Enhanced Machine Learning Algorithms for Real-Time Anomaly Detection in Network Security: Addressing Scalability, Adaptability and Privacy Challenges in the Face of Emerging Cyber Threats

Vaibhav Sharma, Vikas Singh, Vikas Kumar Tiwari, S. Narayanasamy, Shakthi Sharan R., G. V. Rambabu

2025

Abstract

The increasing complexity of cyber threats demands smart, adaptable and efficient network security solutions. In this paper we present an efficient machine learning framework to detect network anomalies in real-time, which overcomes these limitations by existing methods at least in one of the following aspects: misleading false positive rates, lack of generalizability, static threat model sor computational overhead. By adopting hybrid deep learning networks with attention mechanisms and SHAP-based interpretability, the designed system is capable of providing detection accuracy and interpretability. The framework is evaluated on real datasets, integrating federated private-preserving methodologies, and shows evidence of zero-day attack adaptation using continual learning. Additionally, the lightness allows being used in resources lacking and multi-network environments while keeping detection capabilities. This paper presents a scalable, privacy-sensitive, and adaptable solution to protect present networks from recent cyber-attacks.

Paper Citation

in Harvard Style

Sharma V., Singh V., Tiwari V., Narayanasamy S., R. S. and Rambabu G. (2025). Enhanced Machine Learning Algorithms for Real-Time Anomaly Detection in Network Security: Addressing Scalability, Adaptability and Privacy Challenges in the Face of Emerging Cyber Threats. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 807-813. DOI: 10.5220/0013944000004919

in Bibtex Style

@conference{icrdicct`2525,
author={Vaibhav Sharma and Vikas Singh and Vikas Tiwari and S. Narayanasamy and Shakthi R. and G. Rambabu},
title={Enhanced Machine Learning Algorithms for Real-Time Anomaly Detection in Network Security: Addressing Scalability, Adaptability and Privacy Challenges in the Face of Emerging Cyber Threats},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={807-813},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013944000004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25
TI - Enhanced Machine Learning Algorithms for Real-Time Anomaly Detection in Network Security: Addressing Scalability, Adaptability and Privacy Challenges in the Face of Emerging Cyber Threats
SN - 978-989-758-777-1
AU - Sharma V.
AU - Singh V.
AU - Tiwari V.
AU - Narayanasamy S.
AU - R. S.
AU - Rambabu G.
PY - 2025
SP - 807
EP - 813
DO - 10.5220/0013944000004919
PB - SciTePress