In the competitive market of wind power forecasting, ensuring data security while enhancing forecasting accuracy is crucial. Federated learning emerges as a key solution to these challenges, offering a way to preserve data privacy and security across multiple entities. This paper presents a groundbreaking approach that leverages Federated Transfer Learning (FTL) in conjunction with Additive Attention Temporal Neural Network architectures, focusing on the competitive wind power forecasting. Our method integrates the sophisticated capabilities of Gated Residual Networks with a strategic process of feature selection, tailored for the unique demands of secure and efficient power forecasting. Through experiments involving data from various geographical regions, this research uncovers the specific challenges of applying federated learning in a competitive landscape. The resulting FTL-Net model demonstrates exceptional performance, surpassing existing benchmarks with metrics such as a Mean Absolute Error (MAE) of 6.011, Mean Squared Error (MSE) of 13,855.45, Root Mean Squared Error (RMSE) of 11.7, a coefficient of determination (\({R}^{2}\)) of 0.9899, and a correlation coefficient of 0.998. These results not only confirm the FTL-Net model's efficacy but also emphasize the significance of meticulous feature selection in federated learning contexts. Compared to other models, FTL-Net's superior performance, bolstered by feature selection, establishes it as a leading approach in the secure and competitive realm of wind power forecasting. This research underscores the potential of combining advanced neural network architectures with federated learning for future advancements in the field, highlighting the importance of data security and collaborative learning in achieving high accuracy in power forecasting.

Real-time non-intrusive monitoring of wind turbines, blades, and defect surfaces poses a set of complex challenges related to accuracy, safety, cost, and computational efficiency. This work introduces an enhanced deep learning-based framework for real-time detection of wind turbine blade defects. The WindDefNet is introduced, which features the Inception-ResNet modules, Visual Transformer (ViT), and multi-scale attention mechanisms. WindDefNet utilizes modified cross-convolutional blocks, including the powerful Inception-ResNet hybrid, to capture both fine-grained and high-level features from input images. A multi-scale attention module is added to focus on important regions in the image, improving detection accuracy, especially in challenging areas of the wind turbine blades. We employ pertaining to Inception-ResNet and ViT patch embedding architectures to achieve superior performance in defect classification. WindDefNet’s capability to capture and integrate multi-scale feature representations enhances its effectiveness for robust wind turbine condition monitoring, thereby reducing operational downtime and minimizing maintenance costs. Our model WindDefNet integrates a novel advanced attention mechanism, with custom-pretrained Inception-ResNet combining self-attention with a Visual Transformer encoder, to enhance feature extraction and improve model accuracy. The proposed method demonstrates significant improvements in classification performance, as evidenced by the evaluation metrics attain precision, recall, and F1-scores of 0.88, 1.00, and 0.93 for the damage, 1.00, 0.71, and 0.83 for the edge, and 1.00, 1.00, and 1.00 for both the erosion and normal surfaces. The macro-average and weighted-average F1 scores stand at 0.94, highlighting the robustness of our approach. These results underscore the potential of the proposed model for defect detection in industrial applications.

The Software-Defined Networking (SDN) paradigm has significantly improved network efficiency by integrating machine learning (ML) capabilities into the control plane (CP). This integration allows adaptive management of network resources in response to dynamic traffic conditions. However, the significant geographical distance between the CP and data plane (DP) causes considerable round-trip latency, measured in milliseconds, which poses a major challenge to time-sensitive traffic. To address this challenge, our research proposes a novel in-network Reinforcement Learning (RL) inference framework. This framework extends programmability from the CP to the DP, enabling timely and precise control of network resources to meet the stringent Quality of Service (QoS) requirements of mission-critical applications. The in-network RL inference is implemented using match-action tables within the Protocol Independent Switch Architecture (PISA) framework in the DP and validated through hardware deployment of protocol-independent packet processors (P4). To allocate bandwidth precisely based on QoS requirements, a P4 meter extern is used to differentiate the unique demands of individual traffic flows. Our enhanced deep-deterministic policy gradient (eDDPG)-based RL inference achieves superior performance with minimal processing overhead, reducing latency by 88.7% and jitter by 89.1 % compared to systems without RL inference.

Fire detection is crucial for safeguarding lives and property. In recent years, advancements in deep learning and Internet of Things (IoT) architectures have significantly enhanced the accuracy of fire and smoke detection systems. In this study, we introduce FlareNet, a feature fusion based deep learning model that leverages a DenseNet architecture combined with Spatial Pyramid Pooling (SPP) and a Contextual Feature Pyramid Network (CFPN). FlareNet is further augmented with dual attention mechanisms and a Feature Enhancement Attention (FEA) mechanism to selectively emphasize critical features for distinguishing between fire and non-fire scenes. Our proposed model is rigorously evaluated across five diverse datasets: Sharma, Deep Quest, BoWFire, FD dataset, and our novel MixFire dataset, achieving an impressive average accuracy of 99.2%. A comparative evaluation against state-of-the-art (SOTA) algorithms reveals that FlareNet outperforms existing methods with a notable improvement in accuracy, precision, recall, and F1-score, thereby setting a new benchmark in the domain of fire detection. Furthermore, A comprehensive analysis with baseline models such as VGG16, VGG19, ResNet18, MobileNetV2, and DenseNet is also presented. These advancements underscore FlareNet’s capability to enhance fire detection systems for a more sustainable environment. FlareNet code and MixFire dataset can be accessed by https://github.com/adeelferozmirza/FlareNet.

The rapid advancement of deepfake technology presents significant challenges for video authenticity verification, necessitating robust detection mechanisms. This paper introduces a novel framework for deepfake video detection that integrates the Multi-Scale Spectral-Guided Graph Attention Network (MSG-GAT) with the Lightweight Assimilation-Elimination (Lite-ASEL) algorithm. By representing video frames as graphs, the framework effectively captures intricate pixel relationships, enabling the detection of subtle manipulations with enhanced precision. Additionally, the Lite-ASEL algorithm is employed for feature selection, balancing reduced computational complexity with high detection performance. Experimental results demonstrate the superiority of the proposed framework, achieving state-of-the-art performance with an AUC of 99.1 % and a detection accuracy of 99.3%. Furthermore, hyperparameter tuning confirms the framework's robustness and efficiency, consistently achieving an optimal objective score of 98.54%, validating its effectiveness for optimal feature selection and deepfake detection.