Course

Unit 1

Machine Learning Foundations for Security Analytics

Introduction to applied machine learning in cybersecurity, cyber security as a data-driven domain, types of learning – supervised, unsupervised, semi-supervised, reinforcement learning, security dataset characteristics – noise, missing data, imbalance, concept drift, data collection and labeling challenges in security, feature engineering from security events, dimensionality reduction techniques, evaluation metrics – accuracy, precision, recall, F1-score, ROC-AUC, reproducibility in security ML research.

Unit 2

Predictive Modeling and Classification Techniques

Classification workflow for security problems, logistic regression for risk prediction, decision trees and rule learning, random forest and ensemble methods, support vector machines, naïve Bayes for probabilistic detection, model selection and hyperparameter tuning, cross validation strategies, performance optimization and model generalization techniques, cost-sensitive learning in security.

Unit 3

Adversarial Machine Learning

Introduction to adversarial machine learning, threat models in machine learning, evasion attacks, data poisoning attacks, adversarial examples, white-box and black-box attacks, model extraction attacks, membership inference attacks, robustness testing of ML models, attack surface analysis, basic defense techniques against adversarial attacks, secure model development practices, case studies in adversarial machine learning.

Unit 4

Explainable AI, Transfer Learning and Federated Learning

Explainable AI concepts, interpretable models vs black-box models, SHAP and LIME techniques, model explainability in decision-making systems, introduction to transfer learning, pre-trained models, domain adaptation concepts, applications of transfer learning, federated learning architecture and workflow, decentralized model training, privacy-preserving machine learning, applications of federated learning in distributed environments.

Text Books / References

Text Book / References

1. Clarence Chio and David Freeman, Machine Learning and ecurity, O’Reilly Media, 2018.
2. Aurélien Géron, Hands-On Machine Learning wit Scikit-Learn, Keras and TensorFlow, O’Reilly Media, 3rd Edition, 2022.
3. Soma Halder and Sinan Ozdemir, Practical Machine Learning for Cybersecurity, Packt Publishing, 2018.
4. Christoph Molnar, Interpretable Machine Learning, Lulu Publications, 2nd Edition.
5. Sumeet Dua and Xian Du, Data Mining and Machine Learning in Cybersecurity, CRC Press.

Preamble

This course equips M.Tech Cyber Security Systems and Networks students with the theoretical foundation and practical skills to apply machine learning techniques for real-world analytical problems. The course emphasizes predictive modeling, classification, anomaly detection, explainable AI, model robustness, and deployment of machine learning systems in modern computing environments. Students will develop competence in designing intelligent data-driven solutions using contemporary machine learning workflows while addressing challenges such as imbalanced datasets, concept drift, false positives, model interpretability, and trustworthy AI

Course Objectives

• Understand the role of machine learning in modern cybersecurity systems
• Apply supervised and unsupervised learning methods for threat analytics
• Build predictive models for cyber risk scoring and suspicious activity detection
• Develop explainable, robust, and privacy-aware ML solutions for security operations
• Evaluate and deploy ML pipelines for practical cybersecurity environments

Course Outcomes

CO-PO Mapping

Correlation Levels: 3 = High, 2 = Moderate, 1 = Low

Evaluation Pattern – 60:40

• Midterm Exam – 20% (Lab Examination)
• Class Test – 10%
• Lab Assignments – 10%
• Project – 20%

End Sem Exam – 40% (Lab Examination)