Project Detail
Masters Project
Adversarial Vulnerability Analysis of Deep Neural Network-Based Intrusion Detection Systems Using FGSM and PGD Attacks
Deep learning significantly improves Intrusion Detection Systems (IDS) by automatically recognizing complex attack patterns in network traffic, but these models are highly vulnerable to adversarial manipulation.
This paper analyzes how two powerful gradient-based attacks—Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD)—can degrade a deep neural network–based IDS on NSL‑KDD and CICIDS2017, and shows how increasing perturbation levels cause systematic drops in accuracy, precision, recall, and F1‑score.
Project Detail
B.Tech Project
Personality Evaluation and CV Analysis Using Machine Learning Algorithm March, 2023
- Developed a Machine Learning model for automated personality evaluation and CV analysis, integrating Natural Language Processing (NLP) and supervised learning techniques.
- Utilized classification algorithms (e.g., Random Forest, SVM) to predict candidate-role compatibility based on CV content and inferred personality traits.
- Achieved 75%+ accuracy in matching candidates to ideal job profiles, supporting data-driven recruitment and talent screening processes.
Approach And Technology
The system combines Natural Language Processing (NLP) with supervised machine learning to understand both the content and tone of a candidate’s CV. Classification algorithms such as Random Forest and Support Vector Machine (SVM) are used to predict candidate–role compatibility based on extracted features and inferred personality attributes.
Key Features
- Automated CV parsing and text preprocessing using NLP techniques to extract skills, experience, education, and behavioral cues from resume content.
Role compatibility scoring that maps each candidate to the most relevant job profiles using trained classification models.
- Personality trait inference from writing patterns, keywords, and contextual signals to approximate traits like leadership, adaptability, or attention to detail.
Results And Impact
The model achieved over 75% accuracy in matching candidates to ideal job profiles, demonstrating strong potential as a decision-support tool in recruitment pipelines. By standardizing CV and personality evaluation, the system helps HR teams shortlist candidates faster while reducing subjective bias in early screening stages.
Learning And Future Scope
Through this project, the focus was on combining cybersecurity-aware thinking with responsible AI usage, ensuring that candidate data is handled securely and ethically. Future enhancements may include integrating real-time interview data, psychometric assessments, and explainable AI modules to provide transparent reasons behind each recommendation.
Secure Your Future
Ready to strengthen your security posture or collaborate on cutting-edge research?