Search for: All records

Backdoor attacks pose a critical threat by embedding hidden triggers into inputs, causing models to misclassify them into target labels. While extensive research has focused on mitigating these attacks in object recognition models through weight fine-tuning, much less attention has been given to detecting backdoored samples directly. Given the vast datasets used in training, manual inspection for backdoor triggers is impractical, and even state-of-the-art defense mechanisms fail to fully neutralize their impact. To address this gap, we introduce a groundbreaking method to detect unseen backdoored images during both training and inference. Leveraging the transformative success of prompt tuning in Vision Language Models (VLMs), our approach trains learnable text prompts to differentiate clean images from those with hidden backdoor triggers. Experiments demonstrate the exceptional efficacy of this method, achieving an impressive average accuracy of 86% across two renowned datasets for detecting unseen backdoor triggers, establishing a new standard in backdoor defense. 

Large Language Models (LLMs) have demonstrated exceptional capabilities in the field of Artificial Intelligence (AI) and are now widely used in various applications globally. However, one of their major challenges is handling high-concurrency workloads, especially under extreme conditions. When too many requests are sent simultaneously, LLMs often become unresponsive which leads to performance degradation and reduced reliability in real-world applications. To address this issue, this paper proposes a queue-based system that separates request handling from direct execution. By implementing a distributed queue, requests are processed in a structured and controlled manner, preventing system overload and ensuring stable performance. This approach also allows for dynamic scalability, meaning additional resources can be allocated as needed to maintain efficiency. Our experimental results show that this method significantly improves resilience under heavy workloads which prevents resource exhaustion and enables linear scalability. The findings highlight the effectiveness of a queue-based web service in ensuring LLMs remain responsive even under extreme workloads. 

The increasing use of high-dimensional imaging in medical AI raises significant privacy and security concerns. This paper presents a Bootstrap Your Own Latent (BYOL)-based self supervised learning (SSL) framework for secure image processing, ensuring compliance with HIPAA and privacy-preserving machine learning (PPML) techniques. Our method integrates federated learning, homomorphic encryption, and differential privacy to enhance security while reducing dependence on labeled data. Experimental results on the MNIST and NIH Chest Xray datasets demonstrate a classification accuracy of 97.5% and 99.99% (pre-fine-tuning 40%), with improved clustering performance using K-Means (Silhouette Score: 0.5247). These findings validate BYOL’s capability for robust, privacy-preserving image processing while emphasizing the need for fine-tuning to optimize classification performance. 

Large language models (LLMs) are becoming a popular tool as they have significantly advanced in their capability to tackle a wide range of language-based tasks. However, LLMs applications are highly vulnerable to prompt injection attacks, which poses a critical problem. These attacks target LLMs applications through using carefully designed input prompts to divert the model from adhering to original instruction, thereby it could execute unintended actions. These manipulations pose serious security threats which potentially results in data leaks, biased outputs, or harmful responses. This project explores the security vulnerabilities in relation to prompt injection attacks. To detect whether a prompt is vulnerable or not, we follows two approaches: 1) a pre-trained LLM, and 2) a fine-tuned LLM. Then, we conduct a thorough analysis and comparison of the classification performance. Firstly, we use pre-trained XLMRoBERTa model to detect prompt injections using test dataset without any fine-tuning and evaluate it by zero-shot classification. Then, this proposed work will apply supervised fine-tuning to this pre-trained LLM using a task-specific labeled dataset from deep set in huggingface, and this fine-tuned model achieves impressive results with 99.13% accuracy, 100% precision, 98.33% recall and 99.15% F1-score thorough rigorous experimentation and evaluation. We observe that our approach is highly efficient in detecting prompt injection attacks. 

This work introduces an novel approach to improving cybersecurity systems to focus on spam email-based cyberattacks. The proposed technique tackles the challenge of training Machine Learning (ML) models with limited data samples by leveraging Bidirectional Encoder Representations from Transformers (BERT) for contextualized embeddings. Unlike traditional embedding methods, BERT offers a nuanced representation of smaller datasets, enabling more effective ML model training. The methodology will use several pre-trained BERT models for generating contextualized embeddings using data samples, and these embeddings will be fed to various ML algorithms for effective training. This approach demonstrates that even with scarce data, BERT embeddings significantly enhance model performance compared to conventional embedding approaches like Word2Vec. The technique proves especially advantageous for insufficient instances of high-quality dataset. The result of this proposed work outperforms traditional techniques to mitigate phishing attacks with few data samples. This work provides a robust accuracy of 99.25% when we use multilingual BERT (M-BERT) to embed dataset. 

In today’s fast-paced software development environments, DevOps has revolutionized the way teams build, test, and deploy applications by emphasizing automation, collaboration, and continuous integration/continuous delivery (CI/CD). However, with these advancements comes an increased need to address security proactively, giving rise to the DevSecOps movement, which integrates security practices into every phase of the software development lifecycle. DevOps security remains underrepresented in academic curricula despite its growing importance in the industry. To address this gap, this paper presents a handson learning module that combines Chaos Engineering and Whitebox Fuzzing to teach core principles of secure DevOps practices in an authentic, scenario-driven environment. Chaos Engineering allows students to intentionally disrupt systems to observe and understand their resilience, while White-box Fuzzing enables systematic exploration of internal code paths to discover cornercase vulnerabilities that typical tests might miss. The module was deployed across three academic institutions, and both pre- and post-surveys were conducted to evaluate its impact. Pre-survey data revealed that while most students had prior experience in software engineering and cybersecurity, the majority lacked exposure to DevOps security concepts. Post-survey responses gathered through ten structured questions showed highly positive feedback 66.7% of students strongly agreed, and 22.2% agreed that the hands-on labs improved their understanding of secure DevOps practices. Participants also reported increased confidence in secure coding, vulnerability detection, and resilient infrastructure design. These findings support the integration of experiential learning techniques like chaos simulations and white-box fuzzing into security education. By aligning academic training with realworld industry needs, this module effectively prepares students for the complex challenges of modern software development and operations. 

Although software developers of mHealth apps are responsible for protecting patient data and adhering to strict privacy and security requirements, many of them lack awareness of HIPAA regulations and struggle to distinguish between HIPAA rules categories. Therefore, providing guidance of HIPAA rules patterns classification is essential for developing secured applications for Google Play Store. In this work, we identified the limitations of traditional Word2Vec embeddings in processing code patterns. To address this, we adopt multilingual BERT (Bidirectional Encoder Representations from Transformers) which offers contextualized embeddings to the attributes of dataset to overcome the issues. Therefore, we applied this BERT to our dataset for embedding code patterns and then uses these embedded code to various machine learning approaches. Our results demonstrate that the models significantly enhances classification performance, with Logistic Regression achieving a remarkable accuracy of 99.95%. Additionally, we obtained high accuracy from Support Vector Machine (99.79%), Random Forest (99.73%), and Naive Bayes (95.93%), outperforming existing approaches. This work underscores the effectiveness and showcases its potential for secure application development.