Adaptive Defense: Zero-Day Attack Detection in NIDS With Deep Reinforcement Learning

Zero-Day attack detection in Network Intrusion Detection Systems (NIDS) refers to the ability to identify previously unseen attack patterns during testing without having been explicitly trained on those specific attacks, utilizing learned features from other known attacks. In this paper, we propose...

Full description

Saved in:
Bibliographic Details
Main Authors: Khorshed Alam, Md Fahad Monir, Md Junayed Hossain, Mohammad Shorif Uddin, Md. Tarek Habib
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11063272/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849427497659138048
author Khorshed Alam
Md Fahad Monir
Md Junayed Hossain
Mohammad Shorif Uddin
Md. Tarek Habib
author_facet Khorshed Alam
Md Fahad Monir
Md Junayed Hossain
Mohammad Shorif Uddin
Md. Tarek Habib
author_sort Khorshed Alam
collection DOAJ
description Zero-Day attack detection in Network Intrusion Detection Systems (NIDS) refers to the ability to identify previously unseen attack patterns during testing without having been explicitly trained on those specific attacks, utilizing learned features from other known attacks. In this paper, we propose a Deep Reinforcement Learning (DRL)-based NIDS designed for Zero-Day attack detection. We use a stacked LSTM architecture to extend the learning capabilities of the DRL agent. We apply several oversampling techniques to handle the issue of class imbalance since the zero-day attack datasets are not as abundant. We use some of the most widely available benchmark datasets in NIDS domain, which all together cover a wide range of attack types, such as reconnaissance, ddoS, infiltration, injection, password attacks, brute force, dos, backdoor, and benign traffic. For example, we converted attacks to 1 and benign traffic to 0, then excluded certain attack categories (DoS and Backdoor) from the training dataset while keeping them in the test dataset. This makes those attack types zero-day attacks, as they are entirely unseen during training. We also compare which data balancing technique works better among K-means SMOTE, SMOTE, Borderline-SMOTE and ADASYN on the performance of our DRL agent. We then demonstrate how powerful our agent is by validating many datasets for remarkable success in detecting both known and unknown attacks in a zero-day manner. Our work has been made publicly available on GitHub (<uri>https://github.com/codewithkhurshed/ZDAD</uri>) to support researchers in advancing zero-day attack detection in NIDS.
format Article
id doaj-art-56cabc39994d4575bdf07661ef2af658
institution Kabale University
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-56cabc39994d4575bdf07661ef2af6582025-08-20T03:29:02ZengIEEEIEEE Access2169-35362025-01-011311634511636110.1109/ACCESS.2025.358544511063272Adaptive Defense: Zero-Day Attack Detection in NIDS With Deep Reinforcement LearningKhorshed Alam0https://orcid.org/0009-0004-8040-8598Md Fahad Monir1Md Junayed Hossain2Mohammad Shorif Uddin3https://orcid.org/0000-0002-7184-2809Md. Tarek Habib4https://orcid.org/0000-0001-5009-6459Department of Computer Science and Engineering (CSE), Independent University at Bangladesh, Dhaka, BangladeshDepartment of Computer Science and Engineering (CSE), Independent University at Bangladesh, Dhaka, BangladeshDepartment of Computer Science and Engineering (CSE), Independent University at Bangladesh, Dhaka, BangladeshDepartment of Computer Science and Engineering, Jahangirnagar University, Dhaka, BangladeshDepartment of Computer Science and Engineering (CSE), Independent University at Bangladesh, Dhaka, BangladeshZero-Day attack detection in Network Intrusion Detection Systems (NIDS) refers to the ability to identify previously unseen attack patterns during testing without having been explicitly trained on those specific attacks, utilizing learned features from other known attacks. In this paper, we propose a Deep Reinforcement Learning (DRL)-based NIDS designed for Zero-Day attack detection. We use a stacked LSTM architecture to extend the learning capabilities of the DRL agent. We apply several oversampling techniques to handle the issue of class imbalance since the zero-day attack datasets are not as abundant. We use some of the most widely available benchmark datasets in NIDS domain, which all together cover a wide range of attack types, such as reconnaissance, ddoS, infiltration, injection, password attacks, brute force, dos, backdoor, and benign traffic. For example, we converted attacks to 1 and benign traffic to 0, then excluded certain attack categories (DoS and Backdoor) from the training dataset while keeping them in the test dataset. This makes those attack types zero-day attacks, as they are entirely unseen during training. We also compare which data balancing technique works better among K-means SMOTE, SMOTE, Borderline-SMOTE and ADASYN on the performance of our DRL agent. We then demonstrate how powerful our agent is by validating many datasets for remarkable success in detecting both known and unknown attacks in a zero-day manner. Our work has been made publicly available on GitHub (<uri>https://github.com/codewithkhurshed/ZDAD</uri>) to support researchers in advancing zero-day attack detection in NIDS.https://ieeexplore.ieee.org/document/11063272/Zero-day attack detectiondeep reinforcement learningcybersecuritynetwork intrusion detection systemsinternetunseen attack generalization
spellingShingle Khorshed Alam
Md Fahad Monir
Md Junayed Hossain
Mohammad Shorif Uddin
Md. Tarek Habib
Adaptive Defense: Zero-Day Attack Detection in NIDS With Deep Reinforcement Learning
IEEE Access
Zero-day attack detection
deep reinforcement learning
cybersecurity
network intrusion detection systems
internet
unseen attack generalization
title Adaptive Defense: Zero-Day Attack Detection in NIDS With Deep Reinforcement Learning
title_full Adaptive Defense: Zero-Day Attack Detection in NIDS With Deep Reinforcement Learning
title_fullStr Adaptive Defense: Zero-Day Attack Detection in NIDS With Deep Reinforcement Learning
title_full_unstemmed Adaptive Defense: Zero-Day Attack Detection in NIDS With Deep Reinforcement Learning
title_short Adaptive Defense: Zero-Day Attack Detection in NIDS With Deep Reinforcement Learning
title_sort adaptive defense zero day attack detection in nids with deep reinforcement learning
topic Zero-day attack detection
deep reinforcement learning
cybersecurity
network intrusion detection systems
internet
unseen attack generalization
url https://ieeexplore.ieee.org/document/11063272/
work_keys_str_mv AT khorshedalam adaptivedefensezerodayattackdetectioninnidswithdeepreinforcementlearning
AT mdfahadmonir adaptivedefensezerodayattackdetectioninnidswithdeepreinforcementlearning
AT mdjunayedhossain adaptivedefensezerodayattackdetectioninnidswithdeepreinforcementlearning
AT mohammadshorifuddin adaptivedefensezerodayattackdetectioninnidswithdeepreinforcementlearning
AT mdtarekhabib adaptivedefensezerodayattackdetectioninnidswithdeepreinforcementlearning