Safwan Mahmood Al-Selwi
@utp.edu.my
Department of Computer and Information Sciences
Universiti Teknologi Petronas
Safwan Al-Selwi received a BEng in Software Engineering from Taiz University (Yemen), and an MSc in Computer Applications from Bangalore University (India). He is currently a Research and Teaching Assistant at CISD-UTP (Malaysia). He has over eight years of experience in academic institutions and industry. His industry experience includes Android and web development. His research interests include artificial intelligence, machine learning, metaheuristic algorithms, IoT, and optimization.
EDUCATION
MSc in Computer Applications from Bangalore University (India).
BEng in Software Engineering from Taiz University (Yemen).
RESEARCH, TEACHING, or OTHER INTERESTS
Artificial Intelligence, Computer Science
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
Jameel Shehu Yalli, Mohd Hilmi Hasan, Low Tan Jung, and Safwan Mahmood Al-Selwi
Elsevier BV
Shahab Ul Hassan, Said Jadid Abdulkadir, M Soperi Mohd Zahid, and Safwan Mahmood Al-Selwi
Elsevier BV
Jameel S. Yalli, Mohd H. Hasan, Low T. Jung, Abdulrasheed I. Yerima, Dahiru A. Aliyu, Umar D. Maiwada, Safwan M. Al-Selwi, and Mujeeb U. R. Shaikh
Institute of Electrical and Electronics Engineers (IEEE)
Safwan Mahmood Al-Selwi, Mohd Fadzil Hassan, Said Jadid Abdulkadir, Mohammed Gamal Ragab, Alawi Alqushaibi, and Ebrahim Hamid Sumiea
Elsevier BV
Mehak Mushtaq Malik, Abdul Muiz Fayyaz, Mussarat Yasmin, Said Jadid Abdulkadir, Safwan Mahmood Al-Selwi, Mudassar Raza, and Sadia Waheed
Springer Science and Business Media LLC
Safwan Mahmood Al-Selwi, Mohd Fadzil Hassan, Said Jadid Abdulkadir, Amgad Muneer, Ebrahim Hamid Sumiea, Alawi Alqushaibi, and Mohammed Gamal Ragab
Springer Science and Business Media LLC
Ebrahim Hamid Sumiea, Said Jadid Abdulkadir, Hitham Seddig Alhussian, Safwan Mahmood Al-Selwi, Alawi Alqushaibi, Mohammed Gamal Ragab, and Suliman Mohamed Fati
Elsevier BV
Mohammed Gamal Ragab, Kamaluddeen Usman Danyaro, Said Jadid Abdulkadir, Safwan Mahmood Al-Selwi, Ebrahim Hamid Sumiea, and Alawi Alqushaibi
IEEE
Brain tumors present a significant challenge in medical diagnostics due to their complex nature and the critical need for precise detection. This paper introduces a novel approach to brain tumor detection using instance segmentation with the advanced YOLOv8 model in medical imaging. We first outline the limitations of current imaging techniques in accurately identifying brain tumors. Then, we detail our methodology, which includes a comprehensive data preprocessing strategy tailored for medical imaging, followed by an in-depth explanation of the customized YOLOv8 architecture used for this study. Our training and validation process is thoroughly explained, emphasizing the adaptations made for handling medical datasets. The results section demonstrates the model's efficacy through various metrics, including precision with 0.99%, recall 0.90%, and mAP Analysis 0.95%, showcasing significant improvements over existing methods. The paper concludes by emphasizing the potential impact of applying YOLOv8 in medical imaging for brain tumor detection, suggesting a substantial step forward in oncological diagnostics and patient care.
Shahab Ul Hassan, Said Jadid Abdulkadir, Mohd Soper Mohd Zahid, Abdul Muiz Fayyaz, Safwan Mahmood Al-Selwi, and Ebrahim Hamid Sumiea
IEEE
Cardiac arrhythmia is one of the most critical cardiovascular diseases that cause millions of fatalities every year. Early detection of the disease by analyzing the electrocardiogram signals of patients has the potential to save many lives. Deep learning prediction models have gained a lot of attention for arrhythmia prediction. Among them, convolutional neural network (CNN) and long short-term memory (LSTM) techniques are widely used. These models have been recently combined into the CNN-LSTM model to achieve high accuracy and efficiency in arrhythmia prediction. However, there is a lack of studies analyzing the performance of CNN and CNN-LSTM techniques to find the optimal values for key parameters such as the number of filters, kernel size, and layers. This article determines optimized CNN (OCNN) and optimized CNN-LSTM (OCNN-LSTM) models for MIT-BIH arrhythmia datasets by analyzing the models' performance by varying several key parameters. Performance is measured in terms of accuracy, AUC, recall, precision, and specificity. Hence, the aim is to find the optimum values for the key parameters required to develop deep learning models for the MIT-BIH arrhythmia dataset. The finest outcomes attained for the OCNN and OCNN-LSTM models were 99.9% and 98.1% AUC, 99.0%, 96.1% accuracy, 98.5% and 93.3% recall, 98.1%, and 93.5% precision, and 99.2% and 97.2% specificity, respectively.
Mohammed Gamal Ragab, Said Jadid Abdulkadir, Amgad Muneer, Alawi Alqushaibi, Ebrahim Hamid Sumiea, Rizwan Qureshi, Safwan Mahmood Al-Selwi, and Hitham Alhussian
Institute of Electrical and Electronics Engineers (IEEE)
YOLO (You Only Look Once) is an extensively utilized object detection algorithm that has found applications in various medical object detection tasks. This has been accompanied by the emergence of numerous novel variants in recent years, such as YOLOv7 and YOLOv8. This study encompasses a systematic exploration of the PubMed database to identify peer-reviewed articles published between 2018 and 2023. The search procedure found 124 relevant studies that employed YOLO for diverse tasks including lesion detection, skin lesion classification, retinal abnormality identification, cardiac abnormality detection, brain tumor segmentation, and personal protective equipment detection. The findings demonstrated the effectiveness of YOLO in outperforming alternative existing methods for these tasks. However, the review also unveiled certain limitations, such as well-balanced and annotated datasets, and the high computational demands. To conclude, the review highlights the identified research gaps and proposes future directions for leveraging the potential of YOLO for medical object detection.
Safwan Mahmood Al-Selwi, Mohd Fadzil Hassan, Said Jadid Abdulkadir, and Amgad Muneer
Akademia Baru Publishing
Recurrent neural networks (RNNs) are an excellent fit for regression problems where sequential data are the norm since their recurrent internal structure can analyse and process data for long. However, RNNs are prone to the phenomenal vanishing gradient problem (VGP) that causes the network to stop learning and generate poor prediction accuracy, especially in long-term dependencies. Originally, gated units such as long short-term memory (LSTM) and gated recurrent unit (GRU) were created to address this problem. However, VGP was and still is an unsolved problem, even in gated units. This problem occurs during the backpropagation process when the recurrent network weights tend to vanishingly reduce and hinder the network from learning the correlation between temporally distant events (long-term dependencies), that results in slow or no network convergence. This study aims to provide an empirical analysis of LSTM networks with an emphasis on inefficiency in long-term dependencies convergence because of VGP. Case studies on NASA’s turbofan engine degradation are examined and empirically analysed.
Ebrahim Hamid Sumiea, Said Jadid AbdulKadir, Hitham Alhussian, Safwan Mahmood Al-Selwi, Mohammed Gamal Ragab, and Alawi Alqushaibi
IEEE
The optimization of continuous action control tasks is a crucial step in deep reinforcement learning (DRL) applications. The goal is to identify optimal actions through the accumulation of experience in continuous action control tasks. This process can be achieved through DRL, which trains agents to develop a policy that maximizes the cumulative rewards gained from decision-making in dynamic environments. Balancing exploration and exploitation is a crucial challenge in acquiring this policy. To address the exploration-exploitation trade-off, the Exploration Decay Policy (EDP) implements a dynamic exploration noise strategy that adapts to the current training progress, enabling efficient exploration in the initial phases while gradually reducing exploration to focus on exploitation as training progresses. However, the fluctuating training stability across episodes in dynamic environments poses a challenge for exploitation policies to adapt accordingly. In this paper, we propose EDP to address exploration–exploitation trade-off dilemma. The objective is to dynamically modulate the noise scale, gradually decreasing it during periods of high training stability to promote exploration, while reducing it to maintain exploitation during periods of low training stability. The study introduces the EDP-DDPG method, enhancing continuous control tasks in Box2D environments. EDP-DDPG outperforms the standard DDPG by achieving higher rewards and quicker convergence. Its success stems from dynamically adjusting exploration noise every 25 episodes, balancing exploration and exploitation. This adaptive approach, reducing noise by 10% every 25 episodes, evolves from random to strategic limb movements, optimizing policy exploitation and adaptability in dynamic settings.
Alawi Alqushaibi, Mohammed Gamal Ragab, Ebrahim Hamid Sumiea, Safwan Mahmood Al-Selwi, Mohd Hilmi Hasan, Said Jadid Abdulkadir, Hitham Alhussian, and Mahmoud Mustafa Al-Asbahi
IEEE
Detection of heart disease is critical, as it is a common and serious health condition. Predicting its occurrence is vital in the healthcare domain, and machine learning techniques are increasingly being utilized to accurately identify patients at risk. To evaluate how well machine learning models can effectively predict heart disease, we used a pre-existing dataset from the UCI repository’s Cleveland database. Though the dataset originally contained 303 records and 76 attributes, we followed the literature by narrowing the focus to 14 relevant attributes in order to improve the precision of various algorithmic approaches. Our study employed five machine learning classifiers, namely KNearest Neighbors (KNN), CatBoost Classifier (CBC), Decision Tree Classifier (DT), XGBClassifier (XGBC), and Support Vector Classifier (SVM). We evaluated their performance using different metrics like accuracy, precision, AUC, and F1 score on the heart disease dataset. We found that the CatBoost Classifier had the best performance, achieving high accuracy and AUC with 0.88% and 0.94% values, respectively. This suggests that machine learning algorithms could aid in identifying heart disease at an early stage and have significant implications for improving patient outcomes in healthcare.
Ebrahim Hamid Hasan Sumiea, Said Jadid Abdulkadir, Mohammed Gamal Ragab, Safwan Mahmood Al-Selwi, Suliamn Mohamed Fati, Alawi AlQushaibi, and Hitham Alhussian
Institute of Electrical and Electronics Engineers (IEEE)
Deep Reinforcement Learning (DRL) allows agents to make decisions in a specific environment based on a reward function, without prior knowledge. Adapting hyperparameters significantly impacts the learning process and time. Precise estimation of hyperparameters during DRL training poses a major challenge. To tackle this problem, this study utilizes Grey Wolf Optimization (GWO), a metaheuristic algorithm, to optimize the hyperparameters of the Deep Deterministic Policy Gradient (DDPG) algorithm for achieving optimal control strategy in two simulated Gymnasium environments provided by OpenAI. The ability to adapt hyperparameters accurately contributes to faster convergence and enhanced learning, ultimately leading to more efficient control strategies. The proposed DDPG-GWO algorithm is evaluated in the 2DRobot and MountainCarContinuous simulation environments, chosen for their ease of implementation. Our experimental results reveal that optimizing the hyperparameters of the DDPG using the GWO algorithm in the Gymnasium environments maximizes the total rewards during testing episodes while ensuring the stability of the learning policy. This is evident in comparing our proposed DDPG-GWO agent with optimized hyperparameters and the original DDPG. In the 2DRobot environment, the original DDPG had rewards ranging from −150 to −50, whereas, in the proposed DDPG-GWO, they ranged from −100 to 100 with a running average between 1 and 800 across 892 episodes. In the MountainCarContinuous environment, the original DDPG struggled with negative rewards, while the proposed DDPG-GWO achieved rewards between 20 and 80 over 218 episodes with a total of 490 timesteps.