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


Faris S. Alghareb has a total of 17 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    An Optimized Convolutional Neural Network Framework for Cancer Classification Using RNA-Seq Transformation

    by Ahmed A. Mohammed, Faris S. Alghareb, Mohammad J. M. Zedan, Mohammed A. M. Abdullah, Sarmad F. Ismael, Ersin Elbasi, Mohd Asyraf Zulkifley

    Inspired by the hierarchical, receptive-field-based processing of biological vision and by synaptic pruning in the developing brain, this work proposes a bio-inspired optimized convolutional neural network framework for multi-class cancer classification from RNA-Seq gene expression data. Although current deep learning models achieve strong predictive performance, existing approaches typically rely either on large pre-trained networks that cannot be deployed on resource-constrained devices or on image representations that normalize gene expression values to the 8-bit range (0–255), discarding the original biological dynamic range. A variance-driven feature selection step reduces 20,531 gene profiles to 16,384 high-information features, which are then reshaped into 128×128 images while preserving the original log2-normalized expression range (approximately 0–20.78) without rescaling to that 8-bit range. The network architecture is derived from a parametric mathematical model that enables analytical control over parameter count and FLOPs, resulting in a compact network with only 0.16 million parameters, a model size of 1.29 MB, and 0.0372 GFLOPs. On the TCGA PANCAN dataset (five cancer classes), the proposed baseline achieves 99.38% accuracy, matching or exceeding twelve state-of-the-art pre-trained architectures. For real-world deployment, the network was further optimized through two pathways: (A) direct TensorFlow Lite quantization in FP32, FP16, and INT8, and (B) structured channel pruning, analogous to synaptic pruning, followed by TFLite conversion. All variants retained the baseline 99.38% accuracy, with the Pruned TFLite FP16 variant reducing computational cost by 23.4% to 0.0285 GFLOPs, model size by 80.6%, and latency by 75.4%, while achieving the highest throughput of 1256 images/s at only 0.25 MB.

  • Open Access

    A Defensive Distillation Approach Towards Adversarial Attacks Based on Gaussian Blurring

    by Ahmed Ahmed, Mohammed Al-Nuaimi, Faris Alghareb, Ahmad Abdulfattah

    Deep Neural Networks (DNNs) have attained remarkable prediction outcomes in image classification tasks, leading to significant progress in computer vision applications. However, the presence of adversarial examples has emerged as a critical challenge to the robustness and efficiency of deep learning-based image classifiers. Adversarial examples are specially designed perturbations applied to input images to deceive the models into generating inaccurate predictions while appearing indistinguishable to human observers. In this paper, we present a defense mechanism, namely Defensive Distillation with Gaussian Blurring (DDGB), that improves the robustness of deep learning models towards adversarial attacks. First, two models, a teacher and a student model, were utilized to train and validate the presented approach. The teacher model is trained and then leveraged to determine softened probabilities, which are later utilized to train the student model. Second, a feature-squeezing technique based on Gaussian blurring is applied to the adversarial examples generated from the distilled student model as a form of defense mechanism to make the adversarial perturbations less effective. The obtained findings demonstrate that the proposed approach is effective in improving the performance, achieving classification accuracies of 87.61% and 87.48% using the Fast Gradient Sign Method (FGSM) and Basic Iterative Method (BIM) attacks, respectively, based on the CIFAR-10 dataset. In summary, the presented approach achieves a 70.66% reduction in computations for the student model, allowing the model to be deployed on devices with limited resources and provide improved prediction accuracy towards adversarial attacks.

  • Open Access

    Encoding-Based Machine Learning Approach for Health Status Classification and Remote Monitoring of Cardiac Patients

    by Sohaib R. Awad, Faris S. Alghareb

    Remote monitoring of a patient’s vital activities has become increasingly important in dealing with various medical applications. In particular, machine learning (ML) techniques have been extensively utilized to analyze electrocardiogram (ECG) signals in cardiac patients to classify heart health status. This trend is largely driven by the growing interest in computer-aided diagnosis based on ML algorithms. However, there has been inadequate investigation into the impact of risk factors on heart health, which hinders the ability to identify heart-related issues and predict the conditions of cardiac patients. In this context, developing a GUI-based classification approach can significantly facilitate online monitoring and provide real-time warnings by predicting potential complications. In this paper, a general framework structure for medical real-time monitoring systems is proposed for modeling the vital signs of cardiac patients in order to predict the patient’s status. The proposed approach analyzes AI-driven interventions to provide a more accurate cardiac diagnosis and real-time monitoring system. To further demonstrate the validity of the presented approach, we employ it in a LabVIEW-based remote tracking system to predict three healthcare statuses (stable, unstable non-critical, and unstable critical). The developed monitoring system receives various information about patients’ vital signs, and then it leverages a novel encoding-based machine learning algorithm to pre-process, analyze, and classify patient status. The developed ANN classifier and proposed encoding-based ML model are compared to other conventional ML-based models, such as Naive Bayes, SVM, and KNN for model accuracy evaluation. The obtained outcomes demonstrate the efficacy of the presented ANN and encoding-based ML approaches by achieving an accuracy of 98.4% and 98.8% for the developed ANN classifier and the proposed encoding-based technique, respectively, whereas Naive Bayes and quadratic SVM algorithms realize 94.8% and 96%, respectively. In short, this study aims to explore how ML algorithms can enhance diagnostic accuracy, improve real-time monitoring, and optimize treatment outcomes. Meanwhile, the proposed tracking system outperforms most existing monitoring systems by offering high classification accuracy of the heart health status and a user-friendly interactive interface. Therefore, it can potentially be utilized to improve the performance of remote healthcare monitoring for cardiac patients.


  • Open Access

    Multitask Learning-Based Pipeline-Parallel Computation Offloading Architecture for Deep Face Analysis

    by Faris S. Alghareb, Balqees Talal Hasan

    Deep Neural Networks (DNNs) have been widely adopted in several advanced artificial intelligence applications due to their competitive accuracy to the human brain. Nevertheless, the superior accuracy of a DNN is achieved at the expense of intensive computations and storage complexity, requiring custom expandable hardware, i.e., graphics processing units (GPUs). Interestingly, leveraging the synergy of parallelism and edge computing can significantly improve CPU-based hardware platforms. Therefore, this manuscript explores levels of parallelism techniques along with edge computation offloading to develop an innovative hardware platform that improves the efficacy of deep learning computing architectures. Furthermore, the multitask learning (MTL) approach is employed to construct a parallel multi-task classification network. These tasks include face detection and recognition, age estimation, gender recognition, smile detection, and hair color and style classification. Additionally, both pipeline and parallel processing techniques are utilized to expedite complicated computations, boosting the overall performance of the presented deep face analysis architecture. A computation offloading approach, on the other hand, is leveraged to distribute computation-intensive tasks to the server edge, whereas lightweight computations are offloaded to edge devices, i.e., Raspberry Pi 4. To train the proposed deep face analysis network architecture, two custom datasets (HDDB and FRAED) were created for head detection and face-age recognition. Extensive experimental results demonstrate the efficacy of the proposed pipeline-parallel architecture in terms of execution time. It requires 8.2 s to provide detailed face detection and analysis for an individual and 23.59 s for an inference containing 10 individuals. Moreover, a speedup of 62.48% is achieved compared to the sequential-based edge computing architecture. Meanwhile, 25.96% speed performance acceleration is realized when implementing the proposed pipeline-parallel architecture only on the server edge compared to the sever sequential implementation. Considering classification efficiency, the proposed classification modules achieve an accuracy of 88.55% for hair color and style classification and a remarkable prediction outcome of 100% for face recognition and age estimation. To summarize, the proposed approach can assist in reducing the required execution time and memory capacity by processing all facial tasks simultaneously on a single deep neural network rather than building a CNN model for each task. Therefore, the presented pipeline-parallel architecture can be a cost-effective framework for real-time computer vision applications implemented on resource-limited devices.


  • Non-Volatile Spintronic Flip-Flop Design for Energy-Efficient SEU and DNU Resilience

    by Faris S. Alghareb, Ramtin Zand, Ronald F. Demara

    In this paper, low-energy radiation-hardening approaches are proposed to develop non-volatile (NV) flip-flop (FF) circuits using spintronic devices. In particular, spin Hall effect magnetic tunnel junctions are used to design a radiation-hardened NV-latch that is proposed to be utilized as a shadow latch to maintain the data during the standby mode when the circuit is power-gated. Moreover, soft-error resilient complementary metal-oxide-semiconductor-based latching circuits are designed to be leveraged as master and slave latches in the NVFF structure. The proposed hardening techniques are based on using feedback loops and clock-gating Muller C-elements, as well as increasing the charge capacity of the vulnerable nodes. The circuit simulations indicate that the proposed single-event upset and double node upset resilient latching circuits can achieve at least 81% and 24% power-delay product improvement, respectively, while incurring comparable area overhead compared to the previous energy-efficient radiation-hardened latch designs. Finally, the proposed latching circuits are combined to develop four radiation-hardened NVFF designs. The results obtained show that, using the proposed NV latching circuit as a shadow latch can result in two orders of magnitude reduction in energy consumption compared to the FF circuits with an NV master latch. In addition, the proposed latches achieve favorable tradeoffs in terms of minimized performance overheads and maximized robustness (100%) of soft fault coverage to single and double upsets. Thus, the proposed NVFFs can be employed within logic datapaths to ensure data integrity as a potential mainstream solution for aerospace and avionic nanoelectronics.


  • High-Performance Double Node Upset-Tolerant Non-Volatile Flip-Flop Design

    by Faris S. Alghareb, Ramtin Zand, Ronald F. DeMara

    Emerging spin-based devices are introduced as an intriguing candidate to alleviate leakage currents and continue the scalability of CMOS technology. However, their immunity to radiation-induced transient faults needs to be adequately addressed. In this work, a radiation-immune hybrid Spin Transfer Torque Magnetic Tunnel Junction (STT-MTJ)/CMOS flip-flop is designed and evaluated for nonvolatile applications. The proposed nonvolatile flip-flop circuit achieves attractive features, such as low standby power dissipation (21% less than CMOS-based design), high computing performance, and superior soft-error resilience (concurrently can tolerate DNU) to potentially become as a mainstream solution for the aerospace and avionic nanoelectronics.


  • Designing and Evaluating Redundancy-Based Soft-Error Masking on a Continuum of Energy versus Robustness

    by Faris S. Alghareb, Rizwan A. Ashraf, Ronald F. DeMara

    Near-threshold computing is an effective strategy to reduce the power dissipation of deeply-scaled CMOS logic circuits. However, near-threshold strategies exacerbate the impact of delay variations on device performance and increase the susceptibility to soft errors due to narrow voltage margins. The objective of this work is to develop and assess design approaches that leverage tradeoffs between performance and the resilience of fault masking coverage for various soft error mitigation techniques. The primary insights identify those redundancy-based hardening techniques that can deliver increased benefits in terms of the Fault Coverage Energy Ratio (FCER) for the leveraged tradeoffs within iso-energy constraints at Near-Threshold Voltage (NTV). Simulation results demonstrate that temporal redundancy approaches offer favorable tradeoffs in terms of FCER. Meanwhile, it is shown that a hybrid redundancy approach can be used to protect a low-power system to maintain throughput while tolerating soft errors. We demonstrate how the FCER metric can be used as an optimization parameter to guide circuit synthesis to meet performance and robustness goals. Finally, the impact of design diversity on spatial and hybrid redundancy at NTV is assessed in terms of FCER and delay variation, to form overall recommendations regarding soft-error mitigation at NTV.


  • Soft Error Effect Tolerant Temporal Self-Voting Checkers: Energy vs. Resilience Tradeoffs

    by Faris S. Alghareb, Mingjie Lin, Ronald F. DeMara

    Achieving high reliability against transient faults poses significant challenges due to the trends of technology and voltage scaling. Thus, numerous soft error mitigation techniques have been proposed for masking Soft Error Rate (SER) in logic circuits. However, most soft error suppression approaches have significant overheads in terms of area, power consumption, and speed performance degradation. Herein, we propose two circuit-level techniques, namely Temporal Self-Voting Logic (TSVL) and Hybrid Spatial and Temporal Redundancy Double-Error Correction (HSTR-DEC), to prevent the effects of soft errors in logic circuits, occurring due to Single Event Upset (SEU) or Single Event Transient (SET). TSVL and HSTR-DEC circuits can be utilized to improve the reliability of a logic path with minimal impact on circuit delay while achieving a complete and cost-effective SEU handling as compared to traditional spatial or temporal redundancy approach. The primary contribution of the TSVL approach is that it eliminates error masking from the critical datapath, thus, area and energy overheads are significantly reduced. A transient gate-level fault injection and analysis is used to evaluate the capability of soft errors suppression of the proposed hardening approach. Experimental results indicate that TSVL can cover soft errors, on average, roughly by 99% while realizing an amelioration of 22.02% and 2.15% for area and speed degradation as compared to the previous Self-Voting DMR approach. Meanwhile, HSTRDEC approach realizes a complete coverage for single and double SEUs while incurring comparable area and delay overheads as compared to the prior hybrid redundancy approach.