Cybersecurity threats have become increasingly sophisticated and dynamic due to the rapid evolution of malicious software and attack techniques. However, the reliance of traditional IDS/IPS implementations on signature-based detection limits their ability to identify novel, rapidly evolving, and adaptive malware, despite their effectiveness against known cyber threats. This paper presents a unified machine-learning framework for defence against malware-laden URLs, which simultaneously targets intrusion detection (accuracy first, out of band) and intrusion prevention (latency first, inline) through module-aware, bio-inspired feature selection. The detection module employs random forest (RF) with the bat algorithm (BA), achieving 99.52% accuracy under stratified fivefold validation. This approach yields thousands of additional correct decisions per million URLs and reduces false negatives and false positives without adverse throughput effects. The prevention module utilizes a decision tree (DT) with the HHO ∩ BA consensus subset, achieving an inference time of 1 ms per URL with competitive accuracy (98.96%), enabling more than 1,000 URLs per second per core and meeting strict inline decision budgets. Standard metrics, such as accuracy, precision, recall, and F1-score, along with timing analyses, confirm that aligning feature selection with module objectives strikes a practical balance between detection quality and latency. The resulting guidance is direct: employ RF with BA for intrusion detection when optimal discriminative performance is essential, and adopt DT with the HHO ∩ BA subset for intrusion prevention when millisecond-scale latency dictates deployment.

This paper proposes an innovative approach to translating the nonlinear dynamics of a memristive FitzHugh-Nagumo-Hindmarsh-Rose (FN-HR) coupled neuron model into an AI-optimized, resource-efficient VLSI implementation on FPGA platforms, advancing intelligent computing paradigms. The bidirectional memristive synapse coupling FN and HR neurons enables rich dynamic behaviors such as mixed-mode oscillations and chaos, which are harnessed to enhance adaptive machine learning and neural network training. A detailed dynamical analysis, including Lyapunov exponent spectra and synchronization properties, identifies parameter regimes suitable for AI applications. Nonlinear operators are approximated using quantized lookup tables and three-term sinusoidal expansions, achieving RMSE values of 0.0105 (FN) and 0.0114 (HR) while eliminating DSP usage. Synthesized on an AMD Zynq UltraScale+ ZCU104 FPGA, a 50-neuron network utilizes 5.3% LUTs and 7% BRAM, delivering 42 million neuron-updates per second at 210 mW. This work establishes a scalable, low-power platform for real-time AI-driven neuromorphic computing and intelligent adaptive control systems.

Malware continues to pose a critical threat to cybersecurity, necessitating intelligent detection systems capable of adapting to evolving attack strategies. This paper introduces an enhanced malware detection framework that integrates bio-inspired feature selection with advanced gradient boosting classifiers to achieve high accuracy and efficiency. Two metaheuristic algorithms-Harris Hawks Optimization (HHO) and the Bat Algorithm (BA)-are independently applied to extract compact and discriminative feature subsets from the ISCX-URL2016 malware dataset. The reduced feature sets are subsequently evaluated using Gradient Boosting Machine (GBM) and Light Gradient Boosting Machine (LightGBM) classifiers. Performance is rigorously assessed using Accuracy as the primary evaluation metric. Experimental results demonstrate that LightGBM combined with BA achieves the highest performance, reaching an accuracy of 99.52%, precision of 99.48%, recall of 99.48%, and F1-Score of 99.48%. These findings underscore the effectiveness of bio-inspired optimization for feature selection, showing that the proposed framework not only improves predictive performance but also offers a scalable and reliable solution for real-world malware detection.

The exponential growth of malware attacks, particularly those exploiting malicious URLs, poses a significant threat to cybersecurity in real-time digital environments. To address the challenges of high-dimensional feature spaces and the need for fast, accurate detection, this study proposes a hybrid bio-inspired optimization framework that combines Harris Hawks Optimization (HHO) and the Bat Algorithm (BA) for effective feature selection. The framework evaluates two strategies—union (HHO∪BA) and intersection (HHO∩BA)—to balance detection performance and computational efficiency. After feature selection, classifiers including XGBoost and Extra Trees are fine-tuned using Grid Search to ensure optimal performance. Experiments are conducted on the ISCX-URL2016 dataset, which includes a comprehensive set of benign and malware-labeled URLs. Results show that the HHO∪BA approach achieves the highest detection accuracy (up to 99.52%) and robust classification metrics, making it ideal for high-security applications where accuracy is critical. In contrast, the HHO∩BA method offers significantly faster training and inference times, making it more suitable for real-time or resource-constrained environments. These findings highlight the trade-off between accuracy and speed and provide a flexible framework that can be adapted to various cybersecurity deployment scenarios.

In this study, we propose a computationally efficient, multiplier-free digital realization of the four-dimensional Wilson neuron model, a well-established framework for modeling complex neural dynamics. To enable large-scale real-time implementation on digital hardware, we reformulate the nonlinear terms in the model using a hybrid approximation strategy based on power-of-two arithmetic and look-up tables (LUTs), significantly reducing computational complexity while preserving the system’s rich dynamical behavior. This approach leverages advanced optimization techniques to minimize resource usage, making it suitable for Artificial Intelligence applications that require efficient processing of large-scale neural networks. The resulting Digital-Friendly Wilson 4D (DFW4-D) model is validated through several approaches, including numerical error analysis, Lyapunov exponent calculations, and stimulus–response behavior under auditory inputs. The Lyapunov analysis demonstrates that the proposed model retains the stability properties of the original system across a wide range of initial conditions and external currents. Furthermore, extensive FPGA-based synthesis on multiple platforms reveals a drastic improvement in operating frequency—up to 4.8× —with complete elimination of DSP block usage. This work offers a scalable and biologically realistic digital neuron model suitable for neuromorphic signal processing and low-power cognitive computing, with particular implications for embedded biomedical systems and AI-driven applications.

This brief presents a high-efficiency, multiplierless Field-Programmable Gate Array (FPGA) realization of the Rulkov neuron map using a hybrid approximation framework. The proposed approach integrates High-Frequency Point Detection (HFPD) for accurate spiking approximation and Exponential Binary Power (EBP) computed via Coordinate Rotation Digital Computer (CORDIC) for efficient bursting dynamics, fully eliminating multipliers and dividers. Implemented on a Xilinx Virtex-4 FPGA, the design achieves a synthesis frequency of 276.9 MHz, reduces hardware utilization by 30%, and improves speed by 1.27 & times; compared to the baseline. An average Root Mean Squared Error (RMSE) of 0.54 and successful two-coupled and ten-multiplexed implementations confirm the accuracy, scalability, and real-time suitability of the proposed Rulkov-Digital Circuit (R-DC) architecture for neuromorphic computing systems.

This paper presents the implementation of a neuromorphic platform for the Hindmarsh-Rose (HR) neuronal model on a Field-Programmable Gate Array (FPGA). We introduce a modified digital HR (DHR) model that significantly reduces computational overhead while maintaining biological accuracy. The DHR model is validated through comprehensive simulations that compare its spike-timing and dynamical behaviors against the original HR model, demonstrating a high level of similarity across various spiking modes. The hardware implementation is realized on a Virtex 4 FPGA, where the DHR model effectively utilizes only 1% of the FPGA resources and achieves an operational frequency of 480.353 MHz. Additionally, comparisons with existing models highlight the DHR model's superior efficiency and performance, enabling the implementation of large-scale networks. Our model achieves an impressive speed of 480.353 MHz while maintaining a power consumption of only 95.13 mW, resulting in an Efficiency Factor (EF) of 1.54. In comparison to other similar models, our DHR demonstrates superior processing speed and efficiency with acceptable error levels (compared to other similar models), utilizing just 198 flip-flops-which are considerably fewer than many other implementations. This efficient architecture is well-suited for real-time applications. This high-speed, low-power architecture enables real-time neuromorphic processing for autonomous robotics, facilitating rapid decision-making in dynamic environments. In bioengineering, its accurate replication of neuronal dynamics supports advanced neural prostheses and brain-inspired computational models for studying neurological disorders. Additionally, scalable and resource-efficient design paves the way for energy-efficient embedded systems in smart biomedical devices applications.

This paper aims to investigate the application of PKE technologies and its development in the e-government setups along with analyzing the relevance of PKE technologies in supporting secure and efficient communications between government departments and the members of the public. Through analyzing RSA algorithm and advanced lattice-based cryptography, the paper highlights that application of PKE is also significant for defending the newly generated cyber threats while identifies advantages and possible usages of the method. Moreover, the review continues to discuss the actual usability issues associated with the architecture of PKI in egovernment, like key management difficulties and the amount of computation required by the encryption method. Challenges are investigated regarding the ways in which biometric data can be incorporated into e-government services in the future and how quantum-resistant cryptographic methods might potentially be incorporated to improve the stability of e-government services. Thus, when appraising the conceptual topicality of the review based on the analysis of a wide range of sources, it can be noted that the exploration of the importance of PKE in ensuring the provision of safe, effective, and clear e-gov services suggests the necessity of further development and improvement of cryptographic approaches as a response to the dynamic growth of threats in the sphere of the digital environment.

Accurate determination of volume percentages in three-phase fluids is paramount for the success of various industrial processes, ranging from oil and gas production to chemical engineering. This study presents a comprehensive approach to this challenge by leveraging advanced signal processing techniques and machine learning paradigms. Our methodology integrates the time, frequency, and wavelet transform features extracted from X-ray-based measurement systems whose structure consists of an X-ray tube source, two sodium iodide detectors, and a test pipe, all of which were simulated using the Monte Carlo N Particle code. The amalgamation of these features provides a rich representation of the fluid composition that captures both temporal and spectral characteristics. To enhance the discriminative power of the features, we employ a simulated annealing algorithm to strategically reduce their dimensionality and select pertinent features. The simulated annealing unit systematically evaluates the contribution of each feature to predictive accuracy. Further, through iterative elimination and re-evaluation, the algorithm refines the feature set, retaining only those with the highest relevance to the three-phase fluid composition. This feature selection process optimises the performance of subsequent machine learning models, streamlining the input space for enhanced interpretability and efficiency. Finally, to determine the volume percentages, we employ a support vector regression (SVR) neural network, which is trained on a refined dataset with capability to handle complex relationships and high-dimensional data. The proposed approach demonstrates superior accuracy in determining volume percentages of three-phase fluids compared to traditional methods, thereby making it an effective and integrated technique to analyse fluid composition in a variety of industrial settings and applications.

The cochlea, which is a spiral-shaped component of the inner ear, holds significant importance in the auditory process. Its primary function involves transforming sound waves into electrical signals that can be understood by the brain. This paper presents a modified version of 2D artificial Cochlea mathematical modeling using a combined approximation method. The main novelty and contribution of our work is using the trigonometric-based functions and base-2 terms, simultaneously. The method outlined utilizes a precise and optimized approach with minimal errors to estimate a specific nonlinear function with high accuracy called Cochlea Digital Circuit (CDC). Moreover, the intricate multiplication processes in other nonlinear components of the CDC model are streamlined into simpler terms. As a result, this technique can be efficiently applied to FPGAs with minimal hardware resources. The proposed approach utilizes the Coordinate Rotation Digital Computer (CORDIC) algorithm to digitally execute trigonometric functions. Additionally, the rest of the elements in the suggested model can be realized using fundamental digital modules without requiring multiplication operations. The study examines the digital implementation and hardware expenses of the suggested model for extensive deployment. Through hardware synthesis, it is demonstrated that the CDC model can replicate accurate active cochlear behaviors more efficiently and at significantly reduced computational expenses than the original model. We utilized the Virtex-II board for the digital FPGA implementation of the CDC model to validate and synthesize the recommended approach. Our proposed model outperforms the original model in terms of speed and cost efficiency in this context. Additionally, our hardware model can faithfully reproduce the original patterns with minimal discrepancies. Our proposed model achieves a speed-up that is approximately 1.316 times faster than the original model (376.459 MHz versus 286.029 MHz), and the number of fitted CDCs is about 1.457 times higher in our approach (51 compared to 35). As a result, the basic parts of our paper can be presented as five parts. Motivation: Cochlear modeling is vital for understanding auditory processing and developing effective hearing restoration technologies. However, existing cochlear models face challenges related to computational efficiency and accuracy. Methodology: In this study, we introduce a modified 2D artificial cochlea model implemented through a CDC. Our approach utilizes a combination of trigonometric functions and base-2 terms, which eliminates complex multiplication operations and enhances efficiency for digital design. Results: The proposed model achieves a speed-up of approximately 1.316 times compared to the original model. Furthermore, it accommodates 51 fitted CDCs while demonstrating a reduction in computational expense. Significance: The findings from this research have significant implications for the field of auditory research and the development of bio-inspired AI systems, contributing to more effective hearing restoration options for individuals with auditory challenges. In practical scenarios, this method holds promise for advancing the creation of cochlear implants that are both more efficient and cost-effective, offering enhanced hearing restoration options for those with auditory challenges. Furthermore, the insights gained from this research could be utilized to improve the construction and integration of signal processing systems in a range of audio and communication devices, thus laying the groundwork for more sophisticated audio processing technologies that deliver heightened efficiency and minimized hardware expenses. Conclusions: These improvements indicate that our modified cochlea model not only enhances auditory processing technology but also provides a more efficient framework for cochlear implants.

Currently, several algorithms in machine learning (ML) are employed to enhance the detection of attacks on networks. The developers have suggested several solutions for the detection of attacks on networks in the previous works. This paper contributes to the earlier works by employing the Decision Tree (DT) and Naïve Bayes (NB) algorithms on the UNSW-NB15 dataset. In addition, the Bat Algorithm (BA) metaheuristic optimizer is used to lessen the irrelevant and redundant features. The achievement of the DT and NB algorithm classifiers is evaluated in terms of accuracy, Precision, and Recall. The DT achieved 100 % Accuracy, Precision, and Recall. On the other hand, the NB achieved $99.81 \%, 99.87 \%$, and 99.91 % Accuracy, Precision, and Recall, respectively. The results show that DT achieved the optimal results with the BA algorithm, and NB DT achieved nearly optimal results.

Phising attacks continue to pose significant risks to digital security by exploiting user vulnerabilities through deceptive methods. This paper presents a smart firewall model for phishing detection that leverages bio-inspired algorithms to enhance threat identification and response. The model utilizes the Whale Optimization Algorithm (WOA) and Dragonfly Algorithm (DA) independently for effective feature selection, thereby reducing data dimensionality while retaining critical phishing indicators. These optimized features are then processed by advanced Machine Learning (ML) classifiers-Extra Trees (ET), Random Forest (RF), and K-Nearest Neighbors (KNN)-to rigorously evaluate detection accuracy. Experimental results on the ISCX-URL2016 dataset demonstrate that the combination of WOA with the ETs classifier achieves a superior detection accuracy of 98.86%, precision of 99.50%, recall of 99.50%, F1-Score of 99.50%, outperforming alternative configurations and recent methods. This result highlights the potential of bio-inspired optimization combined with ML to develop intelligent, adaptive firewalls capable of effectively mitigating phishing threats.

INTRODUCTION:Underground coal fires pose significant environmental and health risks due to releasing CO2 emissions. Predicting surface CO2 flux accurately in underground coal fire areas is crucial for understanding the distribution of spontaneous combustion zones and developing effective mitigation strategies. In recent years, advanced machine learning techniques have shown promise in various carbon-related studies. This research uses an experimental approach to explore the power of advanced machine learning schemes for predicting CO2 flux in underground coal fire areas. OBJECTIVES:By leveraging the power of advanced machine learning schemes and experimental approaches, this research aims to provide valuable insights into CO2 flux prediction in coal fire areas and inform environmental monitoring and management strategies. METHODS:The study involves the collection of an experimental dataset specific to underground coal fire areas, encompassing various parameters related to CO2 flux and underground coal fire characteristics. Innovative feature engineering techniques are applied to capture the unique characteristics of underground coal fire areas and their impact on CO2 flux. Different machine learning algorithms, including Natural gradient boosting regression (NGRB), Extreme gradient boosting (XGboost), Light gradient boosting (LGRB), and random forest (RF), are evaluated and compared for their predictive capabilities. The models are trained, optimized, and assessed using appropriate performance metrics. RESULTS:The NGRB model yields the best predictive performances with R2 of 0.967 and MAE of 0.234. The novel contributions of this study include the development of accurate prediction models tailored to underground coal fire areas, shedding light on the underlying factors driving CO2 flux. The findings have practical implications for delineating the spontaneous combustion zone and mitigating CO2 emissions from underground coal fires, contributing to global efforts in combating climate change.

This paper presents a pioneering FPGA-based hardware implementation aimed at enhancing insulin delivery systems, a critical component in the management of diabetes. The research introduces a novel Reduced Base-2 Approximation (RB2A) technique for modeling pancreatic /3-cells, focusing on efficient control of blood glucose levels. In extreme manufacturing scenarios where precision and responsiveness are paramount, our approach demonstrates significant improvements in processing speed, power consumption, and resource utilization. Specifically, the new model is shown to consume approximately 32% less power, operate around 2.14 times faster, and require fewer hardware resources compared to conventional models. By simulating the pancreatic functions digitally, this study underscores the potential of integrating advanced biological processes within the realm of extreme manufacturing. The findings suggest that this innovative digital solution can lead to the development of highly efficient, cost-effective insulin delivery systems, paving the way for the integration of smart medical devices capable of real-time monitoring and regulation of critical physiological parameters, thereby enhancing the manufacturing of bioengineered medical solutions. This optimized control model provides the ability to precisely and effectively regulate blood glucose levels, helping diabetes patients gain better control over their health. Additionally, optimizing the blood glucose control model can lead to more efficient implementations that maximize the performance of insulin delivery devices and improve user experiences.

This paper proposes a machine learning-based firewall framework that enhances intrusion detection by applying a union of two metaheuristic algorithms—Bat Algorithm (BA) and Dragonfly Algorithm (DA)—for feature selection. The UNSWNB15 dataset is used, where the union of features selected by both algorithms improves model accuracy while reducing dimensionality. Two classifiers, Naïve Bayes (NB) and Random Forest (RF), are employed to evaluate the performance of the selected features. The Random Search technique is used to optimize hyperparameters. Experimental results show that the RF classifier achieved perfect scores across all evaluation metrics, while NB yielded near-perfect results. These findings confirm the effectiveness of combining BA and DA to improve the precision and efficiency of attack detection. The proposed model offers a promising solution for real-world cybersecurity applications where accurate and fast decision-making is essential.

The advancements in network technology and associated services have increased the volume of data traffic. However, the harmful impact caused by intruders has also increased. Intrusion activities can be addressed by utilizing Network Intrusion Prevention Systems (NIPS). NIPS can detect attacks by either comparing the traffic with predefined patterns or identifying deviations from normal traffic behavior. In this paper, the machine learning algorithms will be used to enhance the performance of the NIPS systems. First, feature selection algorithms will be used to remove irrelevant features from the network traffic data. In particular, the widely used Dragonfly Algorithm (DA) will be used for feature selection. Then, the Gradient-Boosted Trees (GBT) and Logistic Regression (LR) will be used to find the attack data based on the features selected by the DA algorithm. The UNSWNB15 dataset was used to assess the proposed NIPS system. The accuracy of GBT and LR reaches 100 % and 99.93 %, respectively.

This article presents a novel digital design methodology for modeling cochlea-neuron interactions, tailored for applications within the scope of industrial electronics, such as real-time biosensing, smart health interfaces, and resource-aware neural signal processing. The proposed model employs a simplified 2-D cochlear structure based on the Hopf oscillator, optimized using linear shift-adder (ADD)-based functions and look-up table-based sampling to eliminate complex multipliers and achieve hardware-friendly, high-speed performance. This hybrid multiplierless architecture aligns with the journal's focus on efficient digital realization of intelligent systems and field-programmable gate array (FPGA)-based electronic designs. The resulting cochlea-neuron interaction circuit, when implemented on a Xilinx Virtex-II FPGA, demonstrates 1.33x speed-up and supports up to 87 parallel cochlear modules, while maintaining high signal fidelity and neural activation accuracy. Simulation and hardware validation confirm that the proposed system provides a scalable, low-resource solution suitable for emerging industrial biosensing systems, smart auditory devices, and embedded neural interfaces. The methodology contributes to advancing the real-time digital implementation of biologically inspired systems in industrial and biomedical electronics.

Myocardial Infarction (MI) commonly referred to as a heart attack, results from the abrupt obstruction of blood supply to a section of the heart muscle, leading to the deterioration or death of the affected tissue due to a lack of oxygen. MI, poses a significant public health concern worldwide, particularly affecting the citizens of the Chittagong Metropolitan Area. The challenges lie in both prevention and treatment, as the emergence of MI has inflicted considerable suffering among residents. Early warning systems are crucial for managing epidemics promptly, especially given the escalating disease burden in older populations and the complexities of assessing present and future demands. The primary objective of this study is to forecast MI incidence early using a deep learning model, predicting the prevalence of heart attacks in patients. Our approach involves a novel dataset collected from daily heart attack incidence Time Series Patient Data spanning January 1, 2020, to December 31, 2021, in the Chittagong Metropolitan Area. Initially, we applied various advanced models, including Autoregressive Integrated Moving Average (ARIMA), Error-Trend-Seasonal (ETS), Trigonometric seasonality, Box-Cox transformation, ARMA errors, Trend and Seasonal (TBATS), and Long Short Time Memory (LSTM). To enhance prediction accuracy, we propose a novel Myocardial Sequence Classification (MSC)-LSTM method tailored to forecast heart attack occurrences in patients using the newly collected data from the Chittagong Metropolitan Area. Comprehensive results comparisons reveal that the novel MSC-LSTM model outperforms other applied models in terms of performance, achieving a minimum Mean Percentage Error (MPE) score of 1.6477. This research aids in predicting the likely future course of heart attack occurrences, facilitating the development of thorough plans for future preventive measures. The forecasting of MI occurrences contributes to effective resource allocation, capacity planning, policy creation, budgeting, public awareness, research identification, quality improvement, and disaster preparedness.

Within the heart’s ventricle walls are where Purkinje fibers (PFs) are located. They are essential for maintaining a steady cardiac beat because they allow the heart’s conduction system to create synchronized contractions of its ventricles. To replicate or treat some of this organ’s ailments and deficits, hardware representation of the various heart sections is necessary. Models that include a series of differential equations can be used to characterize the performance of Purkinje Fibers (PFs) in Cardiac. To develop hardware that mimics the performance of PF of Cardiac system, the Noble model is updated in this study. Due to the requirement for using units like multipliers, the original model contains non-linear components that are slow and expensive in terms of hardware resources. The main novelty of revised model incorporates non-linear components that have been converted into two base-2 terms accompanied by additional factors. These calculations are executed through efficient and economical digital hardware, including logical shift operations, additions, and subtractions, enabling high-speed processing. To validate the precision and practicality of the suggested model, a digital platform, Virtex-7 FPGA, is employed. The findings demonstrate the suggested model’s simplicity of implementation on this board and its capability to generate various PF output patterns at a maximum frequency of 381.42 MHz. The proposed digital circuit can be applied in application-based fields according to high-speed, low-cost, and accurate design. Based on the high-switching speed of neural data transferring in the human brain, other organs that are realized on hardware platforms need to be designed in high-frequency (speed-up) for adaptation with the brain. Due to removing the non-linear terms, the modified model works 1.95 times faster than the original one and saves the FPGA resources up to 35%. In case of real-world applications, an efficient low-cost hardware design for high-frequency, accurate replication of cardiac Purkinje fiber cells has the potential to revolutionize the development of medical devices for diagnosing and treating cardiac conditions. The speed and cost-effectiveness of design make it particularly promising for the creation of advanced cardiac simulation and treatment systems. The high-speed, low-cost, and accurate hardware design has broader implications for biomedical engineering beyond the cardiac system. The ability to efficiently replicate the functions of biological systems using hardware could open doors for developing high-speed, low-cost, and accurate hardware models for simulating and understanding other biological processes, such as neural data processing in the human brain or the function of other vital organs.