arXiv subjects

Seraj Al Mahmud Mostafa

Publications and source records attributed to Seraj Al Mahmud Mostafa.

10 recordsLinked to original sources

arXiv2026

Causal Feedback Discovery using Convergence Cross Mapping on Sea Ice Data

Identifying causal relationships in climate systems remains challenging due to nonlinear, coupled dynamics that limit the effectiveness of linear and stochastic causal discovery approaches. This study benchmarks Convergence Cross Mapping (CCM) against Granger causality, PCMCI, and VarLiNGAM using both synthetic datasets with ground truth causal links and 41 years of Arctic climate data (1979--2021). Unlike stochastic models that rely on autoregressive residual dependence, CCM leverages Takens' state-space reconstruction and delay-embedding to reconstruct attractor manifolds from time series. Cross mapping between reconstructed manifolds exploits deterministic signatures of causation, enabling the detection of weak and bidirectional causal links that linear models fail to resolve. Results demonstrate that CCM achieves higher specificity and fewer false positives on synthetic benchmarks, while maintaining robustness under observational noise and limited sample lengths. On Arctic data, CCM reveals significant causal interactions between sea ice extent and atmospheric variables like specific humidity, longwave radiation, and surface temperature with a $p$-value of $0.009$, supporting ice-albedo feedbacks and moisture-radiation couplings central to Arctic amplification. In contrast, stochastic approaches miss these nonlinear dependencies or infer spurious causal relations. This work establishes CCM as a robust causal inference tool for nonlinear climate dynamics and provides the first systematic benchmarking framework for method selection in climate research.

stat.AP↗

arXiv2026

TFFM: Topology-Aware Feature Fusion Module via Latent Graph Reasoning for Retinal Vessel Segmentation

Precise segmentation of retinal arteries and veins carries the diagnosis of systemic cardiovascular conditions. However, standard convolutional architectures often yield topologically disjointed segmentations, characterized by gaps and discontinuities that render reliable graph-based clinical analysis impossible despite high pixel-level accuracy. To address this, we introduce a topology-aware framework engineered to maintain vascular connectivity. Our architecture fuses a Topological Feature Fusion Module (TFFM) that maps local feature representations into a latent graph space, deploying Graph Attention Networks to capture global structural dependencies often missed by fixed receptive fields. Furthermore, we drive the learning process with a hybrid objective function, coupling Tversky loss for class imbalance with soft clDice loss to explicitly penalize topological disconnects. Evaluation on the Fundus-AVSeg dataset reveals state-of-the-art performance, achieving a combined Dice score of 90.97% and a 95% Hausdorff Distance of 3.50 pixels. Notably, our method decreases vessel fragmentation by approximately 38% relative to baselines, yielding topologically coherent vascular trees viable for automated biomarker quantification. We open-source our code at https://tffm-module.github.io/.

cs.CV↗

arXiv2025

Visual Bias and Interpretability in Deep Learning for Dermatological Image Analysis

Accurate skin disease classification is a critical yet challenging task due to high inter-class similarity, intra-class variability, and complex lesion textures. While deep learning-based computer-aided diagnosis (CAD) systems have shown promise in automating dermatological assessments, their performance is highly dependent on image pre-processing and model architecture. This study proposes a deep learning framework for multi-class skin disease classification, systematically evaluating three image pre-processing techniques: standard RGB, CMY color space transformation, and Contrast Limited Adaptive Histogram Equalization (CLAHE). We benchmark the performance of pre-trained convolutional neural networks (DenseNet201, Efficient-NetB5) and transformer-based models (ViT, Swin Transformer, DinoV2 Large) using accuracy and F1-score as evaluation metrics. Results show that DinoV2 with RGB pre-processing achieves the highest accuracy (up to 93%) and F1-scores across all variants. Grad-CAM visualizations applied to RGB inputs further reveal precise lesion localization, enhancing interpretability. These findings underscore the importance of effective pre-processing and model choice in building robust and explainable CAD systems for dermatology.

cs.CV↗

arXiv2025

SLA-MORL: SLA-Aware Multi-Objective Reinforcement Learning for HPC Resource Optimization

Dynamic resource allocation for machine learning workloads in cloud environments remains challenging due to competing objectives of minimizing training time and operational costs while meeting Service Level Agreement (SLA) constraints. Traditional approaches employ static resource allocation or single-objective optimization, leading to either SLA violations or resource waste. We present SLA-MORL, an adaptive multi-objective reinforcement learning framework that intelligently allocates GPU and CPU resources based on user-defined preferences (time, cost, or balanced) while ensuring SLA compliance. Our approach introduces two key innovations: (1) intelligent initialization through historical learning or efficient baseline runs that eliminates cold-start problems, reducing initial exploration overhead by 60%, and (2) dynamic weight adaptation that automatically adjusts optimization priorities based on real-time SLA violation severity, creating a self-correcting system. SLA-MORL constructs a 21-dimensional state representation capturing resource utilization, training progress, and SLA compliance, enabling an actor-critic network to make informed allocation decisions across 9 possible actions. Extensive evaluation on 13 diverse ML workloads using production HPC infrastructure demonstrates that SLA-MORL achieves 67.2% reduction in training time for deadline-critical jobs, 68.8% reduction in costs for budget-constrained workloads, and 73.4% improvement in overall SLA compliance compared to static baselines. By addressing both cold-start inefficiency and dynamic adaptation challenges, SLA-MORL provides a practical solution for cloud resource management that balances performance, cost, and reliability in modern ML training environments.

cs.LG↗

arXiv2025

A Multi-Scale Attention-Enhanced Architecture for Gravity Wave Localization in Satellite Imagery

Satellite images present unique challenges due to their high object variability and lower spatial resolution, particularly for detecting atmospheric gravity waves which exhibit significant variability in scale, shape, and pattern extent, making accurate localization highly challenging. This variability is further compounded by dominant unwanted objects such as clouds and city lights, as well as instrumental noise, all within a single image channel, while conventional detection methods struggle to capture the diverse and often subtle features of gravity waves across varying conditions. To address these issues, we introduce YOLO-DCAT incorporating Multi Dilated Residual Convolution (MDRC) and Simplified Spatial and Channel Attention (SSCA), an enhanced version of YOLOv5 specifically designed to improve gravity wave localization by effectively handling their complex and variable characteristics. MDRC captures multi-scale features through parallel dilated convolutions with varying dilation rates, while SSCA focuses on the most relevant spatial regions and channel features to enhance detection accuracy and suppress interference from background noise. In our experiments, the improved model outperformed state-of-the-art alternatives, improving mean Average Precision (mAP) by over 14% and Intersection over Union (IoU) by approximately 17%, demonstrating significantly improved localization accuracy for gravity waves in challenging satellite imagery and contributing to more precise climate research and modeling.

astro-ph.IM↗

arXiv2025

Enhancing Satellite Object Localization with Dilated Convolutions and Attention-aided Spatial Pooling

Object localization in satellite imagery is particularly challenging due to the high variability of objects, low spatial resolution, and interference from noise and dominant features such as clouds and city lights. In this research, we focus on three satellite datasets: upper atmospheric Gravity Waves (GW), mesospheric Bores (Bore), and Ocean Eddies (OE), each presenting its own unique challenges. These challenges include the variability in the scale and appearance of the main object patterns, where the size, shape, and feature extent of objects of interest can differ significantly. To address these challenges, we introduce YOLO-DCAP, a novel enhanced version of YOLOv5 designed to improve object localization in these complex scenarios. YOLO-DCAP incorporates a Multi-scale Dilated Residual Convolution (MDRC) block to capture multi-scale features at scale with varying dilation rates, and an Attention-aided Spatial Pooling (AaSP) module to focus on the global relevant spatial regions, enhancing feature selection. These structural improvements help to better localize objects in satellite imagery. Experimental results demonstrate that YOLO-DCAP significantly outperforms both the YOLO base model and state-of-the-art approaches, achieving an average improvement of 20.95% in mAP50 and 32.23% in IoU over the base model, and 7.35% and 9.84% respectively over state-of-the-art alternatives, consistently across all three satellite datasets. These consistent gains across all three satellite datasets highlight the robustness and generalizability of the proposed approach. Our code is open sourced at https://github.com/AI-4-atmosphere-remote-sensing/satellite-object-localization.

cs.CV↗

arXiv2025

Efficient Leaf Disease Classification and Segmentation using Midpoint Normalization Technique and Attention Mechanism

Enhancing plant disease detection from leaf imagery remains a persistent challenge due to scarce labeled data and complex contextual factors. We introduce a transformative two-stage methodology, Mid Point Normalization (MPN) for intelligent image preprocessing, coupled with sophisticated attention mechanisms that dynamically recalibrate feature representations. Our classification pipeline, merging MPN with Squeeze-and-Excitation (SE) blocks, achieves remarkable 93% accuracy while maintaining exceptional class-wise balance. The perfect F1 score attained for our target class exemplifies attention's power in adaptive feature refinement. For segmentation tasks, we seamlessly integrate identical attention blocks within U-Net architecture using MPN-enhanced inputs, delivering compelling performance gains with 72.44% Dice score and 58.54% IoU, substantially outperforming baseline implementations. Beyond superior accuracy metrics, our approach yields computationally efficient, lightweight architectures perfectly suited for real-world computer vision applications.

cs.CV↗

arXiv2025

Does More Bandwidth Really Not Matter (Much)?

The prevailing wisdom is that more network bandwidth does not matter much and that website performance is primarily limited by network latency. However, as mobile websites become more complex and mobile network performance improves, does this adage continue to hold? To understand the effects of small changes in network bandwidth and latency on website performance, we propose a novel webpage characterization metrics, Critical Path of Improvement (CPI). We compute CPI for 45 websites and analyze it against the network performance of four mobile ISPs in 57 US cities. Our results show that 18% of websites are primarily limited by bandwidth with others limited by bandwidth to some extent. These results show that contrary to accepted wisdom, insufficient bandwidth is a limiting factor in some website/network combinations. We also offer a discussion of approaches website developers and mobile network administrators can follow to understand and mitigate bandwidth limitations to website performance.

cs.NI↗

arXiv2024

YOLO based Ocean Eddy Localization with AWS SageMaker

Ocean eddies play a significant role both on the sea surface and beneath it, contributing to the sustainability of marine life dependent on oceanic behaviors. Therefore, it is crucial to investigate ocean eddies to monitor changes in the Earth, particularly in the oceans, and their impact on climate. This study aims to pinpoint ocean eddies using AWS cloud services, specifically SageMaker. The primary objective is to detect small-scale (<20km) ocean eddies from satellite remote images and assess the feasibility of utilizing SageMaker, which offers tools for deploying AI applications. Moreover, this research not only explores the deployment of cloud-based services for remote sensing of Earth data but also evaluates several YOLO (You Only Look Once) models using single and multi-GPU-based services in the cloud. Furthermore, this study underscores the potential of these services, their limitations, challenges related to deployment and resource management, and their user-riendliness for Earth science projects.

cs.CV↗

arXiv2024

gWaveNet: Classification of Gravity Waves from Noisy Satellite Data using Custom Kernel Integrated Deep Learning Method

Atmospheric gravity waves occur in the Earths atmosphere caused by an interplay between gravity and buoyancy forces. These waves have profound impacts on various aspects of the atmosphere, including the patterns of precipitation, cloud formation, ozone distribution, aerosols, and pollutant dispersion. Therefore, understanding gravity waves is essential to comprehend and monitor changes in a wide range of atmospheric behaviors. Limited studies have been conducted to identify gravity waves from satellite data using machine learning techniques. Particularly, without applying noise removal techniques, it remains an underexplored area of research. This study presents a novel kernel design aimed at identifying gravity waves within satellite images. The proposed kernel is seamlessly integrated into a deep convolutional neural network, denoted as gWaveNet. Our proposed model exhibits impressive proficiency in detecting images containing gravity waves from noisy satellite data without any feature engineering. The empirical results show our model outperforms related approaches by achieving over 98% training accuracy and over 94% test accuracy which is known to be the best result for gravity waves detection up to the time of this work. We open sourced our code at https://rb.gy/qn68ku.

cs.CV↗