arXiv subjects

Ziya Uddin

Publications and source records attributed to Ziya Uddin.

8 recordsLinked to original sources

arXiv2026

Physics Informed Optimal Homotopy Analysis Method (PI-OHAM): A Hybrid Analytical Computational Framework for Solving nonlinear Differential Equations

We present the Physics-Informed Optimal Homotopy Analysis Method (PI-OHAM) for solving nonlinear differential equations. PI-OHAM, based on classical HAM, employs a physics-informed residual loss to optimize convergence-control parameters systematically by combining data, boundary conditions, and governing equations in the manner similar to Physics Informed Neural Networks (PINNs). The combination of the flexibility of PINNs and the analytical transparency of HAM provides the approach with high numerical stability, rapid convergence, and high consistency with traditional numerical solutions. PI-OHAM has superior accuracy-time trade-offs and faster and more accurate convergence than standard HAM and PINNs when applied to the Blasius boundary-layer problem. It is also very close to numerical standards available in the literature. PI-OHAM ensures analytical transparency and interpretability by series-based solutions, unlike purely data-driven or data-free PINNs. Significant contributions are a conceptual bridge between decades of homotopy-based analysis and modern physics-inspired methods, and a numerically aided but analytically interpretable solver of nonlinear differential equations. PI-OHAM appears as a computationally efficient, accurate and understandable alternative to nonlinear fluid flow, heat transfer and other industrial problems in cases where robustness and interpretability are important.

cs.CE↗

arXiv2025

p and hp Spectral Element Methods for Elliptic Boundary Layer Problems

In this article, we propose p and hp least-squares spectral element methods for one-dimensional elliptic boundary layer problems. Stability estimates are derived and we design numerical schemes based on minimizing the residuals in the sense of least-squares in appropriate Sobolev norms. We prove parameter robust uniform error estimates i.e. error in the approximation is independent of the boundary layer parameter. For the p-version we prove a robust uniform convergence rate of O(sqrt(log W)/W), where W denotes the polynomial order used in approximation and for the hp-version the convergence rate is shown to be O(e^(-W/logW)). Numerical results are presented for a number of model elliptic boundary layer problems confirming the theoretical estimates and uniform convergence results for the p and hp versions.

math.NA↗

arXiv2025

Hybrid Deep Learning Framework for Classification of Kidney CT Images: Diagnosis of Stones, Cysts, and Tumors

Medical image classification is a vital research area that utilizes advanced computational techniques to improve disease diagnosis and treatment planning. Deep learning models, especially Convolutional Neural Networks (CNNs), have transformed this field by providing automated and precise analysis of complex medical images. This study introduces a hybrid deep learning model that integrates a pre-trained ResNet101 with a custom CNN to classify kidney CT images into four categories: normal, stone, cyst, and tumor. The proposed model leverages feature fusion to enhance classification accuracy, achieving 99.73% training accuracy and 100% testing accuracy. Using a dataset of 12,446 CT images and advanced feature mapping techniques, the hybrid CNN model outperforms standalone ResNet101. This architecture delivers a robust and efficient solution for automated kidney disease diagnosis, providing improved precision, recall, and reduced testing time, making it highly suitable for clinical applications.

eess.IV↗

arXiv2024

Exploring Structural Dynamics in Retracted and Non-Retracted Author's Collaboration Networks: A Quantitative Analysis

Retractions undermine the reliability of scientific literature and the foundation of future research. Analyzing collaboration networks in retracted papers can identify risk factors, such as recurring co-authors or institutions. This study compared the network structures of retracted and non-retracted papers, using data from Retraction Watch and Scopus for 30 authors with significant retractions. Collaboration networks were constructed, and network properties analyzed. Retracted networks showed hierarchical and centralized structures, while non-retracted networks exhibited distributed collaboration with stronger clustering and connectivity. Statistical tests, including $t$-tests and Cohen's $d$, revealed significant differences in metrics like Degree Centrality and Weighted Degree, highlighting distinct structural dynamics. These insights into retraction-prone collaborations can guide policies to improve research integrity.

cs.IR↗

arXiv2024

Exploring Physics-Informed Neural Networks: From Fundamentals to Applications in Complex Systems

Physics-informed neural networks (PINNs) have emerged as a versatile and widely applicable concept across various science and engineering domains over the past decade. This article offers a comprehensive overview of the fundamentals of PINNs, tracing their evolution, modifications, and various variants. It explores the impact of different parameters on PINNs and the optimization algorithms involved. The review also delves into the theoretical advancements related to the convergence, consistency, and stability of numerical solutions using PINNs, while highlighting the current state of the art. Given their ability to address equations involving complex physics, the article discusses various applications of PINNs, with a particular focus on their utility in computational fluid dynamics problems. Additionally, it identifies current gaps in the research and outlines future directions for the continued development of PINNs.

cs.CE↗

arXiv2024

SIR-RL: Reinforcement Learning for Optimized Policy Control during Epidemiological Outbreaks in Emerging Market and Developing Economies

The outbreak of COVID-19 has highlighted the intricate interplay between public health and economic stability on a global scale. This study proposes a novel reinforcement learning framework designed to optimize health and economic outcomes during pandemics. The framework leverages the SIR model, integrating both lockdown measures (via a stringency index) and vaccination strategies to simulate disease dynamics. The stringency index, indicative of the severity of lockdown measures, influences both the spread of the disease and the economic health of a country. Developing nations, which bear a disproportionate economic burden under stringent lockdowns, are the primary focus of our study. By implementing reinforcement learning, we aim to optimize governmental responses and strike a balance between the competing costs associated with public health and economic stability. This approach also enhances transparency in governmental decision-making by establishing a well-defined reward function for the reinforcement learning agent. In essence, this study introduces an innovative and ethical strategy to navigate the challenge of balancing public health and economic stability amidst infectious disease outbreaks.

cs.LG↗

arXiv2024

Unraveling Retraction Dynamics in COVID-19 Research: Patterns, Reasons, and Implications

Amid the COVID-19 pandemic, while the world sought solutions, some scholars exploited the situation for personal gains through deceptive studies and manipulated data. This paper presents the extent of 400 retracted COVID-19 papers listed by the Retraction Watch database until February 2024. The primary purpose of the research was to analyze journal quality and retraction trends. For all stakeholders involved, such as editors, relevant researchers, and policymakers, evaluating the journal's quality is crucial information since it could help them effectively stop such incidents and their negative effects in the future. The present research results imply that one-fourth of publications were retracted within the first month of their publication, followed by an additional 6\% within six months of publication. One-third of the retractions originated from Q1 journals, with another significant portion coming from Q2 (29.8). A notable percentage of the retracted papers (23.2\%) lacked publishing impact, signifying their publication as conference papers or in journals not indexed by Scopus. An examination of the retraction reasons reveals that one-fourth of retractions were due to numerous causes, mostly in Q2 journals, and another quarter were due to data problems, with the majority happening in Q1 publications. Elsevier retracted 31 of the papers, with the majority published in Q1, followed by Springer (11.5), predominantly in Q2. Retracted papers were mainly associated with the USA, China, and India. In the USA, retractions were primarily from Q1 journals followed by no-impact publications; in China, it was Q1 followed by Q2, and in India, it was Q2 followed by no-impact publications. The study also examined author contributions, revealing that 69.3 were male contributors, with females (30.7) mainly holding middle author positions.

cs.DL↗

arXiv2023

Measuring the continuous research impact of a researcher: The Kz index

The ongoing discussion regarding the utilization of individual research performance for academic hiring, funding allocation, and resource distribution has prompted the need for improved metrics. While traditional measures such as total publications, citations count, and the h-index provide a general overview of research impact, they fall short of capturing the continuous contribution of researchers over time. To address this limitation, we propose the implementation of the Kz index, which takes into account both publication impact and age. In this study, we calculated Kz scores for 376 research profiles. Kz reveals that the researchers with the same h-index can exhibit different Kz scores, and vice versa. Furthermore, we observed instances where researchers with lower citation counts obtained higher Kz scores, and vice versa. Interestingly, the Kz metric follows a log-normal distribution. It highlights its potential as a valuable tool for ranking researchers and facilitating informed decision-making processes. By measuring the continuous research impact, we enable fair evaluations, enhance decision-making processes, and provide focused career advancement support and funding opportunities.

cs.DL↗