The growing sophistication of network attacks makes it difficult for cybersecurity systems mainly based on predefined rules and static traffic representations. In our framework, we initially used Random Forest, XGBoost and Isolation Forest for attack classification and anomaly detection. To overcome the limitations of the above approach, the extended framework suggests the use of Artificial Neural Network (ANN) and Convolutional Neural Network (CNN) models which are capable to learn complex relationships in network traffic. The framework retains the existing pre-processing, class balancing, Explainable Artificial Intelligence, and Security Operations Center components and incorporates a deep learning layer. The performance of the model is evaluated by accuracy, precision, recall, f1 score and confusion-matrix analysis. Furthermore, a Human-in-the-Loop mechanism enables analysts to validate predictions and correct false detections. Their decisions are stored as structured feedback to enable controlled model improvement. The proposed framework aims to offer a more adaptive and reliable method for cybersecurity monitoring by combining machine learning, deep learning, explainability and human expertise.

Dinesh Kumaar R, Nandeesha M H, R Harsha,"Augmented Intelligence in Machine Learning for Cybersecurity Enhancing Threat Detection and Human-Machine Collaboration" International Journal of Advanced Multidisciplinary Research and Educational Development, V2(5): Page(1053-1062) September - October 2026. ISSN: 3107-6513. www.ijamred.com. Published by Scientific and Academic Research Publishing.