Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Khalil M. Abdelnaby has a total of 14 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Label Noise and the Accuracy Ceiling in Arabic Sentiment Benchmarks: A Reconciled Error Audit of HARD with Replication on ASTD and ArSAS

    by Mohammad O. Alhawarat, Ayman J. Alnsour, Khalil M. Abdelnaby, Mohammed A. F. Al-Husainy

    Arabic sentiment benchmarks are widely used to rank models, yet how much of their residual error reflects genuine model failure rather than incorrect ground-truth labels is rarely audited directly. We report a reconciled, two-annotator audit of every error (n = 401) made by the best-performing model on the Hotel Arabic Reviews Dataset (HARD), with moderate item-level agreement (Cohen’s κ = 0.41–0.42), categorizing each as label noise, genuine model error, or irreducibly ambiguous. We find that 43.4% of residual errors (49.2% of false positives, 32.6% of false negatives) are attributable to incorrect ground truth. Instantiating the random classification noise model with this measurement implies, for any classifier whose errors are independent of the label noise (a conditional assumption we flag, not a demonstrated fact), a measurable-accuracy ceiling of about 98.35% on HARD derived from a lower-bound noise estimate, leaving at most roughly 2.11 points of headroom for future work. Against that ceiling we place a five-seed comparison of four ensemble-combination rules over four Arabic transformer backbones, in which stacking’s gain over the best single model is modest and not robust: 96.40% vs. 96.24% macro-F1, nominally p = 0.023 but not significant after multiple-comparison correction, and no ensemble rule reliably beats the best backbone. Replicating the full protocol on ASTD and ArSAS shows that neither the backbone ranking nor the ensemble benefit generalizes to informal-register data. We release the reconciled annotations, protocol, and evaluation code.

  • Open Access

    A Systematic Multi-Paradigm Evaluation Framework for Network Intrusion Detection in Fog-IoT Environments: Deep Learning, Transformer, and Ensemble Methods Across Deployment Tiers

    by Khalil M. Abdelnaby

    This paper presents the Adaptive Confidence-Gated Ensemble (ACGE) framework for Network Intrusion Detection Systems (NIDSs) in resource-heterogeneous fog-IoT deployments. No single algorithmic paradigm simultaneously achieves high minority-class detection accuracy, sub-millisecond inference latency, and generalization across heterogeneous IoT traffic at resource-constrained fog-IoT nodes. To address this challenge, ACGE integrates a Sequence Transformer Encoder with LightGBM via a learned and sample-adaptive gating network trained on inter-model error complementarity, formally specified within a three-mode tiered deployment architecture. Both models execute in parallel for every sample; the gating network assigns transformer-dominant weight (α ≈ 1) to 18% of samples where attention-based processing yields the greatest architectural advantage. Under true parallel inference, ACGE latency is 1.45 ms (bounded by the transformer); the measured wall-clock latency under CPU-GPU pipeline parallelism is 1.52 ms. The framework is evaluated under 5-fold stratified cross-validation with McNemar’s significance testing on IoTID20 (primary IoT benchmark) and NSL-KDD (reproducibility anchor) and validated through zero-shot transfer on CICIoT2023 and Edge-IIoTset. Notably, the Sequence Transformer Encoder and ACGE process L = 5 consecutive flow records per decision, providing five times the information available to single-flow baseline models; approximately 57% of the transformer’s performance advantage is attributable to architectural design rather than extended input access. On IoTID20 multiclass classification, ACGE is the only evaluated configuration yielding a negative accuracy-to-macro-F1 gap (−0.51 pp), reaching 99.12 ± 0.03% accuracy and 98.67 ± 0.08% balanced accuracy, demonstrating genuine minority-class improvement. LightGBM achieves the optimal fog-tier profile (0.97 ms latency, concept drift sensitivity of −0.53 pp over three months). On Edge-IIoTset zero-shot transfer, ACGE achieves 93.89% binary accuracy and 85.78% multiclass macro-F1. Ransomware detection (12.34–66.34% F1 across all models) is identified as the primary capability gap; targeted data collection, few-shot adaptation, and federated learning are recommended as the most critical future directions. All inter-paradigm performance differences are reported as the mean ± SD and verified by McNemar’s test with Bonferroni correction.

  • Open Access

    Event-Driven Highlight Generation in Football Broadcasts: A Context-Aware Temporal Framework for Six-Class Action Spotting

    by Khalil M. Abdelnaby

    This paper presents an end-to-end system for the automatic summarization of football matches through temporally accurate event spotting across six action categories: goal, card, substitution, kick-off, direct free-kick, and corner. These six classes were selected from the SoccerNet-v2 annotation taxonomy as a non-standard evaluation subset: the three standard SoccerNet-v2 classes (goal, card, substitution) are retained and extended with three additional context-dependent set-piece classes (kick-off, direct free-kick, and corner) that represent the primary unresolved challenge in temporal action spotting. We developed and empirically evaluated a framework based on a Temporal Convolutional Network (TCN) with capsule-based relational encoding applied to class-specific Time-Shift Encoded (TSE) labels using a composite segmentation-detection loss. Evaluated on the SoccerNet-v2 benchmark (300/100/100 train/validation/test splits) using the official PCA-compressed ResNet-152 features sampled at 2 FPS, the implemented TCN-Capsule baseline achieved a test-set mean Average Precision (mAP) of 58.16% at a 10 s tolerance threshold (δ = 20 frames), with a validation mAP of 58.55%, indicating strong generalization with only a 0.39 pp performance gap. The six per-class AP values underlying this result are Goal: 75.6%, Corner: 78.0%, Substitution: 56.9%, Card: 53.9%, Direct Free-kick: 52.8%, and Kick-off: 34.3%; the headline mAP is computed by the official SoccerNet area-under-precision-recall-curve script and is not a simple arithmetic average of these six values. In addition, five complementary evaluation metrics are introduced: Average Temporal Error (ATE), Temporal Recall@K (TR@K), Latency-to-Event (L2E), Temporal Causality Consistency (TCC), and Annotation Noise Sensitivity. The analysis reveals that visually salient events such as goals and corners are detected with high accuracy (AP > 75%, ATE < 5 s, AUC > 0.95 at frame-level), whereas context-dependent events, including kick-offs and direct free-kicks, remain challenging (AP < 55%, ATE > 8 s, AUC = 0.68–0.72). As future work, two architectural extensions are formally specified but not empirically evaluated: a Temporal Segment Network with Non-Local Blocks and a hybrid Transformer-CNN framework with optical flow fusion. Furthermore, an OpenCV-based highlight generation prototype is implemented; formal quality evaluation is planned as future work.

  • Open Access

    Optimizing Energy-Efficient Resource Allocation in 5G Autonomous Vehicle Networks Through Deep Reinforcement Learning

    by Khalil M. Abdelnaby, Mohammed A. F. Al-Husainy, Mohammad O. Alhawarat, Mohamed A. Rohaim, Khairy M. Assar, Khaled A. Elshafey

    AVs are also bound to capitalize on 5G networks, which creates crucial challenges in the adaptable management of resources because they need very low latency, a high-speed connection, and energy-efficient functionality. Older approaches to optimizing resource allocation in the high-frequency changing vehicle environment fail to deliver as mobility trends and network status constantly adapt and change. To overcome these problems, we suggest a new Deep Reinforcement Learning (DRL)-based algorithm, which is aimed at optimizing the allocation of resources to AVs. This model combines a Spatiotemporal Graph Convolution Network (ST-GCN), Gated Recurrent Units (GRU), and Multi-Agent Deep Deterministic Policy Gradient (MADDPG) to create a unified model. The ST-GCN is successful at both capturing the dynamic space relationship between vehicles and between vehicles and roadside infrastructure, and also gives a complete picture of network topology. GRU uses traffic and communication information to forecast future mobility patterns and bandwidth demand of each agent and therefore allocate resources proactively. The MADDPG algorithm is used to enable decentralized but coordinated decision-making among AVs, which enables the realization of dynamic policies of bandwidth allocation in real-time. Simulations using such aspects as a realistic Rayleigh fading channel model, a node density of 100 vehicles/km2, and 100 MHz of bandwidth prove the effectiveness of the framework extensively. We find that the end-to-end latency increase is reduced by up to 30%, and the system throughput is increased by up to 28, and the energy efficiency is increased by an average of 40 percent in comparison with the baseline techniques. Such results confirm our framework to be a plausible solution to building effective and sustainable communication systems to enable AVs to cooperate in the information exchange of important data.

  • Open Access

    Energy-efficient wireless network control via spatio-temporal deep learning and multi-agent reinforcement learning

    by Khalil M. Abdelnaby, Mohammed A. F. Al-Husainy, Mohammad O. Alhawarat, Mohamed A. Rohaim, Khairy M. Assar

    The rapid expansion of wireless data traffic is placing increasing strain on the energy consumption of current communication networks, intensifying the tension between performance and sustainability objectives. In interference-intensive multiple access scenarios such as power-domain non-orthogonal multiple access (NOMA), energy-efficient optimization is particularly challenging due to the strong coupling between power control and resource allocation decisions. In order to solve this issue, this paper introduces an AI-Enhanced Energy Optimization Framework (AEEOF), which uses deep spatio-temporal learning and reinforcement learning to provide adaptive and energy-aware network control. The proposed framework incorporates a Spatio-Temporal Graph Convolutional Network (ST-GCN) to learn spatial interference relationships and a Gated Recurrent Unit (GRU) to capture temporal traffic dynamics, embedding the resulting representations into a Multi-Agent Deep Deterministic Policy Gradient (MADDPG) controller to support sequential decision-making. Such a design provides the framework with the ability to dynamically distribute power and timing policies based on changes in network conditions. Extensive simulations in a realistic 5G-oriented environment of high interference levels prove the significant performance improvement. The suggested solution can save up to 15% of energy and make the energy use more efficient by about 40%, which is the number of bits delivered per joule. The overall system throughput goes up by 6.25%, the cell-edge user data rate goes up by up to 60%, the fairness goes up by 20%, and the chance of an outage goes down by 70%. A systematic ablation study with three architectural variants validates the individual contribution of each core component - the ST-GCN spatial module, the GRU temporal module, and the MADDPG reinforcement learning controller. Comparative evaluation against conventional orthogonal and non-AI-assisted baselines further supports the effectiveness of the proposed framework within the studied setting. These findings indicate that intelligent spatio-temporal learning is a promising approach for improving energy efficiency and network performance in interference-intensive wireless environments, as demonstrated within the studied 5G-oriented simulation setting.

  • Open Access

    Limit-Cycle Proliferation Under Parametric Delayed Feedback in a Conductance-Based Neuron: Bifurcation Landscape, Orbit Catalog, and Capacity Analysis

    by Mohammad O. Alhawarat, Ayman J. Alnsour, Mohammed A. F. Al-Husainy, Khalil M. Abdelnaby

    We show that a single Hodgkin–Huxley (HH) neuron with Pyragas-type delayed feedback control (DFC) can store multiple symbols as stable periodic orbits, where the specific orbit is selected by tuning the DFC gain K and time delay τ. Sweeping the (K,τ) parameter plane at fixed bias current Ibias = 10.0 μA/cm2 reveals 207 orbit types across 12 topological categories, with inter-spike interval (ISI) means from 5.9 to 56.9 ms. We establish: (i) a write protocol that reliably locks orbits with 13.9 ms median settling time; (ii) a novel Pattern-Oriented Limit-cycle Decoder (POLD) that reads orbits at 100% accuracy from only five observed ISIs (1200 trials across 12 orbits; Wilson 95% CI: 99.7–100%); (iii) a complete single-symbol write–read–erase (W–R–E) cycle with 100% read accuracy, 92% erase verification, and no decay over hold durations up to 50 s; and (iv) a fully validated 12-symbol memory capacity with a read-discriminable upper bound of 67 symbols (11.2× over rate coding; write viability confirmed only for the conservative 12-symbol subset). Reliable orbit addressing needs delay precision of ±2%, which constitutes a write-precision specification and not a fundamental capacity limit. These findings show that parametric delayed feedback is a viable mechanism for limit-cycle-based information storage in conductance-based spiking neurons. The biological interpretation is analogical, not direct: the ±2% delay-precision requirement exceeds what has been demonstrated for biological autaptic variability, and the orbit-coded memory framing is best understood as a computational proof-of-principle aimed at neuromorphic engineering, not as a claim about biological working memory.


  • Open Access

    Causality-Aware and Explainable Self-Supervised Spatio-Temporal Graph Learning for Hardware Trojan Detection

    by Khalil M. Abdelnaby

    As hardware Trojans (HTs) are becoming increasingly stealthy in global semiconductor supply chains, the need for both robust and explainable detection methods is pressing. The use of deep learning models (e.g., Siamese networks, Transformer models) in side-channel signals has shown promising detection accuracy. Yet, they are black-box, data-intensive, and do not expose the causal, structural, and temporal relationships that indicate the presence of HTs. In this paper, we present a causality-focused and explainable detection framework that goes beyond pattern matching. We develop a Self-Supervised Spatio-Temporal Graph Neural Network (SST-GNN) that embeds spatio-temporal side-channel information. Our approach builds a graph that models gate-level components as nodes with temporal power and electromagnetic (EM) features, and functional and physical connections as edges. To address label scarcity, a common problem in real-world applications, we leverage a self-supervised pretraining approach. In particular, a context-aware contrastive loss allows the model to differentiate valid augmentations of benign subgraphs and their side-channel signatures, thus capturing general representations of benign components without Trojan labels. This involves a Causality-Aware GNN (CA-GNN) layer, which embeds differentiable causal discovery into graph learning. This process decouples correlation from causation, identifying the pathways potentially affected by HT trigger and payload. To explain decision making, a gradient-based graph explainer localizes minimal decisive subcircuits and pivotal time windows, generating intuitive detection reports. We evaluated our method on the IEEE Hardware Trojan Side-Channel Dataset (with netlist data), achieving state-of-the-art results (F1 > 0.98). In particular, the model achieves over 60% improvement in Trojan localization precision and false-positive rate, compared to Transformer-based approaches, with high label efficiency and adversarial robustness.

  • Open Access

    GIS-Centric Operational Control of Medium-Voltage Distribution Networks: A Cost-Effective Framework Eliminating ADMS Dependency Through Embedded Switching Intelligence and Real-Time Topological Visualization

    by Khalil M. Abdelnaby

    The operational control of medium-voltage (MV) distribution networks has conventionally relied on a tightly integrated, multi-platform architecture comprising a Supervisory Control and Data Acquisition (SCADA) system, an Advanced Distribution Management System (ADMS), and a Geographic Information System (GIS), interconnected through middleware integration layers. This architecture imposes substantial capital expenditure—typically USD 3.5–4.5 million per control center deployment—and introduces structural data divergence between the ADMS operational model and the GIS geographic representation, with synchronization lags ranging from 24 h to seven days under standard batch update configurations. This paper proposes, develops, and validates a GIS-native operational control framework for MV distribution networks that eliminates the structural dependency on a standalone ADMS by embedding switching intelligence, real-time topology processing, and georeferenced operational visualization directly within the GIS platform. The framework comprises four tightly integrated components: a Unified Spatial Data Model (USDM) serving as the single authoritative network state store; an Embedded Topology Engine (ETE) implementing a loop-safe Breadth-First Search algorithm for real-time energization state computation; a Real-Time Visualization Engine (RTVE) providing continuous georeferenced display of the live network operational state; and a Switching Control Module (SCM) with a Three-State Switch Position Logic (TSPL) conflict resolution mechanism ensuring switching state integrity under concurrent RTU and operator command conditions. The framework was validated on a live operational Egyptian 11 kV distribution network comprising 312 switching elements and 42,650 customers across seven representative switching scenarios. Validation results demonstrate: zero switching state divergence (δ(t) = 0) across all 200 verification points; 100% topological correctness across all 37 switching steps; end-to-end processing latency consistently below 400 milliseconds per switching operation, representing a 14×–67× improvement over the conventional batch GIS synchronization latency; an 88–89% reduction in deployment CAPEX relative to the conventional multi-platform architecture; and a 74–75% reduction in ten-year total cost of ownership inclusive of platform licensing, custom development maintenance, and operational expenditure. The single-platform architecture additionally eliminates 100% of inter-system integration interfaces, removing the primary class of synchronization failure modes inherent to multi-platform deployments. These results establish the proposed framework as a technically rigorous and economically viable operational control solution for MV distribution utilities operating under capital-constrained conditions, with direct applicability to distribution utility sectors across Egypt, the broader MENA region, and developing-world utility environments.

  • Open Access

    Limit-Cycle Proliferation Under Parametric Delayed Feedback in a Conductance-Based Neuron: Bifurcation Landscape, Orbit Catalog, and Capacity Analysis

    by Mohammad O. Alhawarat, Ayman J. Alnsour, Mohammed A. F. Al-Husainy, Khalil M. Abdelnaby

    We show that a single Hodgkin–Huxley (HH) neuron with Pyragas-type delayed feedback control (DFC) can store multiple symbols as stable periodic orbits, where the specific orbit is selected by tuning the DFC gain K and time delay τ. Sweeping the (K,τ) parameter plane at fixed bias current Ibias=10.0μA/cm2 reveals 207 orbit types across 12 topological categories, with inter-spike interval (ISI) means from 5.9 to 56.9 ms. We establish: (i) a write protocol that reliably locks orbits with 13.9 ms median settling time; (ii) a novel Pattern-Oriented Limit-cycle Decoder (POLD) that reads orbits at 100% accuracy from only 5 observed ISIs; (iii) a full write–read–erase (W–R–E) cycle with 100% read accuracy, 92% erase verification, and no decay over hold durations up to 50 s; and (iv) a fully validated 12-symbol memory capacity, with a read-discriminable upper bound of 67 symbols (11.2× over rate coding) pending write-viability confirmation for the extended set. Reliable orbit addressing needs delay precision of ±2%, which constitutes a write-precision specification and not a fundamental capacity limit. These findings show that parametric delayed feedback is a viable mechanism for limit-cycle-based information storage in conductance-based spiking neurons. The biological interpretation is analogical, not direct: the ±2% delay-precision requirement exceeds what has been demonstrated for biological autaptic variability, and the orbit-coded memory framing is best understood as a computational proof-of-principle aimed at neuromorphic engineering, not as a claim about biological working memory.

  • Open Access

    The Intelligent Home: A Systematic Review of Technological Pillars, Emerging Paradigms, and Future Directions

    by Khalil M. Abdelnaby, Mohammed A. F. Al-Husainy, Mohammad O. Alhawarat, Mohamed A. Rohaim, Khairy M. Assar, Khaled A. Elshafey

    Home automation is undergoing a paradigm shift from connected IoT environments with rule-based control to intelligent homes exhibiting ambient intelligence and proactive adaptation. Artificial intelligence, privacy-preserving sensing, and converging connectivity standards are the primary forces driving this transition. This systematic literature review synthesizes the technological foundations, architectural developments, emerging paradigms, and socio-technical challenges characterizing the next generation of smart homes, evaluated against the original Ambient Intelligence (AmI) vision. Following PRISMA 2020 guidelines, searches were conducted across four databases—IEEE Xplore, ACM Digital Library, Scopus, and Web of Science—covering studies published between January 2020 and June 2025. From 3450 records, 113 studies were selected through a two-reviewer screening procedure with inter-rater reliability assessments. Quality was assessed using a modified JBI Critical Appraisal Checklist, and findings were synthesized through thematic analysis. Three converging technological pillars were identified: multi-modal privacy-preserving sensing including mmWave radar; a hierarchical cloud-edge-TinyML intelligence engine; and unified connectivity through the Matter/Thread standard. Emerging paradigms include LLM-based cognitive orchestration, hyper-personalization, Digital Twin simulation, and grid-interactive prosumer energy management. Realizing that the intelligent home vision requires addressing the privacy–security–trust trilemma, algorithmic bias, system reliability, and human–agent collaboration, a research roadmap encompassing explainable AI, privacy-by-design, lifelong learning, and standardized ethical auditing is proposed.