrophes. Artificial intelligence techniques in this application heavily depend on the quality of RS images. The low-quality RS images, such as those based on deep learning architectures trained on the data, impact the AI model’s performance. Fusing features and models increases the pattern redundancy, sometimes affecting the model performance and leading to overfitting. This article proposes a novel batch normalization deep bottleneck residual architecture to classify aerial scene recognition from low-resolution RS images accurately. The proposed architecture is based on the seven-batch-normalized residual blocks that accurately extract deeper information from the raw data. These blocks extract contextual and spatial characteristics from satellite image data. In the training of the proposed model, hyperparameters are selected through Bayesian optimization, which makes it smoother than manual selection. Finally, the GradCAM visualization technique tests and interprets the trained model. Four datasets are employed in this work for the experimental process, such as EuroSAT, NWPU-RESISC45, and SIRI-WHU. A customized CoastalD dataset is also prepared to validate the proposed model for predicting coastal regions. The accuracy of these datasets is 94.62%, 92.88%, 99.75%, and 90.0%, respectively. In addition, several ablation studies and comparisons were conducted, and the proposed model outperformed the classification of aerial scenes and coastal regions using low-contrast RS images.

Effective task offloading is crucial for overcoming constraints like resource limitations, latency, and energy consumption in Internet of Things (IoT)-enabled Uncrewed Aerial Vehicles (UAVs) running in edge-cloud computing environments. Due to ineffective handling of task offloading many issue can arise, such as higher latency, increased computational power, and utilization of extra resources. It is essential to efficiently optimize task offloading. This can be resolved by optimizing task allocation and resource utilization, which is achieved by the dynamic and intelligent integration of Double Deep Q Networks (DDQN) and Software-Defined Networking (SDN). In this article, we present a hybrid SDQNEC, a new architecture that uses DDQN to optimize task offloading in actual time and SDN for centralized network control and administration. The Markov Decision Process (MDP) formulation of the task offloading issue allows the DDQN to evaluate the environment and identify the best way to distribute tasks across edge and cloud resources. Minimizing total costs and delays, improving resource utilization, and minimizing task rejection rates are the objectives of the suggested approach. Results from simulations show that SDQNEC provides a 40% improvement in resource utilization over baseline DDQNEC models and a 50% lower task rejection rate than classic DQNEC approaches. Furthermore, SDQNEC guarantees effective path selection for task offloading and significantly decreases system costs, attaining up to 55% optimum path utilization under high task loads. These outcomes demonstrate how well the framework works in dynamic and resource-constrained environments to increase task acceptance rates, optimize resource efficiency, and reduce delays. SDQNEC guarantees an effective trade-off between latency and resource cost by strategically assigning tasks to edge or cloud servers according to their availability and resource requirements. This research offers a strong basis for developing edge-cloud computing in Internet of Things networks, with possible uses in vital fields including disaster recovery, industrial automation, and healthcare.

Worldwide, cancer is one of the leading causes of death in humans. Interobserver variability and specialized experience are key factors in diagnosing gastrointestinal tract (GIT) abnormalities using endoscopic procedures. Due to this diversity, small lesions may go unnoticed, leading to a delay in early diagnosis. Therefore, it is essential to design a computer-aided diagnosis (CAD) system for the detection and classification of GIT diseases at the early stages. This paper proposes a CAD system that combines the feature fusion of modified deep learning models with optimal feature selection. Three publicly available datasets, including Kvasir V1, Kvasir V2, and Hyperkvasir, are utilized in the experimental process. In the proposed method, a contrast enhancement step is performed using the fusion of the top-bottom filtering technique. In the next step, two deep learning models (ResNet18 and ResNet50) are modified with a new layer called entropic field propagation (EFP). The pooling layers are replaced with EFP layers in both models, which are then trained on the selected datasets. In the testing process, trained models are employed, and features are extracted from the deeper layers, which are further refined using the Newton-Raphson Marine Predator Optimization (NRMPO) algorithm. The selected features from both models are finally fused using a novel mean threshold-based fusion approach and passed to machine learning classifiers. The proposed CAD system achieved accuracies of 99.0, 89.6, and 82.7% for Kvasir V1, Kvasir V2, and HyperKvasir, respectively. A detailed ablation study is also conducted for the middle steps that validate these reported accuracies. Conclusion: A comparison is performed with state-of-the-art (SOTA) techniques, showing that the proposed method achieves improved accuracy and precision rates.

Convolutional neural networks (CNNs), in particular, demonstrate the remarkable power of feature learning in remote sensing for land use and cover classification, as demonstrated by recent deep learning techniques driven by vast amounts of data. In this work, we proposed a new network-level fusion deep architecture based on 16-tiny Vision Transformer and SIBNet. In the initial phase, data augmentation has been performed to resolve the problem of data imbalances. In the next step, we proposed a self-attention bottleneck-based inception CNN network named SIBNet. In this network, two architectures are followed. The blocks are designed using inception architecture, and each inception module is created with bottleneck blocks. The 16-tiny vision transformer architecture has been implemented for RS images and fused using a network-level fusion with SIBNet for the first time. Hyperparameters of the proposed model have been initialized using Bayesian Optimization for better training on the RS images. After the fusion, the model was on RS image datasets and extracted deep features from the self-attention layer. The extracted features are classified using a neural network classifier with multiple hidden layers. The experimental process of the proposed architecture has been performed on two publically available datasets, such as EuroSAT and NWPU, and obtained an accuracy of 97.8 and 98.9%, respectively. A detailed ablation study has been performed to test the proposed models and shows that the fusion model achieved improved accuracy. In addition, a comparison is conducted with recent techniques and proposed methods, showing improved precision, recall, and accuracy.

Software project outcomes heavily depend on natural language requirements, often causing diverse interpretations and issues like ambiguities and incomplete or faulty requirements.Researchers are exploring machine learning to predict software bugs, but a more precise and general approach is needed.Accurate bug prediction is crucial for software evolution and user training, prompting an investigation into deep and ensemble learning methods.However, these studies are not generalized and efficient when extended to other datasets.Therefore, this paper proposed a hybrid approach combining multiple techniques to explore their effectiveness on bug identification problems.The methods involved feature selection, which is used to reduce the dimensionality and redundancy of features and select only the relevant ones; transfer learning is used to train and test the model on different datasets to analyze how much of the learning is passed to other datasets, and ensemble method is utilized to explore the increase in performance upon combining multiple classifiers in a model.Four National Aeronautics and Space Administration (NASA) and four Promise datasets are used in the study, showing an increase in the model's performance by providing better Area Under the Receiver Operating Characteristic Curve (AUC-ROC) values when different classifiers were combined.It reveals that using an amalgam of techniques such as those used in this study, feature selection, transfer learning, and ensemble methods prove helpful in optimizing the software bug prediction models and providing high-performing, useful end mode.

Cloud solutions accelerate the large‐scale acceptance of IoT projects. By diminishing the need for maintaining on‐premises infrastructure, the cloud has enabled corporations to surpass the traditional applications of IoT (e.g., in‐home appliances) and opened the doors for large‐scale deployment of IoT applications on the cloud. However, shifting legacy systems to the cloud environment can be considerably difficult. Accordingly, this article proposes a method that may support organizations in deciding to modernize their legacy systems. The main concept of this study is to discuss the modernization strategies in detail and to support organizations in selecting the most accurate and appropriate cloud migration strategy, based on their requirements of the legacy system. This article introduces a novel research process, called the K‐means cosine cloud clustering method (K3CM). K3CM is a statistical knowledge‐based method for identifying and clustering the most relevant and similar cloud migration strategies. The quality of a cluster is evaluated by measuring intra‐cohesiveness. Simulation experiments statistically analyzed, evaluated, and verified the quality of K3CM clusters. Correspondence analysis explored the similarity and relationship among cloud migration frameworks and validated the proposed technique. The statistical and simulation results of this study focus on the analytics and decision support system implementation that provides a reliable, valid, and robust clustering method for modernizing the legacy system.

A serious, all-encompassing, and deadly cancer that affects every part of the body is skin cancer. The most prevalent causes of skin lesions are UV radiation, which can damage human skin, and moles. If skin cancer is discovered early, it may be adequately treated. In order to diagnose skin lesions with less effort, dermatologists are increasingly turning to machine learning (ML) techniques and computer-aided diagnostic (CAD) systems. This paper proposes a computerized method for multiclass lesion classification using a fusion of optimal deep-learning model features. The dataset used in this work, ISIC2018, is imbalanced; therefore, augmentation is performed based on a few mathematical operations. After that, two pre-trained deep learning models (DarkNet-19 and MobileNet-V2) have been fine-tuned and trained on the selected dataset. After training, features are extracted from the average pool layer and optimized using a hybrid firefly optimization technique. The selected features are fused in two ways: (i) original serial approach and (ii) proposed threshold approach. Machine learning classifiers are used to classify the chosen features at the end. Using the ISIC2018 dataset, the experimental procedure produced an accuracy of 89.0%. Whereas, 87.34, 87.57, and 87.45 are sensitivity, precision, and F1 score respectively. At the end, comparison is also conducted with recent techniques, and it shows the proposed method shows improved accuracy along with other performance measures.