arXiv subjects

Muhammad Nawaz

Publications and source records attributed to Muhammad Nawaz.

7 recordsLinked to original sources

arXiv2026

Impact of a Cold Dark Matter Halo on Magnetic Reconnection and Energy Extraction from Kerr-like Black Holes

Recently, Comisso and Asenjo introduced a new energy extraction mechanism based on magnetic reconnection. In this paper, we investigate the power and efficiency of magnetic reconnection energy extraction in a rotating black hole surrounded by a cold dark matter halo. We first examine the properties of the underlying spacetime and its physical quantities, including the event horizon, photon sphere, and ergosphere. We then analyze the allowed energy extraction region and the corresponding energy extraction efficiency for circular orbits. Our results show that energy extraction is feasible for black holes with a spin parameter.

gr-qc↗

arXiv2026

Magnetic Reconnection and Energy Extraction from a Rotating Black Hole in the Einstein-AdS SU(N)-Nonlinear Sigma Model

In the present study, we analyze the power and efficiency of energy extraction via magnetic reconnection in the rotating Einstein-AdS-SU($N$)-NLSM black hole, both in the circular-orbit regime and in the plunging region. Initially, we define the background properties of this spacetime, and then analyze the physical quantities such as the size of the ergoregion, the event horizon, and the boundaries of the ergosphere. We analyze the magnetic reconnection process within circular orbits. We plot energy-extraction parameter diagrams and analyze the power and efficiency of energy extraction. Our results indicate that energy extraction remains feasible even at a spin parameter as low as $0.7$, significantly below previously reported thresholds, and the extracted power can exceed that of the Blandford-Znajek mechanism with specific constraints. The coupling constant $K$, AdS radius $l$ and the flavors number $N$, collectively participate in lowering the spin threshold for energy extraction. Consequently, we further investigate the permissible energy extraction region for the energy extraction mechanism in the plunging region, as well as the corresponding power output and efficiency. We observe that the energy extraction is possible even at a spin as low as $0.2$. Importantly, the parameters $N,~K$ and $l$ mainly contribute to lowering the energy extraction spin threshold. This behavior is similar to that in circular orbits. Finally, comparing the plunging region with the circular orbits, we observe that the energy extraction power in the plunging region is higher than in the circular orbits.

gr-qc↗

arXiv2026

Magnetic Reconnection and Energy Extraction from a Rotating Black Hole in a Four-dimensional Einstein-Gauss-Bonnet Gravity

Recently, Comisso and Asenjo proposed a new mechanism for energy extraction based on magnetic reconnection of plasma within the ergosphere. In this paper, we analyze the power and efficiency of energy extraction through magnetic reconnection in a rotating four-dimensional Einstein-Gauss-Bonnet gravity black hole. Firstly, we analyze the background properties of this spacetime and its physical quantities, including the event horizon, the boundary of the ergosphere, and the size of the ergosphere. We analyze the magnetic reconnection in circular orbits and investigate the energy extraction region, as well as the power and efficiency of energy extraction. Our results indicate that energy extraction remains feasible even for a low spin parameter of $0.4$, which is well below the previously reported threshold. We also find that the energy extraction power exceeds that of the Blandford-Znajek mechanism. The Gauss-Bonnet coupling parameters $\alpha$ lower the spin threshold for energy extraction. Similarly, we analyze the energy extraction region in the plunging regime, together with the corresponding power and efficiency. We find that energy extraction is possible even for a low spin parameter as $0.22$. We also observe that the Gauss-Bonnet coupling parameter $\alpha$ further lowers the spin threshold. This behavior is consistent with that observed in the circular orbit case. Finally, we compare the energy extraction power in the plunging and circular orbit regimes and find that the plunging regime yields a higher energy extraction power than the circular orbit case.

gr-qc↗

arXiv2025

Weakly Supervised Pixel-Level Annotation with Visual Interpretability

Medical image annotation is essential for diagnosing diseases, yet manual annotation is time-consuming, costly, and prone to variability among experts. To address these challenges, we propose an automated explainable annotation system that integrates ensemble learning, visual explainability, and uncertainty quantification. Our approach combines three pre-trained deep learning models - ResNet50, EfficientNet, and DenseNet - enhanced with XGrad-CAM for visual explanations and Monte Carlo Dropout for uncertainty quantification. This ensemble mimics the consensus of multiple radiologists by intersecting saliency maps from models that agree on the diagnosis while uncertain predictions are flagged for human review. We evaluated our system using the TBX11K medical imaging dataset and a Fire segmentation dataset, demonstrating its robustness across different domains. Experimental results show that our method outperforms baseline models, achieving 93.04% accuracy on TBX11K and 96.4% accuracy on the Fire dataset. Moreover, our model produces precise pixel-level annotations despite being trained with only image-level labels, achieving Intersection over Union IoU scores of 36.07% and 64.7%, respectively. By enhancing the accuracy and interpretability of image annotations, our approach offers a reliable and transparent solution for medical diagnostics and other image analysis tasks.

cs.CV↗

arXiv2025

TagGAN: A Generative Model for Data Tagging

Precise identification and localization of disease-specific features at the pixel-level are particularly important for early diagnosis, disease progression monitoring, and effective treatment in medical image analysis. However, conventional diagnostic AI systems lack decision transparency and cannot operate well in environments where there is a lack of pixel-level annotations. In this study, we propose a novel Generative Adversarial Networks (GANs)-based framework, TagGAN, which is tailored for weakly-supervised fine-grained disease map generation from purely image-level labeled data. TagGAN generates a pixel-level disease map during domain translation from an abnormal image to a normal representation. Later, this map is subtracted from the input abnormal image to convert it into its normal counterpart while preserving all the critical anatomical details. Our method is first to generate fine-grained disease maps to visualize disease lesions in a weekly supervised setting without requiring pixel-level annotations. This development enhances the interpretability of diagnostic AI by providing precise visualizations of disease-specific regions. It also introduces automated binary mask generation to assist radiologists. Empirical evaluations carried out on the benchmark datasets, CheXpert, TBX11K, and COVID-19, demonstrate the capability of TagGAN to outperform current top models in accurately identifying disease-specific pixels. This outcome highlights the capability of the proposed model to tag medical images, significantly reducing the workload for radiologists by eliminating the need for binary masks during training.

eess.IV↗

arXiv2022

Mg-doping and free-hole properties of hot-wall MOCVD GaN

The hot-wall metal-organic chemical vapor deposition (MOCVD), previously shown to enable superior III-nitride material quality and high performance devices, has been explored for Mg doping of GaN. We have investigated the Mg incorporation in a wide doping range ($2.45\times{10}^{18}~cm^{-3}$ up to $1.10\times{10}^{20}~cm^{-3}$) and demonstrate GaN:Mg with low background impurity concentrations under optimized growth conditions. Dopant and impurity levels are discussed in view of Ga supersaturation which provides a unified concept to explain the complexity of growth conditions impact on Mg acceptor incorporation and compensation. The results are analysed in relation to the extended defects, revealed by scanning transmission electron microscopy (STEM), X-ray diffraction (XRD), and surface morphology, and in correlation with the electrical properties obtained by Hall effect and capacitance-voltage (C-V) measurements. This allows to establish a comprehensive picture of GaN:Mg growth by hot-wall MOCVD providing guidance for growth parameters optimization depending on the targeted application. We show that substantially lower H concentration as compared to Mg acceptors can be achieved in GaN:Mg without any in-situ or post-growth annealing resulting in p-type conductivity in as-grown material. State-of-the-art $p$-GaN layers with a low-resistivity and a high free-hole density (0.77 $\Omega$.cm and $8.4\times{10}^{17}~cm^{-3}$, respectively) are obtained after post-growth annealing demonstrating the viability of hot-wall MOCVD for growth of power electronic device structures.

physics.app-ph↗

arXiv2018

Search Based Code Generation for Machine Learning Programs

Machine Learning (ML) has revamped every domain of life as it provides powerful tools to build complex systems that learn and improve from experience and data. Our key insight is that to solve a machine learning problem, data scientists do not invent a new algorithm each time, but evaluate a range of existing models with different configurations and select the best one. This task is laborious, error-prone, and drains a large chunk of project budget and time. In this paper we present a novel framework inspired by programming by Sketching and Partial Evaluation to minimize human intervention in developing ML solutions. We templatize machine learning algorithms to expose configuration choices as holes to be searched. We share code and computation between different algorithms, and only partially evaluate configuration space of algorithms based on information gained from initial algorithm evaluations. We also employ hierarchical and heuristic based pruning to reduce the search space. Our initial findings indicate that our approach can generate highly accurate ML models. Interviews with data scientists show that they feel our framework can eliminate sources of common errors and significantly reduce development time.

cs.SE↗