Publications Scientifiques

Permanent URI for this communityhttps://dspace.univ-boumerdes.dz/handle/123456789/10

Browse

Subject: search.filters.subject.Deep learning

Search Results

Now showing 1 - 10 of 22
  • No Thumbnail Available
    Item
    Enhancing Data Privacy in Intrusion Detection: A Federated Learning Framework With Differential Privacy
    (John Wiley and Son, 2025) Saidi, Ahmed; Khouri, A. Ouadoud
    The rise of cyber threats has underscored the critical need for robust intrusion detection systems (IDS). While traditional approaches, including statistical, knowledge-based, and AI-driven methods, have been pivotal, they often face limitations such as data privacy concerns, scalability challenges, and low detection accuracy on unfamiliar threats. This paper addresses these issues by adopting a federated learning (FL) paradigm for collaborative intrusion detection, allowing data to remain local and enhancing privacy protection. The proposed solution integrates advanced encryption techniques and differential privacy to safeguard confidentiality while ensuring system scalability and adaptability. By introducing a robust separation of agents' roles and leveraging FL's decentralized architecture, the system overcomes the limitations of centralized learning, including single points of failure and communication overhead. Experimental results validate the proposed architecture, demonstrating significant improvements in performance and offering a promising direction for modern network security. This work not only highlights the potential of FL-based IDS but also explores the integration of distributed ledger technologies to further enhance trust and security. These findings contribute to the growing field of privacy-preserving computing and lay the groundwork for future innovations in scalable, secure, and efficient intrusion detection systems
  • Thumbnail Image
    Item
    Seasonal quantile forecasting of solar photovoltaic power using Q-CNN-GRU
    (Nature Research, 2025) Ait Mouloud, Louiza; Kheldoun, Aissa; Oussidhoum, Samira; Alharbi, Hisham; Alotaibi, Saud; Alzahrani, Thabet
    Accurately predicting solar power is essential for ensuring electric grid reliability and integrating renewable energy sources. This paper presents a novel approach to probabilistic solar power forecasting by combining Convolutional Neural Networks (CNN) with Gated Recurrent Units (GRU) into a hybrid Quantile-CNN-GRU model. The proposed model generates intra-day probabilistic quantile forecasts and is rigorously evaluated using datasets from geographically and climatically diverse regions and hemispheres: the Netherlands (temperate maritime climate), Alice Springs (arid desert climate), and Hebei (humid subtropical climate). These datasets cover varied temporal horizons (1-hour, 6-hour, 12-hour, and 24-hour predictions) and seasonal conditions (summer, fall, spring, and winter), highlighting the model’s adaptability to different scenarios. The performance of the proposed Quantile-CNN-GRU model is benchmarked against state-of-the-art deep learning models, including standalone quantile-based architectures such as Quantile-GRU and Quantile-Long Short Term Memory (LSTM). A comprehensive evaluation framework is applied, employing probabilistic tools like the Continuous Ranked Probability Score (CRPS) for assessing forecast reliability, sharpness, and reliability diagrams with consistency bars to evaluate the calibration of the predictions. Results demonstrate that the proposed Quantile-CNN-GRU model consistently outperforms its counterparts in terms of CRPS, across varying forecast horizons and seasonal conditions. To further enhance performance, a multivariate case study incorporating exogenous inputs, specifically Numerical Weather Prediction (NWP) data, is conducted. Through sensitivity analysis, the influence of these additional inputs on forecast horizons and seasonal variability is systematically explored. The study reveals that integrating NWP data significantly improves the model’s predictive skill, particularly for longer forecast horizons and during transitional seasons like spring and fall, when solar variability is higher.
  • No Thumbnail Available
    Item
    Enhancing Data Privacy in Intrusion Detection: A Federated Learning Framework With Differential Privacy
    (John Wiley and Sons Ltd, 2025) Saidi, Ahmed; Khouri, A. Ouadoud
    The rise of cyber threats has underscored the critical need for robust intrusion detection systems (IDS). While traditional approaches, including statistical, knowledge-based, and AI-driven methods, have been pivotal, they often face limitations such as data privacy concerns, scalability challenges, and low detection accuracy on unfamiliar threats. This paper addresses these issues by adopting a federated learning (FL) paradigm for collaborative intrusion detection, allowing data to remain local and enhancing privacy protection. The proposed solution integrates advanced encryption techniques and differential privacy to safeguard confidentiality while ensuring system scalability and adaptability. By introducing a robust separation of agents' roles and leveraging FL's decentralized architecture, the system overcomes the limitations of centralized learning, including single points of failure and communication overhead. Experimental results validate the proposed architecture, demonstrating significant improvements in performance and offering a promising direction for modern network security. This work not only highlights the potential of FL-based IDS but also explores the integration of distributed ledger technologies to further enhance trust and security. These findings contribute to the growing field of privacy-preserving computing and lay the groundwork for future innovations in scalable, secure, and efficient intrusion detection systems
  • No Thumbnail Available
    Item
    Transformer-Based Approach for Intrusion Detection System
    (Institute of Electrical and Electronics, 2025) Senoussi, Nour El Houda; Salmi, Cheikh; Banouh, Nassim; Khalfi, Adem
    Intrusion Detection Systems (IDS) have been used for years to protect enterprise hosts from cyberattacks. Traditional IDSs are usually based on simple methods, such as signatures or heuristics, that do not adapt to reactions against new threats that are constantly increasing. The objective of this paper is to develop an IDS based on a deep learning technique which is transformers. Unlike conventional models and thanks to their self-attention mechanism, transformers are characterized by an excellent ability to support complex patterns by very accurately modeling the context in sequential data. A host-based dataset containing system logs and network activities is used to train the transformer model that forms the core of the developed IDS. A detailed evaluation is used to compare our approach against existing methods based on machine learning and deep learning, showing significant improvements in precision, recall, and false positive rate. These results are very encouraging for developing robust IDSs that can be fine-tuned in real time to take into account new attacks
  • Thumbnail Image
    Item
    Deep Learning Models to Analyze Sentiments of People Regarding New Vaccines
    (Institute of Electrical and Electronics Engineers, 2025) Khoudi, Asmaa; Draoui, Yasmine; Aoutou, Nadjet
    The COVID-19 pandemic has generated a vast corpus of online conversations regarding vaccines, predominantly on social media platforms like X (formerly known as Twitter). However, analyzing sentiment in Arabic text is challenging due to the diverse dialects and lack of readily available sentiment analysis resources for the Arabic language. This paper proposes an explainable Deep Learning (DL) approach designed for sentiment analysis of Arabic tweets related to COVID-19 vaccinations. The proposed approach utilizes a Bidirectional Long Short-Term Memory (BiLSTM) network with Multi-Self-Attention (MSA) mechanism for capturing contextual impacts over long spans within the tweets, while having the sequential nature of Arabic text constructively learned by the BiLSTM model. Moreover, the XLNet embeddings are utilized to feed contextual information into the model. Subsequently, two essential Explainable Artificial Intelligence (XAI) methods, namely Local Interpretable Model-Agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP), have been employed for gaining further insights into the features’ contributions to the overall model performance and accordingly achieving reasonable interpretation of the model’s output. Obtained experimental results indicate that the combined XLNet with BiLSTM model outperforms other implemented state-of-the-art methods, achieving an accuracy of 93.2% and an F-measure of 92% for average sentiment classification. The integration of LIME and SHAP techniques not only enhanced the model’s interpretability, but also provided detailed insights into the factors that influence the classification of emotions. These findings underscore the model’s effectiveness and reliability for sentiment analysis in low-resource languages such as Arabic
  • Thumbnail Image
    Item
    A Residual Temporal Convolutional with Attention Neural Network for Electromyogram-Based Hand Gesture Recognition
    (2025) Namane, Rachid; Boutellaa, Elhocine; Salem, Sif Eddine; Babaci, Yassine
    Electromyography (EMG)-based hand gesture classification is a developing core technology for designing intuitive and responsive human-computer interaction, notably for prosthetic control. EMG signals, which reflect muscle activity during contraction, offer a non-invasive and effective method for capturing user gestures. However, because of their natural variability, noise, and temporal richness pose significant hurdles to precise gesture recognition. In this paper, we investigate the use of causal convolutional layers, which are suitable for sequential data, to improve hand gesture recognition from raw EMG signals. We propose a deep neural network which bases on temporal convolutions and integrates residual connections and contextual attention in an end to end hand gesture recognition system. Furthermore, we apply multiple data augmentation techniques to mitigate intra-subject variability and enhance model generalization. Our approach is evaluated on the benchmark NinaProDB1 dataset. The proposed model show impressive classification performance with an average accuracy of 95.31% and where the majority of the gestures from various subjects were accurately recognized. These results demonstrate the effectiveness of causal convolutions and attention mechanisms for robust EMG-based gesture recognition.
  • Thumbnail Image
    Item
    Deep Learning-Based Fish Health Monitoring and Diagnosis: A Review
    (IEEE, 2025) Kheriji, Lazhar; Kouadri, Abdelmalek; Mansouri,Majdi
    Fish in aquaculture systems face health challenges influenced by aging, water quality, and environmental conditions. These issues affect critical components like feeding and filtration, potentially reducing efficiency and causing system failure. Effective Health Monitoring and Diagnosis (HMD) relies on high-quality features such as behavior, physical condition, feeding habits, and water parameters. However, traditional hand-crafted approaches often fail to capture the complex and nonlinear interactions between biological and environmental factors, limiting their adaptability to sudden changes in water conditions or disease outbreaks. This gap motivates the use of intelligent, multimodal learning strategies that integrate diverse data sources for more robust and reliable analysis. Advances in computing power, large datasets, and sophisticated algorithms have made deep learning (DL) a transformative tool in this field. By combining DL with multimodal data integration, it becomes possible to learn high-level representations directly from heterogeneous inputs such as water quality measures, behavioral signals, and visual observations, thereby overcoming the limitations of conventional feature-based methods. This paper reviews DL-based multimodal approaches in aquaculture HMD, comparing recent techniques, their strengths, and limitations. We also discuss future directions, emphasizing multimodal data fusion to enhance DL-driven health monitoring. This review provides a concise resource for researchers and practitioners aiming to advance aquaculture health monitoring.
  • Thumbnail Image
    Item
    Review on deep learning optimization using knowledge and dataset distillation in medical imaging diagnostics
    (2025) Laribi, Nor-Elhouda; Gaceb, Djamel; Rezoug, Abdellah; Touazi, Faycal
    The integration of deep learning-based artificial intelligence solutions in hospital environments introduces significant challenges, including data privacy restrictions, limited computational resources, and constraints related to the quality and simplicity of the models used. In this review, we highlight the recent advancements in knowledge distillation and dataset distillation as emerging solutions to these challenges in the field of medical imaging. These techniques offer practical benefits in clinical settings by enabling faster training, reduced model size, improved inference speed, and enhanced accuracy, while supporting privacy-preserving learning across decentralized systems and edge devices. Knowledge distillation transfers knowledge from a complex to a simple model, enabling efficient deployment without high loss in diagnostic performance. Dataset distillation, by contrast, focuses on synthesizing datasets that match the pretrained model on real data, reducing data storage requirements. Together, these methods improve learning efficiency, model accuracy, and resource optimization in hospital workflows. However, their integration into medical environments also presents limitations. Challenges such as pipeline complexity, scalability issues, and performance inconsistency across architectures or high-resolution tasks still persist. Overall, this review provides a comprehensive overview of potential and limitations of these two types of distillations in healthcare, offering insights into how these methods can support more scalable, accurate, and privacy-aware AI solutions for medical imaging.
  • No Thumbnail Available
    Item
    Deep Learning for Sustainable Aquaculture: Opportunities and Challenges
    (Institute of Electrical and Electronics Engineers Inc., 2025) Kheriji, Lazhar; Kouadri, Abdelmalek; Mansouri, Majdi
    With the rising global demand for aquatic products, aquaculture has become a cornerstone of food security and sustainability. This review comprehensively analyzes the application of deep learning in sustainable aquaculture, covering key areas such as fish detection and counting, growth prediction and health monitoring, intelligent feeding systems, water quality forecasting, and behavioral and stress analysis. The study discusses the suitability of deep learning architectures, including CNNs, RNNs, GANs, Transformers, and MobileNet, under complex aquatic environments characterized by poor image quality and severe occlusion. It highlights ongoing challenges related to data scarcity, real-time performance, model generalization, and cross-domain adaptability. Looking forward, the paper outlines future research directions including multimodal data fusion, edge computing, lightweight model design, synthetic data generation, and digital twin-based virtual farming platforms. Deep learning is poised to drive aquaculture toward greater intelligence, efficiency, and sustainability
  • No Thumbnail Available
    Item
    Offline Arabic handwritten character recognition: from conventional machine learning system to deep learning approaches
    (2022) Faouci, Soumia; Gaceb, Djamel; Haddad, Mohammed
    Researchers have made great strides in the area of Arabic handwritten character recognition in the last decades especially with the fast development of deep learning algorithms. The characteristics of Arabic manuscript text pose several problems for a recognition system. This paper presents a conventional machine learning system based on the extraction of a set of preselected features and an SVM classifier. In the second part, a simplified convolutional neural network (CNN) model is proposed, which is compared to six other CNN models based on the pre-trained architectures. The suggested methods were tested using three databases: two versions of the OIHACDB dataset and the AIA9K dataset. The experimental results show that the proposed CNN model obtained promising results, as it is able to recognise 94.7%, 98.3%, and 95.6% of the test set of the three databases OIHACDB-28, OIHACDB-40, and AIA9K, respectively.