Sana Ashraf

Dr. Sana Ashraf

Computer Science and Software Engineering, Shenzhen University, Shenzhen, Guangdong, China
orcid

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 July 2024 in Sustainability
Verified
Journal Article
Open access
PDF file
QZQingyu Zhang
Qingyu Zhang
Aman UllahAman Ullah
Aman Ullah
SASana Ashraf
Sana Ashraf
MAMuhammad Abdullah
Muhammad Abdullah
The crucial shift towards sustainable and circular practices is evident in the dynamic and environmentally focused landscape of the fast fashion industry. This study explores how internet of things (IOT) and big-data-driven supply chain (BDSC) technologies mitigate the environmental footprint of fast fashion manufacturing. A survey was conducted on individuals (n = 342) employed in the fast fashion manufacturing sector in Guangzhou, China. Deriving from the resource-based view (RBV) theory, IOT shows a significantly positive influence on firm performance through the mediation of green supply chain management (GSCM) and circular economy (CE) practices. Moreover, moderation of BDSC further strengthens the influence of IOT, GSCM and CE practices. Our empirical analysis investigates the synergistic impact of IOT and BDSC technologies to enhance the overall performance of CE supply chains. The research provides novel insights to improve the environmental sustainability of fast fashion sector.

5 April 2023 in Applied Sciences
Verified
Journal Article
Open access
PDF file
DADr. Saqib Ali
Dr. Saqib Ali
SASana Ashraf
Sana Ashraf
MYMuhammad Sohaib Yousaf
Muhammad Sohaib Yousaf
SRShazia Riaz
Shazia Riaz
Guojun WangGuojun Wang
Guojun Wang
The successful outcomes of deep learning (DL) algorithms in diverse fields have prompted researchers to consider backdoor attacks on DL models to defend them in practical applications. Adversarial examples could deceive a safety-critical system, which could lead to hazardous situations. To cope with this, we suggested a segmentation technique that makes hidden trigger backdoor attacks more robust. The tiny trigger patterns are conventionally established by a series of parameters encompassing their DNN size, location, color, shape, and other defining attributes. From the original triggers, alternate triggers are generated to control the backdoor patterns by a third party in addition to their original designer, which can produce a higher success rate than the original triggers. However, the significant downside of these approaches is the lack of automation in the scene segmentation phase, which results in the poor optimization of the threat model. We developed a novel technique that automatically generates alternate triggers to increase the effectiveness of triggers. Image denoising is performed for this purpose, followed by scene segmentation techniques to make the poisoned classifier more robust. The experimental results demonstrated that our proposed technique achieved 99% to 100% accuracy and helped reduce the vulnerabilities of DL models by exposing their loopholes.

18 March 2026 in Scientific Reports
Verified
Journal Article
Open access
Aman UllahAman Ullah
Aman Ullah
SASana Ashraf
Sana Ashraf
EAEman Abdullah Aldakheel
Eman Abdullah Aldakheel
MAMuhammad Abdullah
Muhammad Abdullah
DKDoaa Sami Khafaga
Doaa Sami Khafaga
This study explores the application of the equivalence class transformation (ECLAT) algorithm for frequent itemset mining and association rule discovery in the analysis of employee mental health data within the workplace. As mental health continues to be recognized as a crucial factor influencing workplace productivity, understanding the factors that impact employee well-being is of paramount importance. To address this, we apply the ECLAT algorithm, a technique known for its efficiency in handling large datasets through vertical transaction representation and depth-first search methods. Unlike traditional algorithms, ECLAT minimizes computational complexity and overhead, allowing for the rapid identification of frequent itemsets. By analyzing the dataset, the ECLAT algorithm identifies the most frequent combinations of these factors, and association rules are derived to uncover the relationships between various mental health elements. These rules are assessed using support, confidence, and lift metrics to determine their significance. The results offer critical insights into the influence of workplace policies on employee mental health, providing a foundation for organizations to create more supportive environments that promote mental well-being. Ultimately, this research contributes to advancing workplace well-being initiatives, fostering a mentally healthy workforce, and driving organizational success.

20 May 2026 in Expert Systems with Applications
Verified
Journal Article
Closed access
Aman UllahAman Ullah
Aman Ullah
QZQingyu Zhang
Qingyu Zhang
SASana Ashraf
Sana Ashraf
Horst TreiblmaierHorst Treiblmaier
Horst Treiblmaier

5 August 2026 in Corporate Social Responsibility and Environmental Management
Verified
Journal Article
Closed access
PDF file
Aman UllahAman Ullah
Aman Ullah
QZQingyu Zhang
Qingyu Zhang
SASana Ashraf
Sana Ashraf
Horst TreiblmaierHorst Treiblmaier
Horst Treiblmaier

5 September 2025 in IEEE Transactions on Dependable and Secure Computing
Verified
Journal Article
Closed access
SASana Ashraf
Sana Ashraf
QYQiao Yan
Qiao Yan
SLSizhe Liang
Sizhe Liang
Aman UllahAman Ullah
Aman Ullah
Few-shot learning (FSL) has demonstrated tremendous potential when challenged with limited training data, but the assessment of its vulnerability to backdoor attacks is still at an early stage. However, recent research revealed this deep learning framework is susceptible to backdoor attack. Existing backdoor techniques attacked FSL by manipulating triggers in the feature space, resulting in overfitting to specific perturbations, poor tolerance to real-world variability, and easy detection. Limited training samples expose these triggers or dirty labels. In this paper, we propose a novel technique that leverages latent embedding to successfully implant backdoor attack on Prototypical Network–based few-shot classification model (prototype-based FSL). Our attack is specifically designed to target the prototype-based FSL due to its effectiveness in constructing fixed class prototypes from limited examples, making it uniquely susceptible to subtle prototype shifts. Since prototypical networks do not require backpropagation on testing samples, there's less chance to detect latent backdoors. The latent mechanism serve as a crucial enhancement to traditional perturbation-based backdoors. For this purpose, our proposed approach utilizes conditional variational autoencoder (CVAE) along with latent attention mechanism and regularization terms to seamlessly encode backdoor trigger within the original image feature space, offering a robust and reconstructed poisoned representation while preserving data integrity. In this setup, the reconstructed poisoned subset by CVAE, combined with clean images, serves as the support set for computing class prototypes. Besides, our strategy enhances generalization by focusing on high-level abstractions and aligns well with the objectives of prototype-based FSL. The experiment results reveal that our poisoning technique achieves high Attack Success Rate (ASR) in FSL challenges while ensuring benign accuracy (BA) and preserving stealthiness. Consequently, this approach outperforms earlier techniques in terms of efficiency and performance, leading to enhanced robustness against detection while assuring that the trigger remains smoothly integrated into the data distribution. This study demonstrates that latent backdoor attacks pose a persistent and significant threat to prototype-based FSL, underscoring an urgent need for robust security measures to protect against these vulnerabilities.