DeepAID: Interpreting and Improving Deep Learning-based Anomaly Detection in Security Applications
Open accessThe DOI matches its Crossref record; 78 of the 78 references it cites were found in a registry or an index.
Abstract
Unsupervised Deep Learning (DL) techniques have been widely used in various security-related anomaly detection applications, owing to the great promise of being able to detect unforeseen threats and superior performance provided by Deep Neural Networks (DNN). However, the lack of interpretability creates key barriers to the adoption of DL models in practice. Unfortunately, existing interpretation approaches are proposed for supervised learning models and/or non-security domains, which are unadaptable for unsupervised DL models and fail to satisfy special requirements in security domains.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Similarly, a recent systematic review [31] highlighted that research on XAI for cybersecurity shows that many XAI applications are crafted without a thorough understanding of their integration into analyst workflows.”
From HuntGPT: Integrating Machine Learning-Based Anomaly Detection and Explainable AI with Large Language Models (LLMs) (2026) · cites this paper as [31]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Moreover, security literature frequently fails to differentiate between diverse use cases or clearly separate the roles of model users and designers, which could potentially lead to diminished adoption.”
From HuntGPT: Integrating Machine Learning-Based Anomaly Detection and Explainable AI with Large Language Models (LLMs) (2026) · cites this paper as [31]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 1×
“This problem can be solved by gradient-based optimization approaches (we use Adam optimizer [28] in this study) since the objective function is fully differentiable.”
MethodPyTorch: An Imperative Style, High-Performance Deep Learning Library
Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al. (2019) · arXiv · doi:10.48550/arxiv.1912.01703 · Cited by 31,473 · cited 2×
“, PyTorch [40]) set the embedding layer to be indifferentiable.”
BackgroundDistributed Representations of Words and Phrases and their Compositionality
Mikolov, T., Sutskever, I., Chen, K., Corrado, G., & Dean, J. (2013) · arXiv · doi:10.48550/arxiv.1310.4546 · Cited by 24,533 · cited 1×
“However, embeddings of inputs are derived by an indifferentiable look-up operation [34] in many cases.”
BackgroundA Unified Approach to Interpreting Model Predictions
Lundberg, S., & Lee, S.-I. (2017) · arXiv · doi:10.48550/arxiv.1705.07874 · Cited by 23,707 · cited 1×
“For example, SHAP interpretation [32] is adapted to (single) Autoencoder in [2], which cannot be applied to Kitsune (ensemble Autoencoders) or other DL models.”
Method"Why Should I Trust You?": Explaining the Predictions of Any Classifier
Ribeiro, M. T., Singh, S., & Guestrin, C. (2016) · Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2939672.2939778 · Cited by 17,555
Gauriat, C.-M., Pencolé, Y., Ribot, P., & Brouillet, G. (2024) · HAL (Le Centre pour la Communication Scientifique Directe) · doi:10.4230/oasics.dx.2024.27 · Cited by 12,586
Chandola, V., Banerjee, A., & Kumar, V. (2009) · ACM Computing Surveys · doi:10.1145/1541880.1541882 · Cited by 11,687
Learning Deep Features for Discriminative Localization
Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., & Torralba, A. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.319 · Cited by 11,428 · cited 2×
“Secondly, most prior methods are designed for non-security domains such as computer vision [16, 44, 62, 65] to understand the mechanism of DNNs.”
DeepWalk: online learning of social representations
Perozzi, B., Al-Rfou, R., & Skiena, S. (2014) · Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/2623330.2623732 · Cited by 10,123 · cited 2×
“The training data is collected from the purely benign authentication graphs through DeepWalk [41], which is a NN-based GE method.”
MethodTowards Evaluating the Robustness of Neural Networks
Carlini, N., & Wagner, D. (2017) · 2017 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2017.49 · Cited by 8,859 · cited 1×
“We eliminate the box constraint (5) through the idea of change-of-variables [8] with tanh function.”
Method
Research works printed but not linked: 2
- Alexander D. Kent. 2015. Cybersecurity Data Sources for Dynamic Network Research. In Dynamic Networks in Cybersecurity. Imperial College Press.
- Limin Yang, Wenbo Guo, Qingying Hao, Arridhana Ciptadi, Ali Ahmadzadeh, Xinyu Xing, and Gang Wang. 2021. CADE: Detecting and Explaining Concept Drift Samples for Security Applications. In 30th USENIX Security Symposium (USENIX Security).
Cited by
Wang, Y., Sun, T., Li, S., Yuan, X., Ni, W., Hossain, E., & Vincent Poor, H. (2023) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2023.3319492 · Cited by 161
Kairos: Practical Intrusion Detection and Investigation using Whole-system Provenance
Cheng, Z., Lv, Q., Liang, J., Wang, Y., Sun, D., Pasquier, T., & Han, X. (2024) · 2024 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp54263.2024.00005 · Cited by 132
“This is because attack reconstruction (1) establishes trust on the decisions, (2) facilitates the necessary human-in-the-loop component in understanding system anomalies, and (3) expedites the process of identifying and reducing FPs (§5.2) [66].”
BackgroundInfluentialArreche, O., Guntur, T. R., & Abdallah, M. (2024) · Applied Sciences · doi:10.3390/app14104170 · Cited by 224
E-XAI: Evaluating Black-Box Explainable AI Frameworks for Network Intrusion Detection
Arreche, O., Guntur, T. R., Roberts, J. W., & Abdallah, M. (2024) · IEEE Access · doi:10.1109/access.2024.3365140 · Cited by 217
“Similarly, the work [31] contributes a new framework called DeepAID, aiming to address the limitations of unsupervised DL-based anomaly detection systems.”
MethodFlash: A Comprehensive Approach to Intrusion Detection via Provenance Graph Representation Learning
Rehman, M. U., Ahmadi, H., & Hassan, W. U. (2024) · IEEE Symposium on Security and Privacy (S&P) · doi:10.1109/sp54263.2024.00139 · Cited by 262
“DeepAid [29] uses deep neural networks to detect anomalies in network traffic.”
BackgroundXAI-IoT: An Explainable AI Framework for Enhancing Anomaly Detection in IoT Systems
Gummadi, A. N., Napier, J. C., & Abdallah, M. (2024) · IEEE Access · doi:10.1109/access.2024.3402446 · Cited by 158
“Such a process has been applied in several prior works [58], [59], [60], [61].”
ResultAkhunzada, A., Al‐Shamayleh, A. S., Zeadally, S., Almogren, A., & Abu-Shareha, A. A. (2024) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2024.109609 · Cited by 109
Anomaly Detection in the Open World: Normality Shift Detection, Explanation, and Adaptation
Han, D., Wang, Z., Chen, W., Wang, K., Yu, R., Wang, S., Zhang, H., Wang, Z., Jin, M., Yang, J., et al. (2023) · Network and Distributed System Security Symposium · doi:10.14722/ndss.2023.24830 · Cited by 126
“Similar to [29], we tailor GLGV by replacing the anomaly detector with an Autoencoder as the original work use a supervised classifier in the last step which is not suitable for this study.”
MethodSelf-citationSoK: Explainable Machine Learning for Computer Security Applications
Nadeem, A., Vos, D., Cao, C., Pajola, L., Dieck, S., Baumgartner, R., & Verwer, S. (2023) · IEEE European Symposium on Security and Privacy · doi:10.1109/eurosp57164.2023.00022 · Cited by 115
“DeepAID [16] is an explanation method for unsupervised settings, such as anomaly detection.”
MethodAmirkhani, A., Karimi, M. P., & Banitalebi-Dehkordi, A. (2022) · The Visual Computer · doi:10.1007/s00371-022-02660-6 · Cited by 82
“The attacks generated in the Evaporate method are iterative in nature [53], and they will be repeated in a next steps if the obtained images in the previous steps are not good enough.”
MethodNascita, A., Aceto, G., Ciuonzo, D., Montieri, A., Persico, V., & Pescapè, A. (2024) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2024.3504955 · Cited by 102
“The LIME technique is the next most frequently employed interpretability tool, and it is also used in a significant number of NTA-related works [82], [85], [89], [96], [98], [105], [114], [115], [117], [120], [121], [123], [124], [135], [136], [140], [141], [143], [160], [165], [166], [167].”
InfluentialExplainable AI and Random Forest based reliable intrusion detection system
Wali, S. W., Farrukh, Y. A., & Khan, I. (2025) · Computers & Security · doi:10.1016/j.cose.2025.104542 · Cited by 104
Farrukh, Y. A., Wali, S. W., Khan, I., & Bastian, N. D. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.128089 · Cited by 94
AI for DevSecOps: A Landscape and Future Opportunities
Fu, M. C., Pasuksmit, J., & Tantithamthavorn, C. K. (2025) · ACM Transactions on Software Engineering and Methodology · doi:10.1145/3712190 · Cited by 113
“Our investigation reveals that while the majority of AI-driven anomaly detection approaches focused on performance improvement, only a few studies [3, 71] prioritized model explainability.”
Farrukh, Y. A., Wali, S., Khan, I., & Bastian, N. D. (2024) · arXiv · doi:10.48550/arxiv.2408.16021 · Cited by 56
“Han et al. (2021) developed the DeepAID framework to interpret unsupervised deep learning-based anomaly detection systems for cybersecurity.”
Anomaly detection using unsupervised machine learning algorithms: A simulation study
Agyemang, E. F. (2024) · Scientific African · doi:10.1016/j.sciaf.2024.e02386 · Cited by 103
SoK: Pragmatic Assessment of Machine Learning for Network Intrusion Detection
Apruzzese, G., Laskov, P., & Schneider, J. (2023) · IEEE European Symposium on Security and Privacy · doi:10.1109/eurosp57164.2023.00042 · Cited by 118
Lucid: A Non-intrusive, Scalable and Interpretable Scheduler for Deep Learning Training Jobs
Hu, Q., Zhang, M., Sun, P., Wen, Y., & Zhang, T. (2023) · International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) · doi:10.1145/3575693.3575705 · Cited by 82
“Metis [63], DeepAid [34] and Lemna [33] design toolkits to improve system transparency by interpreting black-box ML models.”
BackgroundTrident: A Universal Framework for Fine-Grained and Class-Incremental Unknown Traffic Detection
Zhao, Z., Li, Z., Song, Z., Li, W., & Zhang, F. (2024) · International World Wide Web Conference (WWW) · doi:10.1145/3589334.3645407 · Cited by 76
“With encrypted traffic transmission becoming ubiquitous in practice, the proposed approaches gradually evolve from signature-based [26] to machine-learning-based (ML-based) detection in the traffic analysis landscape [16].”
BackgroundMethodInfluentialFu, C., Li, Q. L., Xu, K., & Wu, J. (2023) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3576915.3616631 · Cited by 62
Ali, T., Kostakos, P., & Sheikhi, S. (2026) · Telecom · doi:10.3390/telecom7030073 · Cited by 104
EXPLORA: AI/ML EXPLainability for the Open RAN
Fiandrino, C., Bonati, L., D’Oro, S., Polese, M., Melodia, T., & Widmer, J. (2023) · Proceedings of the ACM on Networking · doi:10.1145/3629141 · Cited by 70
“Ensuring model robustness is important and can be done through anomaly detection [24] or formal verification [12, 33].”
BackgroundLiu, H., Wu, Y., Yu, Z., & Zhang, N. (2024) · IEEE Symposium on Security and Privacy (S&P) · doi:10.1109/sp54263.2024.00120 · Cited by 85
“Due to its unique ability to facilitate trust establishment between users and machine [8], explainable ML has been employed in various critical domains, such as computer security [9], [10], [11], medical diagnosis [12], [13], and model debugging [14].”
Advanced Persistent Threat Attack Detection Systems: A Review of Approaches, Challenges, and Trends
Buchta, R., Gkoktsis, G., Heine, F., & Kleiner, C. (2024) · Digital Threats Research and Practice · doi:10.1145/3696014 · Cited by 61
Explainable Machine Learning in Critical Decision Systems: Ensuring Safe Application and Correctness
Wiggerthale, J., & Reich, C. (2024) · AI · doi:10.3390/ai5040138 · Cited by 61
Advanced AI-driven intrusion detection for securing cloud-based industrial IoT
Tunio, S., He, J., Qureshi, S. U., Zhu, N., Wajahat, A., Nazir, A., Ullah, F., & Wadud, A. (2025) · Egyptian Informatics Journal · doi:10.1016/j.eij.2025.100644 · Cited by 41
Large Language Models can Deliver Accurate and Interpretable Time Series Anomaly Detection
Liu, J., Zhang, C., Qian, J., Ma, M., Qin, S., Bansal, C., Lin, Q., Rajmohan, S., & Zhang, D. (2025) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3711896.3737239 · Cited by 129
An Explanation of the LSTM Model Used for DDoS Attacks Classification
Bashaiwth, A., Binsalleeh, H., & AsSadhan, B. (2023) · Applied Sciences · doi:10.3390/app13158820 · Cited by 31
“[36] developed an approach to explain and enhance anomaly detection based on DL in security fields called DeepAID.”
MethodENIDrift: A Fast and Adaptive Ensemble System for Network Intrusion Detection under Real-world Drift
Wang, X. (2022) · Annual Computer Security Applications Conference · doi:10.1145/3564625.3567992 · Cited by 42
“Like most ML-based NIDSs [21, 57], they have two phases, training and executing.”
BackgroundMethodUnmasking the Internet: A Survey of Fine-Grained Network Traffic Analysis
Feng, Y., Li, J., Mirković, J., Wu, C., Wang, C., Ren, H., Xu, J., & Liu, Y. (2025) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2025.3545541 · Cited by 47
“Although some rule-based and statistical FGTA approaches demonstrate satisfactory performance in terms of explainabil-ity [19], [113], the majority of FGTA techniques heavily rely on machine learning, with only a limited number emphasizing explainability [151], [163], [300].”
Detecting Tunneled Flooding Traffic via Deep Semantic Analysis of Packet Length Patterns
Fu, C., Li, Q. L., Shen, M., & Xu, K. (2024) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3658644.3670353 · Cited by 29
The Missing Link in Network Intrusion Detection: Taking AI/ML Research Efforts to Users
Dietz, K., Mühlhauser, M., Kögel, J., Schwinger, S., Sichermann, M., Seufert, M., Herrmann, D., & Hoßfeld, T. (2024) · IEEE Access · doi:10.1109/access.2024.3406939 · Cited by 38
Bad Design Smells in Benchmark NIDS Datasets
Flood, R., Engelen, G., Aspinall, D., & Desmet, L. (2024) · IEEE European Symposium on Security and Privacy · doi:10.1109/eurosp60621.2024.00042 · Cited by 77
Zhao, Z., Li, Z., Xie, X., Yu, J., Zhang, F., Zhang, R., Chen, B., Luo, X., Hu, M., & Ma, W. (2024) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2024.3413789 · Cited by 26
“They are also known as “zero-positive” learning [17], [18] due to solely using benign samples for training.”
BackgroundBARS: Local Robustness Certification for Deep Learning based Traffic Analysis Systems
Wang, K., Wang, Z., Han, D., Chen, W., Yang, J., Shi, X., & Yin, X. (2023) · Network and Distributed System Security Symposium · doi:10.14722/ndss.2023.24508 · Cited by 30
“However, DLbased models are highly nonlinear, so the classification boundary cannot be formulated trivially [26], [27], [79].”
BackgroundSelf-citationSun, P., Li, S., Xie, J., Xu, H., Cheng, Z., & Yang, R. (2023) · Computers & Security · doi:10.1016/j.cose.2023.103257 · Cited by 21
Mink, J., Benkraouda, H., Yang, L., Ciptadi, A., Ahmadzadeh, A., Votipka, D., & Wang, G. (2023) · IEEE Symposium on Security and Privacy (S&P) · doi:10.1109/sp46215.2023.10179321 · Cited by 52
“As such, security researchers have worked to tailor explanation methods for security applications [3], [15], [16], [25], [26], understand the robustness of the explanation methods [27],”
BackgroundAttributions for ML-based ICS Anomaly Detection: From Theory to Practice
Fung, C., Zeng, E., & Bauer, L. (2024) · Network and Distributed System Security Symposium · doi:10.14722/ndss.2024.23216 · Cited by 36
“For example, explanation methods have been used in security-relevant ML domains to generate training data for model fine-tuning [33], to automatically detect bias [36], and to detect and explain concept drift [32], [69].”
MethodSTEAMCODER: Spatial and Temporal Adaptive Dynamic Convolution Autoencoder for Anomaly Detection
Xu, P., Gan, H., Fu, H., & Zhang, Z. (2023) · Knowledge-Based Systems · doi:10.1016/j.knosys.2023.110929 · Cited by 19
Szymoniak, S., Depta, F., Karbowiak, Ł., & Kubanek, M. (2023) · Applied Sciences · doi:10.3390/app132112068 · Cited by 24
A comparative analysis of DNN-based white-box explainable AI methods in network security
Arreche, O., & Abdallah, M. (2025) · EURASIP Journal on Information Security · doi:10.1186/s13635-025-00201-x · Cited by 27
“Meanwhile, DeepAID [37] tackled some shortcomings in unsupervised Deep Learning binary detection systems.”
Yao, W., Peng, H., Li, Q., & Shen, X. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3568503 · Cited by 23
“…increasingly sophisticated attacks, various machine learning (ML) and deep-learning (DL)-based NIDSs have been developed in recent years [3], [4], [5], including multilayer perceptron (MLP) [6], convolutional neural networks (CNNs) [7], long short-term memory (LSTM) networks [7], deep…”
Bhusal, D., Shin, H. J., Shewale, A. A., Veerabhadran, M. K., Clifford, M., Rampazzi, S., & Rastogi, N. (2023) · International Conference on Availability, Reliability and Security (ARES) · doi:10.1145/3600160.3600193 · Cited by 36
“In the given test sequence, DeepAID highlights event ID-9 as the point where an anomaly occurs in the sequence.”
MethodInfluentialNetwork anomaly detection via similarity-aware ensemble learning with ADSim
Chen, W., Wang, Z., Chang, L., Wang, K., Zhong, Y. C., Han, D., Duan, C., Yin, X., Yang, J., & Shi, X. (2024) · Computer Networks · doi:10.1016/j.comnet.2024.110423 · Cited by 16
Self-citationDAN: Neural network based on dual attention for anomaly detection in ICS
Xu, L., Wang, B., Zhao, D., & Wu, X.-M. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.125766 · Cited by 21
Is It Overkill? Analyzing Feature-Space Concept Drift in Malware Detectors
Chen, Z., Zhang, Z., Kan, Z., Yang, L., Cortellazzi, J., Pendlebury, F., Pierazzi, F., Cavallaro, L., & Wang, G. (2023) · IEEE Security and Privacy Workshops (SPW) · doi:10.1109/spw59333.2023.00007 · Cited by 13
“In practice, there are various techniques [7, 8, 9, 10] to use a smaller number of labeled samples for retraining, which is not the focus of this paper.”
BackgroundMethodInfluentialFan, M., Zuo, J., Zhu, J., & Lu, Y. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3542142 · Cited by 17
“These have been successfully applied to security applications [10], [21], [22], [23], offering a form of indirect explainability and enhancing the transparency of IDS.”
A Survey on Malware Detection based on API Calls
Chang, K., Zhao, N., & Kou, L. (2022) · International Conference on Dependable Systems and Their Applications (DSA) · doi:10.1109/dsa56465.2022.00067 · Cited by 14
An Explainable-by-Design Ensemble Learning System to Detect Unknown Network Attacks
Minh, C., Vermeulen, K., Lefebvre, C., Owezarski, P., & Ritchie, W. (2023) · International Conference on Network and Service Management (CNSM) · doi:10.23919/cnsm59352.2023.10327818 · Cited by 12
“Han et al. [4] addressed ML explainability by proposing a system to interpret existing unsupervised, DL-based NIDS.”
BackgroundSecEG: A Secure and Efficient Strategy against DDoS Attacks in Mobile Edge Computing
Huang, H., Meng, T., Guo, J., Wei, X., & Jia, W. (2024) · ACM Transactions on Sensor Networks · doi:10.1145/3641106 · Cited by 13
A Dynamic Analysis-Powered Explanation Framework for Malware Detection
Zhu, H., Chen, X., Wang, L., Xu, Z., & Sheng, V. S. (2024) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2024.3436891 · Cited by 12
Genos: General In-Network Unsupervised Intrusion Detection by Rule Extraction
Li, R., Li, Q., Zhang, Y., Zhao, D., Xiao, X., & Jiang, Y. (2024) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom52122.2024.10621157 · Cited by 21
“Existing works in this field [30]–[33] often involve fitting a local linear model (e.g., lasso) to uncover the feature weights, which can be slow and unsuitable for delay-sensitive detection tasks, e.g., real-time online security analysis.”
Mateen: Adaptive Ensemble Learning for Network Anomaly Detection
Alotaibi, F., & Maffeis, S. (2024) · International Symposium on Research in Attacks, Intrusions and Defenses (RAID) · doi:10.1145/3678890.3678901 · Cited by 24
Zhao, J., Li, Q., Han, Z., Fu, J., Nan, G., Shen, M., & Bhargava, B. (2024) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2024.3515821 · Cited by 14
A systematic evaluation of white-box explainable AI methods for anomaly detection in IoT systems
Gummadi, A. N., Arreche, O., & Abdallah, M. (2025) · Internet of Things · doi:10.1016/j.iot.2025.101505 · Cited by 16
Gholian, S. M., Fiandrino, C., Vallina-Rodríguez, N., Fiore, M., & Widmer, J. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3531544 · Cited by 17
“DeepAID [45] introduces a framework tailored for interpreting unsupervised deep learning models in security domains suitable for anomaly detection applications.”
Training with Only 1.0 ‰ Samples: Malicious Traffic Detection via Cross-Modality Feature Fusion
Fu, C., Li, Q. L., Bertino, E., & Xu, K. (2025) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3719027.3765143 · Cited by 16
Liu, S., Chen, J., Liu, Z., Wang, J., & Wang, Z. J. (2024) · Reliability Engineering & System Safety · doi:10.1016/j.ress.2024.110784 · Cited by 10
Self-citationDeep anomaly detection of temporal heterogeneous data in AIOps: a survey
Gui, J., Ma, Z., Zhou, H., Su, Y., Zhang, M., Yu, K., & Wu, X. (2025) · Frontiers of Information Technology & Electronic Engineering · doi:10.1631/fitee.2400467 · Cited by 12
“Han et al. (2021) devised an interpreter and distiller to explain anomalies for analysts.”
He, Y., Lou, J., Qin, Z., & Ren, K. (2023) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3576915.3616599 · Cited by 24
DUdetector: A dual-granularity unsupervised model for network anomaly detection
Geng, H., Ma, Q., Chi, H., Zhang, Z., Yang, J., & Yin, X. (2024) · Computer Networks · doi:10.1016/j.comnet.2024.110937 · Cited by 10
Self-citationLewandowski, B., & Paffenroth, R. C. (2022) · International Conference on Machine Learning and Applications · doi:10.1109/icmla55696.2022.00212 · Cited by 7
“One work that uses autoencoder feature residuals is DeepAID [11], which is proposed as a framework to add interpretability to deep learning-based anomaly detection systems.”
MethodProcess-Aware Intrusion Detection in MQTT Networks
Empl, P., Böhm, F., & Pernul, G. (2024) · ACM Conference on Data and Application Security and Privacy (CODASPY) · doi:10.1145/3626232.3653271 · Cited by 17
Fair XIDS: Ensuring fairness and transparency in intrusion detection models
Chinu, & Bansal, U. (2024) · Concurrency and Computation Practice and Experience · doi:10.1002/cpe.8268 · Cited by 9
Jia, X., Cao, D., Zhuang, N., Peng, W., Zhao, B., Peng, X., Li, H., & Shen, C. (2025) · Neural Networks · doi:10.1016/j.neunet.2025.107903 · Cited by 7
Visually Analyze SHAP Plots to Diagnose Misclassifications in ML-based Intrusion Detection
Mia, M., Pritom, M. M. A., Islam, T., & Hasan, K. (2024) · IEEE ... International Conference on Data Mining workshops · doi:10.1109/icdmw65004.2024.00088 · Cited by 10
“Lastly, Han et al. proposed DeepAID [44] that performs poorly when there are feature dependencies in the dataset which is often the case in IDS.”
Reference-Based AI Decision Support for Cybersecurity
Lee, H.-W., Han, T.-H., & Lee, T. (2023) · IEEE Access · doi:10.1109/access.2023.3342868 · Cited by 11
“Furthermore, a comparison of the proposed Reference generation logic with the Reference generation techniques of prior research, such as that of Han [7], is presented in Table 8.”
MethodInfluentialGe, Y., Gao, Y., Li, X., Cai, B., Xi, J., & Yu, S. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3413580 · Cited by 8
“Han et al. [21] introduced a generic framework called DeepAID, leveraging unsupervised deep neural networks (DNNs) to enhance the interpretability of anomaly detection systems in the field of security.”
BackgroundMethodCombating trade in illegal wood and forest products with machine learning
Datta, D., Simeone, J., Meadows, A., Outhwaite, W., Chen, H., Self, N., Walker, L., & Ramakrishnan, N. (2025) · PLoS ONE · doi:10.1371/journal.pone.0311982 · Cited by 8
Finding Component Relationships: A Deep-Learning-Based Anomaly Detection Interpreter
Xu, L., Han, Z., Wang, Z., & Zhao, D. (2024) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2024.3360435 · Cited by 7
“6) DeepAID [24] is an interpretation framework designed specifically for the security domain, which helps security experts to understand why anomalies occur, proposing distillers to improve the utility of the interpreter.”
BackgroundMethodSelf-citationHan, D., Wang, Z., Feng, R., Jin, M., Chen, W., Wang, K., Wang, S., Yang, J., Shi, X., Yin, X., et al. (2024) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3658644.3670375 · Cited by 11
Self-citationAIChronoLens: AI/ML Explainability for Time Series Forecasting in Mobile Networks
Pérez, P. F., Fiandrino, C., Gómez, E. P., Mohammadalizadeh, H., Fiore, M., & Widmer, J. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3554035 · Cited by 5
“Prominent XAI techniques like LayeR-wise backPropagation (LRP) [23], SHapely Additive exPlanations (SHAP) [24], Local Interpretable Model-agnostic Explanations (LIME) [25], DeepLIFT [26] have been adapted for time series or are also applicable to time series like DeepAID [27].”
Improving Network Intrusion Detection Using Autoencoder Feature Residuals
Lewandowski, B., Paffenroth, R. C., & Campbell, K. (2022) · International Conference on Data Intelligence and Security (ICDIS) · doi:10.1109/icdis55630.2022.00013 · Cited by 7
“Its scenarios, which include attacks such as brute force attacks, denial of service, botnet, and web attacks, have been used as a benchmark cybersecurity dataset for a number of works [22], [23].”
MethodResultLewandowski, B., & Paffenroth, R. C. (2023) · Machine Learning and Knowledge Extraction · doi:10.3390/make5030046 · Cited by 7
ProDE: Interpretable APT Detection Method Based on Encoder-decoder Architecture
Zhou, F., Chang, B., Wen, Y., & Meng, D. (2023) · IEEE International Conference on Parallel and Distributed Systems (ICPADS) · doi:10.1109/icpads60453.2023.00059 · Cited by 4
“However, such methods lack the explanatory power to answer critical questions such as why a particular vector is considered abnormal and what factors contribute to its abnormality [11].”
BackgroundZhao, J., Cui, Z., Fu, J., Shen, M., & Li, Q. (2025) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2025.3582871 · Cited by 5
“To adapt to the adversarial threat landscape, it is imperative to exploit the full potential of GNNs on noisy graphs for robust encryptedmalicious trafficdetectioninadverse environments [13], [14].”
CAE-IF: An Anomaly Detection Approach Based on Temporal Representation of the Reconstruction Error
Geng, H., Zhang, Z., Qian, Y., Yang, B., Ma, Q., Chi, H., Yang, J., & Yin, X. (2025) · Tsinghua Science & Technology · doi:10.26599/tst.2024.9010215 · Cited by 6
Self-citationIL-IDS: an incremental learning approach with confined data streams for intrusion detection
Li, J., Wang, Y., Jia, Y., Zeng, L., Feng, W., Jing, X., Luo, C., Gu, Z., & Fang, B. (2025) · Cybersecurity · doi:10.1186/s42400-025-00359-4 · Cited by 4
“In addition to incremental-based and three-way-deci-sion-based methods, deep learning approaches have gained popularity in intrusion detection (Rizvi et al 2022; Han et al 2021; Du et al 2017; Bansal and Mahapa-tra 2017; Yang et al 2021; Latif et al 2024; Moustakidis and Karlsson 2020).”
CANDICE: An explainable and intelligent framework for network intrusion detection
Li, S., Du, R., Chen, J., He, K., Wu, C., & Feng, Y. (2025) · Future Generation Computer Systems · doi:10.1016/j.future.2025.108059 · Cited by 7
BadDGA: Backdoor Attack on LSTM-Based Domain Generation Algorithm Detector
Zhai, Y., Yang, L., Yang, J., He, L., & Li, Z. (2023) · Electronics · doi:10.3390/electronics12030736 · Cited by 5
Self-citationAn Interpretable Method for Anomaly Detection in Multivariate Time Series Predictions
Tang, S., Ding, Y., & Wang, H. (2025) · Applied Sciences · doi:10.3390/app15137479 · Cited by 3
Feng, J., Zhang, J., Zhang, W., & Han, G. (2024) · Journal of Information Security and Applications · doi:10.1016/j.jisa.2024.103831 · Cited by 2
Xu, S., Fu, C., Li, Q. L., & Xu, K. (2025) · IEEE Transactions on Networking · doi:10.1109/ton.2025.3546735 · Cited by 2
“Note that, existing FP reduction workflows incur huge labor consumption [8], [15], [16], [68], i.e., experts should manually identify FPs, before they can retrain the model using identified FPs [30], [45].”
An ensemble learning framework for enhanced anomaly and failure detection in IoT systems
Bibers, I., & Abdallah, M. (2025) · Cyber Security and Applications · doi:10.1016/j.csa.2025.100105 · Cited by 5
RAAM: A Restricted Adversarial Attack Model with Adding Perturbations to Traffic Features
Sun, P., Li, S., Xie, J., Xu, H., Cheng, Z., & Qin, R. (2022) · IFIP advances in information and communication technology · doi:10.1007/978-3-031-06975-8_8 · Cited by 2
Arreche, O., Guntur, T. R., & Abdallah, M. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4567885 · Cited by 2
A Weakly Supervised Method for Encrypted Traffic Classification in the Dark Web
Han, L., Qian, X., Yang, Z., Yao, Y., Wang, X., & Xu, S. (2024) · Communications in computer and information science · doi:10.1007/978-981-97-5803-6_16 · Cited by 1
Every Sherlock Needs a Watson: Practical Semi-realtime Attack Elaboration System
Umayya, Z., Nandi, A., Roy, A., & Chakravarty, S. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-3531-3_14 · Cited by 1
TeRed: Normal Behavior-Based Efficient Provenance Graph Reduction for Large-Scale Attack Forensics
Li, X., Jiang, X., Wan, H., & Zhao, X. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3601381 · Cited by 1
“Intrusion detection [2], [6], [7], [8], [9], [10], [11] aims to monitor and detect malicious activities within a system by analyzing system events.”
Few-Shot Learning based Anomaly Detection in Security Applications
Sureshan, S., & Das, D. (2023) · Joint International Conference on Data Science & Management of Data (CODS-COMAD) · doi:10.1145/3570991.3571040 · Cited by 1
A Shapley-based Lightweight Global Explainer for Network Intrusion Detection System
Zhang, H., Chen, Y., Liu, W., Zhuang, S., Sun, J., Liu, Y., & Geng, L. (2022) · IEEE International Conference on High Performance Computing and Communications; IEEE International Conference on Smart City; IEEE International Conference on Data Science and Systems (HPCC/SmartCity/DSS) · doi:10.1109/hpcc-dss-smartcity-dependsys57074.2022.00101 · Cited by 1
“Besides, Han et al. [9] proposed DeepAID that can meet several special the security requirements for security applications based on unsupervised deep learning models.”
BackgroundSelf-citationAnomaly Detection in Online Data Streams Using Deep Belief Neural Networks
Talapula, D. K., Kumar, A., Ravulakollu, K. K., & Kumar, M. (2023) · Lecture notes in networks and systems · doi:10.1007/978-981-99-3716-5_59 · Cited by 2
Adjacent Image Augmentation and Its Framework for Self-Supervised Learning in Anomaly Detection
Kwon, G. S., & Choi, Y. S. (2024) · Sensors · doi:10.3390/s24175616 · Cited by 1
GNNexPIDS: An Interpretation Method for Provenance-Based Intrusion Detection Based on GNNExplainer
Yu, Z., Li, W., Ma, X., Zheng, B., Han, X., Li, N., Lv, Q., & Huang, W. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-2417-1_13 · Cited by 3
PGAid: An Interpreter for Provenance-Based Intrusion Detection Based on DeepAid
Yu, Z., Wei, X., Dai, R., Lv, Q., Li, N., & Zhang, M. (2025) · International Conference on Computer Supported Cooperative Work in Design (CSCWD) · doi:10.1109/cscwd64889.2025.11033537 · Cited by 1
“We choose DeepAid [30], GNNEx-plainer [19], and GraphLIME [15] as baselines, corresponding to the IDS interpreter, the general graph interpreter, and the gradient-based graph interpreter, respectively.”
Arreche, O., Guntur, T. R., & Abdallah, M. (2024) · Research Square · doi:10.21203/rs.3.rs-4701709/v1 · Cited by 1
A Comparative Analysis of DNN-based White-Box Explainable AI Methods in Network SecurityPreprint
Arreche, O., & Abdallah, M. (2024) · Research Square · doi:10.21203/rs.3.rs-4790512/v1 · Cited by 1
Building efficient and practical machine learning systems
Hu, Q. (2023) · DR-NTU (Nanyang Technological University) · doi:10.32657/10356/172372
Attack Group Analysis and Identification based on Heterogeneous Graph
Han, T.-H., Hwang, S.-Y., & Lee, T. (2024) · Silicon Valley Cybersecurity Conference (SVCC) · doi:10.1109/svcc61185.2024.10637313
Achieving Counterfactual Explanation for Sequence Anomaly Detection
Cheng, H., Xu, D., Yuan, S., & Wu, X. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-70371-3_2 · Cited by 2
Metadata
- Authors
- Han, D., Wang, Z., Chen, W., Zhong, Y., Wang, S., Zhang, H., Yang, J., Shi, X., & Yin, X.
- Affiliations
- Tsinghua University (CN)
- ORCID
- Venue
- Proceedings of the 2021 ACM SIGSAC Conference on Computer and Communications Security
- Publisher
- ACM
- Published
- 2021-11-12
- Pages
- 3197-3217
- ISSN
- 1543-7221 (ISSN-L)
- OpenAlex
- W3200583622
- Cited by
- 121
- References
- 76
- Citations by year
- 2026: 9 · 2025: 38 · 2024: 42 · 2023: 25 · 2022: 7
- FWCI
- 8.53
- Type
- conference paper
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- ACM SIGSAC Conference on Computer and Communications Security (CCS) (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Key Research and Development Program of China (2018YFB1800200)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Physical Sciences › Computer Science › Artificial Intelligence › Adversarial Robustness in Machine Learning
- Physical Sciences › Computer Science › Artificial Intelligence › Anomaly Detection Techniques and Applications
- Keywords
- anomaly detection, interpretability, unsupervised deep learning, deep neural network, security applications, explainable AI
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.