Continual Learning with Network Intrusion Dataset
Lists
Only you see this
Abstract
Deep learning-based cybersecurity applications should be able to continually accumulate threat knowledge for new types of threats over time while maintaining the knowledge of threats already exposed to the application. This paper proposes episodic memory management for continual learning with network intrusion datasets. For new attacks, the number of samples may not be sufficiently large for training, and thus the memory management algorithm should retain as many samples as possible instead of random sampling in the episodic memory for continual learning. The experiment results indicated that the proposed algorithm outperforms offline learning in terms of average per-class accuracy in a continual scenario with a network intrusion dataset.
BibTeX
@inproceedings{kim2022continual,
author = {Hyejin Kim and Dan Dongseong Kim and Jin-Hee Cho and Terrence J. Moore and Frederica F. Nelson and Hyuk Lim},
title = {{Continual Learning with Network Intrusion Dataset}},
booktitle = {IEEE Big Data},
year = {2022},
doi = {10.1109/bigdata55660.2022.10020865},
}Loading related content...