AI for Anomaly Detection in Cybersecurity
Year: 2025
Type: article
Abstract: – Anomaly detection is essential withinside the majority of dynamic domain names, along with cybersecurity, healthcare, and commercial monitoring, in which machine conduct is constantly converting. traditional anomaly detection techniques select offline gaining knowledge of the use of static statistics or on line gaining knowledge of that adapts to converting statistics over the years. the new idea of lifelong gaining knowledge of, however, introduces a promising alternative, in which fashions study and adapt constantly, converting to conform to new statistics at the same time as keeping expertise from beyond experiences. while promising, lifelong gaining knowledge of has been extensively researched in domain names along with photograph type and reinforcement gaining knowledge of, with constrained extension to anomaly detection. anomaly detection, however, gives demanding situations particular to the domain, along with the dynamic nature of regular conduct and the sparsity of anomalous events, a long way from the instances addressed with the aid of using lifelong gaining knowledge of in different domain names. this paper explores the idea of lifelong anomaly detection, figuring out the considerable demanding situations and possibilities for the development of fashions able to gaining knowledge of from converting statistics over the years at the same time as keeping expertise from beyond experiences. we advocate a framework for lifelong anomaly detection, describe the situations and techniques appropriate for this method, and description a state of affairs era method for experimentation. through experimentation with conventional anomaly detection techniques, we exhibit the overall performance hole among conventional techniques and techniques the use of lifelong gaining knowledge of. our consequences emphasize the significance of making use of strategies of lifelong gaining knowledge of to assemble extra sturdy anomaly detection structures able to non-stop edition and keeping expertise withinside the lengthy term. the findings display that life on. anomaly detection holds the capability for boosting version overall performance, presenting a extra complete image of dynamic environments.
Authors: R .Bhuvaneswari, Lakshman, Saaketha, Raju, Karthikeya, Somesh
Institution: SRM Institute of Science and Technology
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6949624375
FWCI:
Cites:
Cited by:
Related to: