Pith. sign in

REVIEW 4 major objections 6 minor 52 references

Detecting Zero-Day Web Attacks with an Ensemble of LSTM, GRU, and Stacked Autoencoders

T0 review · 4 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read An ensemble of LSTM, GRU, and stacked autoencoders, trained only on normal web requests, is claimed to detect zero-day web attacks with 97.58% accuracy, 99.76% specificity, and a 0.2% false-positive rate on CSIC2012.

desk verdict The ensemble idea is reasonable, but the reported 97.58% accuracy is mostly a prevalence artifact of a 97.5%-malicious test set, and the decision threshold is tuned on the same data used to report the headline FPR. read the letter →

arxiv 2504.14122 v1 pith:67ME6CAM submitted 2025-04-19 cs.CR

Vahid Babaey , Hamid Reza Faragardi This is my paper · ORCID
classification cs.CR
keywords zero-daywebattacksanomalydetectionautoencoderensembleLSTMGRUstackedtokenizationCSIC2012
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper aims to show that an ensemble of three autoencoders, LSTM, GRU, and stacked, trained exclusively on normal web requests, can detect zero-day web attacks such as SQL injection, cross-site scripting, and buffer overflow without ever seeing an attack sample. The authors report 97.58% accuracy, 97.52% recall, 99.76% specificity, and 99.99% precision on the CSIC2012 benchmark, with a false-positive rate of only 0.2%. A reader should care because a detector that needs no attack labels can be pointed at novel threats with minimal retraining, and a very low false-positive rate means legitimate users are rarely blocked. The evidence comes from six repeated runs on CSIC2012, comparing the ensemble against each sub-model and against earlier CSIC-based detectors.

What carries the argument

The load-bearing mechanism is a one-class reconstruction ensemble. Each sub-model is a four-layer autoencoder with two encoder layers of 50 and 25 units and two symmetric decoder layers, so each compresses a tokenized request into a latent vector. The innovation is that the three latent vectors are not averaged or majority-voted; they are concatenated and squeezed through a 50-unit dense layer, forcing the ensemble to keep only mutually informative features. At test time the decoded request's Mean Absolute Error ($\mathrm{MAE}=\frac{1}{n}\sum_i |\hat{x}_i-x_i|$) is thresholded, and values above the empirically selected threshold of roughly 4.09 are flagged as malicious. The character-class tokenizer is what turns variable-length raw requests into compact, structured numeric sequences the autoencoders can reconstruct.

What would settle it

Run the trained ensemble on a naturally balanced or held-out set of HTTP requests without re-tuning the 4.09 MAE threshold, and record the fraction of normal requests whose reconstruction error exceeds the threshold; if that fraction is substantially above 0.2% or recall falls well below 97.5%, the zero-day detection claim as stated does not generalize.

Watch

Extended reading notes

Core claim

The central claim is that zero-day web attack detection can be treated as a one-class reconstruction problem: the model learns the distribution of normal requests only, and any request whose reconstruction error is too large is declared malicious. The proposed pipeline tokenizes each request by classifying characters into classes such as numeric, lowercase, uppercase, and special, maps the resulting token sequence to numbers, and feeds it to three autoencoders in parallel. Their latent vectors are concatenated and then compressed through a dense layer, and the reconstruction error, measured as Mean Absolute Error, is compared with a threshold of about 4.09. On CSIC2012 the ensemble reports a false-positive rate of 0.2%, meaning 3 of 1,299 normal requests are misclassified, while catching 48,934 of 50,174 malicious requests.

Load-bearing premise

The headline figures rest on a test set where 50,174 of 51,473 requests are malicious and on a reconstruction-error threshold chosen by experimenting on that same data, so the 0.2% false-positive rate may not transfer to a real traffic mix or an independently chosen threshold.

Editorial extensions

If this is right

  • Because no attack labels are needed for training, the same pipeline can be retargeted to a new web application by collecting only its normal request logs.
  • At the reported false-positive rate, only about 0.2% of legitimate requests would be blocked, which is low enough to deploy in front of real users without constant manual review.
  • The concatenate-and-compress ensemble achieves better accuracy, recall, and false-positive rate than any of the three autoencoders alone, suggesting that combining diverse one-class representations is the source of the gain.
  • With roughly 20 seconds of training and 5 seconds of testing on the reported setup, the method is computationally light enough for inline or near-real-time deployment.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The headline accuracy mostly reflects that 50,174 of 51,473 test requests are malicious, so the number that matters in deployment is the 0.2% false-positive rate, and that is the claim most worth stress-testing.
  • A natural extension is to run the fixed threshold of 4.09 on requests from a different dataset such as HTTPParams or FWAF without retuning; if the false-positive rate climbs, the model is overfit to CSIC2012's request grammar rather than to normality in general.
  • The concat-and-compress ensemble suggests a general recipe: train several cheap one-class reconstructors on the same normal data and let a learned compression layer reconcile their disagreements, rather than averaging their scores.
  • Because the threshold appears to be selected after experimenting on the same data that is later scored, a deployment would need the threshold fixed on separate validation traffic before seeing attack data to know whether the metrics transfer.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper proposes an ensemble of LSTM, GRU, and stacked autoencoders for detecting zero-day web attacks, trained exclusively on normal web requests from the CSIC2012 dataset. Input requests are tokenized at word level, mapped to numeric sequences, encoded and decoded by the three autoencoders, and the concatenated latent representations are compressed through a dense layer. Anomaly detection is performed by thresholding the mean absolute error (MAE) between the reconstructed and original request. The authors report accuracy 97.58%, recall 97.52%, specificity 99.76%, precision 99.99%, and a false positive rate of 0.2%, and claim superiority over each individual sub-model and over prior CSIC-based detectors.

Significance. If the reported evaluation were valid, the proposed method would be a practically attractive low-FPR anomaly detector for web traffic, and the tokenization-plus-ensemble-compression idea would be a useful design pattern. The authors also deserve credit for explicitly reporting FPR, which many prior web-attack detection studies omit. However, the evaluation protocol as presented does not support the headline claims because the classification threshold is selected on the same data used to compute the metrics, and because the test set is so imbalanced that accuracy and precision are dominated by the majority class.

major comments (4)
  1. [5.3, Table 2] The test set contains 50,174 malicious and only 1,299 normal requests, so the reported accuracy of 0.9758 is only 0.10 percentage points above the always-malicious baseline of 50,174/51,473 = 0.9748. Similarly, the 99.99% precision is largely a consequence of the 97.5% prior probability of the malicious class. The central claim of 'remarkable detection metrics' is therefore prevalence-driven; the paper should report balanced metrics such as balanced accuracy, PR-AUC, and metrics evaluated under a realistic normal-majority test distribution, and should compare explicitly against the majority-class baseline.
  2. [5, after Eq. (7) and Figure 6] The classification threshold (approximately 4.09) is determined by inspecting the MAE density plot (Figure 6) and by 'iterative experimentation' on the same data that is later used to produce Table 3. No separate validation split is described for threshold selection. Consequently, the reported FPR of 0.2% and recall of 97.52% are fitted quantities rather than unbiased performance estimates. The authors should select the threshold on a held-out validation set (or via nested resampling) and report the resulting test metrics.
  3. [5.1 vs. Table 2] The text states that 'The dataset comprises approximately 16,000 instances labeled as anomalous,' but Table 2 reports 50,174 malicious requests in the test set and a total of 51,473 requests. This inconsistency makes the provenance of the test set unclear and prevents reproducibility. The authors must clarify the exact data split, how the anomalous samples were selected, and reconcile the anomaly count with the reported table.
  4. [5.3] The paper states that 'The reported results represent the average performance obtained over six independent runs of the model,' but Table 3 reports only point estimates with no variance, and no statistical comparison is made against the sub-models. Since the FPR difference between the ensemble (3 false positives) and the LSTM/GRU sub-models (14 false positives) rests on small counts, the authors should report the distribution of metrics across runs and show whether the ensemble advantage is stable.
minor comments (6)
  1. [4.1.1] The sentence 'this method addresses the challenges of training neural network-based models for web security, which arise from the inherent variability in the length and structure of web requests' is repeated almost verbatim in consecutive sentences; please remove the duplication.
  2. [2 and Abstract] The approach is described as 'unsupervised' in Section 2 but as a 'one-class' method in the Abstract; please use consistent terminology.
  3. [Table 1] Table 1 lists MAE as the 'Evaluation Metric for measuring prediction accuracy,' but MAE is a reconstruction error used for anomaly scoring, not an accuracy metric; please reword.
  4. [3, reference [9]] The related work states that Ingham et al. [9] uses 'Transformer models,' but the cited bibliographic entry is 'Learning DFA representations of HTTP' (Computer Networks, 2007); the citation appears mismatched.
  5. [5.2] Please report all hyperparameters for reproducibility, including the exact number of units in the final dense compression layer, the sequence padding length, and the random seeds used; the current architecture description omits some of these details.
  6. [Figure 9] The caption relies on color ('blue points', 'red line'); please ensure the figure is legible in grayscale or add markers.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline metrics are threshold-fitted: the MAE threshold is chosen by iterative experimentation on the same MAE distribution later used to compute Table 3, so FPR/accuracy are selection results rather than independent predictions.

  1. fitted input called prediction [Section 5 (Evaluation and Results), after Eq. (7); Figure 6; Tables 2 and 3]
    "The threshold value is determined using Figure 6, which visualizes the density distribution of web requests based on the MAE metric. ... The optimal threshold is determined empirically through iterative experimentation. The analysis suggests that a threshold value of 4.09 provides optimal detection performance, as requests with an MAE of 5 or higher are observed infrequently. However, setting the threshold too high (e.g., at 7) may improve training results but could fail to detect certain malicious requests with reconstruction errors in the range of 5 to 7."

    The classification rule is exactly 'MAE above threshold => malicious' (Eq. 7 and surrounding text). The threshold (about 4.09) is selected by inspecting the MAE density of the very requests whose labels are then counted in Table 2 and reported as Table 3 metrics, with no separate validation split for threshold tuning. Consequently, the headline 97.58% accuracy, 99.76% specificity, 99.99% precision, and 0.2% FPR are evaluations of a fitted decision boundary: they measure how well the chosen threshold separates the observed MAE distribution, not how the model would perform with a pre-specified or held-out threshold. The reported numbers are therefore a selection result, statistically forced by the threshold search, rather than an independent prediction of zero-day detection performance.

full rationale

The only substantiated circularity is threshold fitting to the evaluation data. The autoencoder ensemble, tokenization, and normal-only training protocol are not circular: the reconstruction error of a request is computed from the model, and classifying by reconstruction error is a standard anomaly-detection paradigm. No load-bearing self-citation was found; references [48,49] are prior work by the authors but are not used to justify the central detection claim. The dataset inconsistency (Section 5.1 says about 16,000 anomalous samples while Table 2 lists 50,174 positives) and the 97.5%-malicious test-set composition are serious validity concerns, but they are correctness/robustness issues rather than definitional circularity. The circular step is that the decision threshold is chosen by iterative experimentation on the same MAE distribution used to produce the reported FPR, recall, and accuracy, so the headline metrics are partly a consequence of the threshold search. This warrants a score of 6: one reported 'prediction' (the detection metrics) reduces by construction to a parameter fitted on the evaluation data.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The central quantitative claims hinge on the threshold fitted to the MAE distribution and on the benchmark dataset standing in for real zero-day traffic. No new physical or conceptual entities are introduced; the 'novel tokenization' is a set of character classes rather than an invented entity.

free parameters (3)
  • MAE classification threshold = 4.09
    Selected by inspecting the MAE density distribution and by 'iterative experimentation' on the evaluation data; the reported FPR, accuracy, recall, and specificity all depend on this value.
  • Autoencoder layer sizes and dense compression size = 50/25/25/50 and 50
    Chosen by hand with no ablation study; the architecture dimensions are design choices that affect reconstruction quality and ensemble performance.
  • Sequence padding length = not specified
    All requests are padded to a fixed length, but the actual length is not reported; this preprocessing choice affects what the model sees.
assumptions (3)
  • domain assumption CSIC2012 normal requests are representative of all normal web traffic for the deployment scenario.
    The paper trains and evaluates only on CSIC2012 and generalizes to real-world zero-day detection without testing other traffic distributions (Sections 5.1 and 6.2).
  • domain assumption The malicious requests in CSIC2012 can stand in for zero-day attacks even though they are known attack types relabeled for the experiment.
    Section 1 states SQLi, XSS, and buffer overflow are 'treated as zero-day attacks within the dataset'; no truly novel attack traffic is used.
  • ad hoc to paper Reconstruction MAE is a sufficient statistic for anomaly detection with an arbitrary threshold.
    The MAE threshold is set empirically from the data distribution (Figure 6), with no principled calibration or statistical justification.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Detecting Zero-Day Web Attacks with an Ensemble of LSTM, GRU, and Stacked Autoencoders." pith.science (2026). https://pith.science/paper/67ME6CAM

@misc{pith2026250414122,
  author       = {Pith},
  title        = {Pith review of: Detecting Zero-Day Web Attacks with an Ensemble of LSTM, GRU, and Stacked Autoencoders},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/67ME6CAM}},
  note         = {Machine review of arXiv:2504.14122}
}
read the original abstract

The rapid growth in web-based services has significantly increased security risks related to user information, as web-based attacks become increasingly sophisticated and prevalent. Traditional security methods frequently struggle to detect previously unknown (zero-day) web attacks, putting sensitive user data at significant risk. Additionally, reducing human intervention in web security tasks can minimize errors and enhance reliability. This paper introduces an intelligent system designed to detect zero-day web attacks using a novel one-class ensemble method consisting of three distinct autoencoder architectures: LSTM autoencoder, GRU autoencoder, and stacked autoencoder. Our approach employs a novel tokenization strategy to convert normal web requests into structured numeric sequences, enabling the ensemble model to effectively identify anomalous activities by uniquely concatenating and compressing the latent representations from each autoencoder. The proposed method efficiently detects unknown web attacks while effectively addressing common limitations of previous methods, such as high memory consumption and excessive false positive rates. Extensive experimental evaluations demonstrate the superiority of our proposed ensemble, achieving remarkable detection metrics: 97.58% accuracy, 97.52% recall, 99.76% specificity, and 99.99% precision, with an exceptionally low false positive rate of 0.2%. These results underscore our method's significant potential in enhancing real-world web security through accurate and reliable detection of web-based attacks.

Figures

Figures reproduced from arXiv: 2504.14122 by the authors.

Figure 1
Figure 1. The proposed model in the training phase During the training phase, the model is exclusively trained on normal web requests to establish a baseline pattern of legitimate traffic. In this phase, the model has full access to the training dataset, allowing it to learn the distribution of normal request patterns [35]. Conversely, in the testing phase, both normal and malicious web requests are input into the model for e… view at source ↗
Figure 2
Figure 2. The proposed model in the test phase 4.1.1. Tokenization A key innovation of the proposed model is the introduction of a novel tokenization technique for web requests, applied at the word level to both normal and malicious in￾puts [36]. Due to the inherent variability in the length and structure of web requests, this method addresses the challenges of training neural network-based models for web secu￾rity, which ari… view at source ↗
Figure 3
Figure 3. A Tokenized request. 4.1.2. Numerical Sequence Following the tokenization of web requests on a word-by-word basis, each token must be mapped to a corresponding numerical value, as neural networks operate on numbers rather than raw text. This transformation is a critical step, as the varying range of input features necessitates data scaling before being processed by the model [39]. The tokenized text is converted int… view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Mapping words to numbers [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Structure of the proposed model. The primary objective of the decoding process is to validate the quality and representa￾tiveness of the extracted features. The outputs from all three autoencoders are subsequently concatenated to form a unified feature representation, …
Figure 6
Figure 6. Figure 6: Density diagram according to MAE of the proposed model. 5.1. Data Collection Previous research on detecting malicious and zero-day web requests has primarily relied on two well-established datasets: CSIC [42] and HTTPPARAMS [42]. These datasets provide a comprehensive …
Figure 7
Figure 7. Figure 7: Ensemble model architecture. The evaluation results of the proposed model, as depicted in [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]
Figure 8
Figure 8. Figure 8: Training Process of the Proposed Model Based on MAE and Number of Epochs [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: Comparison of the Calculated MAE Value for Each Web Request with the Threshold Value. reducing false positives, albeit at the expense of slightly lower accuracy and recall when used independently [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: Comparison of the Performance of the Proposed Model with the Sub-models in the form of a bar chart. form of a bar chart. In the figure, models that utilized the CSIC2012 dataset are highlighted in red. The primary comparison focuses on studies that have employed the C…
Figure 11
Figure 11. Figure 11: Comparison of the Performance of the Proposed Model with Previously Designed Models in the form of a bar chart. 6. Discussion According to the obtained results, the proposed model demonstrates strong perfor￾mance across all evaluation metrics, particularly in terms of…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 46 canonical work pages

  1. [1]

    Zero-day attack detection: a systematic literature review

    Ahmad, R.; Alsmadi, I.; Alhamdani, W.; Tawalbeh, L. Zero-day attack detection: a systematic literature review. Artificial Intelligence Review 2023, 56, 10733–10811

    work page 2023

  2. [2]

    Deep learning technique-enabled web application firewall for the detection of web attacks

    Dawadi, B.R.; Adhikari, B.; Srivastava, D.K. Deep learning technique-enabled web application firewall for the detection of web attacks. Sensors 2023, 23, 2073

    work page 2023

  3. [3]

    A systematic literature review of information security in chatbots.Applied Sciences 2023, 13, 6355

    Yang, J.; Chen, Y.L.; Por, L.Y.; Ku, C.S. A systematic literature review of information security in chatbots.Applied Sciences 2023, 13, 6355

    work page 2023

  4. [4]

    Machine learning for web vulnerability detection: the case of cross-site request forgery

    Calzavara, S.; Conti, M.; Focardi, R.; Rabitti, A.; Tolomei, G. Machine learning for web vulnerability detection: the case of cross-site request forgery. IEEE Security & Privacy 2020, 18, 8–16

    work page 2020

  5. [5]

    Investigating the Impact of Heuristic Algorithms on Cyberthreat Detection

    Kalla, D.; Mohammed, A.S.; Boddapati, V .N.; Jiwani, N.; Kiruthiga, T. Investigating the Impact of Heuristic Algorithms on Cyberthreat Detection. In Proceedings of the 2024 2nd International Conference on Advances in Computation, Communication and Information Technology (ICAICCIT), 2024, Vol. 1, pp. 450–455

    work page 2024

  6. [6]

    A survey of network anomaly detection techniques

    Ahmed, M.; Mahmood, A.N.; Hu, J. A survey of network anomaly detection techniques. Journal of Network and Computer Applications 2016, 60, 19–31

    work page 2016

  7. [7]

    A hybrid unsupervised clustering-based anomaly detection method

    Pu, G.; Wang, L.; Shen, J.; Dong, F. A hybrid unsupervised clustering-based anomaly detection method. Tsinghua Science and Technology 2020, 26, 146–153

    work page 2020

  8. [8]

    An efficient algorithm and tool for detecting dangerous website vulnerabilities

    Long, H.V .; Tuan, T.A.; Taniar, D.; Can, N.V .; Hue, H.M.; Son, N.T.K. An efficient algorithm and tool for detecting dangerous website vulnerabilities. International Journal of Web and Grid Services 2020, 16, 81–104. Version April 22, 2025 submitted to Journal Not Specified 21 of 22

    work page 2020

Show all 52 references
  1. [9]

    Learning DFA representations of HTTP for protecting web applications.Computer Networks 2007, 51, 1239–1255

    Ingham, K.L.; Somayaji, A.; Burge, J.; Forrest, S. Learning DFA representations of HTTP for protecting web applications.Computer Networks 2007, 51, 1239–1255

    2007

  2. [10]

    Web intrusion detection using character level machine learning approaches with upsampled data

    Sivri, T.T.; Akman, N.P .; Berkol, A.; Peker, C. Web intrusion detection using character level machine learning approaches with upsampled data. Annals of Computer Science and Information Systems 2022, 32

    2022

  3. [11]

    PF-TL: Payload feature-based transfer learning for dealing with the lack of training data

    Jung, I.; Lim, J.; Kim, H.K. PF-TL: Payload feature-based transfer learning for dealing with the lack of training data. Electronics 2021, 10, 1148

    2021

  4. [12]

    An anomaly detection method to detect web attacks using stacked auto-encoder

    Vartouni, A.M.; Kashi, S.S.; Teshnehlab, M. An anomaly detection method to detect web attacks using stacked auto-encoder. In Proceedings of the 2018 6th Iranian Joint Congress on Fuzzy and Intelligent Systems (CFIS). IEEE, 2018, pp. 131–134

    2018

  5. [13]

    HMMPayl: An intrusion detection system based on Hidden Markov Models

    Ariu, D.; Tronci, R.; Giacinto, G. HMMPayl: An intrusion detection system based on Hidden Markov Models. computers & security 2011, 30, 221–241

    2011

  6. [14]

    Anomaly-based web attack detection: a deep learning approach

    Liang, J.; Zhao, W.; Ye, W. Anomaly-based web attack detection: a deep learning approach. In Proceedings of the Proceedings of the 2017 VI International Conference on Network, Communication and Computing, 2017, pp. 80–85

    2017

  7. [15]

    DeepWAF: detecting web attacks based on CNN and LSTM models

    Kuang, X.; Zhang, M.; Li, H.; Zhao, G.; Cao, H.; Wu, Z.; Wang, X. DeepWAF: detecting web attacks based on CNN and LSTM models. In Proceedings of the Cyberspace Safety and Security: 11th International Symposium, CSS 2019, Guangzhou, China, December 1–3, 2019, Proceedings, Part ...

    2019

  8. [16]

    Zerowall: Detecting zero-day web attacks through encoder-decoder recurrent neural networks

    Tang, R.; Yang, Z.; Li, Z.; Meng, W.; Wang, H.; Li, Q.; Sun, Y.; Pei, D.; Wei, T.; Xu, Y.; et al. Zerowall: Detecting zero-day web attacks through encoder-decoder recurrent neural networks. In Proceedings of the IEEE INFOCOM 2020-IEEE Conference on Computer Communications. IEE...

    2020

  9. [17]

    Robust ensemble machine learning model for filtering phishing URLs: Expandable random gradient stacked voting classifier (ERG-SVC)

    Indrasiri, P .L.; Halgamuge, M.N.; Mohammad, A. Robust ensemble machine learning model for filtering phishing URLs: Expandable random gradient stacked voting classifier (ERG-SVC). Ieee Access 2021, 9, 150142–150161

    2021

  10. [18]

    Model uncertainty based annotation error fixing for web attack detection

    Gong, X.; Lu, J.; Zhou, Y.; Qiu, H.; He, R. Model uncertainty based annotation error fixing for web attack detection. Journal of Signal Processing Systems 2021, 93, 187–199

    2021

  11. [19]

    A novel architecture for web-based attack detection using convolutional neural network

    Tekerek, A. A novel architecture for web-based attack detection using convolutional neural network. Computers & Security 2021, 100, 102096

    2021

  12. [20]

    SWAF: a smart web application firewall based on convolutional neural network

    Jemal, I.; Haddar, M.A.; Cheikhrouhou, O.; Mahfoudhi, A. SWAF: a smart web application firewall based on convolutional neural network. In Proceedings of the 2022 15th International Conference on Security of Information and Networks (SIN). IEEE, 2022, pp. 01–06

    2022

  13. [21]

    Web attacks detection using stacked generalization ensemble for LSTMs and word embedding

    Alaoui, R.L.; et al. Web attacks detection using stacked generalization ensemble for LSTMs and word embedding. Procedia Computer Science 2022, 215, 687–696

    2022

  14. [22]

    MC-MLDCNN: Multichannel Multilayer Dilated Convolutional Neural Networks for Web Attack Detection

    Moarref, N.; Sandıkkaya, M.T. MC-MLDCNN: Multichannel Multilayer Dilated Convolutional Neural Networks for Web Attack Detection. Security and Communication Networks 2023, 2023, 2415288

    2023

  15. [23]

    A Static Detection Method for SQL Injection Vulnerability Based on Program Transformation

    Yuan, Y.; Lu, Y.; Zhu, K.; Huang, H.; Yu, L.; Zhao, J. A Static Detection Method for SQL Injection Vulnerability Based on Program Transformation. Applied Sciences 2023, 13

    2023

  16. [24]

    Synthesis of Allowlists for Runtime Protection against SQLi

    Vorobyov, K.; Gauthier, F.; Krishnan, P . Synthesis of Allowlists for Runtime Protection against SQLi. In Proceedings of the Proceedings of the 2024 ACM/IEEE 44th International Conference on Software Engineering: New Ideas and Emerging Results. Association for Computing Machin...

    2024

  17. [25]

    Splendor: Static Detection of Stored XSS in Modern Web Applications

    Su, H.; Li, F.; Xu, L.; Hu, W.; Sun, Y.; Sun, Q.; Chao, H.; Huo, W. Splendor: Static Detection of Stored XSS in Modern Web Applications. In Proceedings of the Proceedings of the 32nd ACM SIGSOFT International Symposium on Software Testing and Analysis. Association for Computin...

    2023

  18. [26]

    Towards a SQL Injection Vulnerability Detector Based on Session Types

    Silvestre, A.; Medeiros, I.; Mordido, A. Towards a SQL Injection Vulnerability Detector Based on Session Types. In Proceedings of the Proceedings of the 19th International Conference on Evaluation of Novel Approaches to Software Engineering - Volume 1: ENASE. INSTICC, SciTePre...

    2024

  19. [27]

    Towards a Zero-Day Anomaly Detector in Cyber Physical Systems Using a Hybrid VAE-LSTM-OCSVM Model

    Yatagha, R.; Nebebe, B.; Waedt, K.; Ruland, C. Towards a Zero-Day Anomaly Detector in Cyber Physical Systems Using a Hybrid VAE-LSTM-OCSVM Model. In Proceedings of the Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, 2024, pp. 5038–5045

    2024

  20. [28]

    One-class IoT anomaly detection system using an improved interpolated deep SVDD autoencoder with adversarial regularizer

    Katbi, A.; Ksantini, R. One-class IoT anomaly detection system using an improved interpolated deep SVDD autoencoder with adversarial regularizer. Digital Signal Processing 2025, p. 105153

    2025

  21. [29]

    Stacking an autoencoder for feature selection of zero-day threats

    Tokmak, M.; Nkongolo, M. Stacking an autoencoder for feature selection of zero-day threats. arXiv preprint arXiv:2311.00304 2023

    2023 arXiv

  22. [30]

    Deep learning architecture for detecting SQL injection attacks based on RNN autoencoder model

    Alghawazi, M.; Alghazzawi, D.; Alarifi, S. Deep learning architecture for detecting SQL injection attacks based on RNN autoencoder model. Mathematics 2023, 11, 3286

    2023

  23. [31]

    Ae-net: Novel autoencoder-based deep features for sql injection attack detection

    Thalji, N.; Raza, A.; Islam, M.S.; Samee, N.A.; Jamjoom, M.M. Ae-net: Novel autoencoder-based deep features for sql injection attack detection. IEEE Access 2023, 11, 135507–135516

    2023

  24. [32]

    A lightweight intelligent network intrusion detection system using one-class autoencoder and ensemble learning for IoT

    Yao, W.; Hu, L.; Hou, Y.; Li, X. A lightweight intelligent network intrusion detection system using one-class autoencoder and ensemble learning for IoT. Sensors 2023, 23, 4141

    2023

  25. [33]

    Multi-Class Intrusion Detection System using Deep Learning

    Mohamed, S.M.; Rohaim, M.A. Multi-Class Intrusion Detection System using Deep Learning. Journal of Al-Azhar University Engineering Sector 2023, 18, 869–883

    2023

  26. [34]

    An enhanced deep learning based framework for web attacks detection, mitigation and attacker profiling

    Shahid, W.B.; Aslam, B.; Abbas, H.; Khalid, S.B.; Afzal, H. An enhanced deep learning based framework for web attacks detection, mitigation and attacker profiling. Journal of Network and Computer Applications 2022, 198, 103270. Version April 22, 2025 submitted to Journal Not S...

    2022

  27. [35]

    Dynamic defenses and the transferability of adversarial examples

    Thomas, S.; Koleini, F.; Tabrizi, N. Dynamic defenses and the transferability of adversarial examples. In Proceedings of the 2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA). IEEE, 2022, pp. 276–284

    2022

  28. [36]

    Predicting web vulnerabilities in web applications based on machine learning

    Khalid, M.N.; Farooq, H.; Iqbal, M.; Alam, M.T.; Rasheed, K. Predicting web vulnerabilities in web applications based on machine learning. In Proceedings of the Intelligent Technologies and Applications: First International Conference, INTAP 2018, Bahawalpur, Pakistan, October...

    2018

  29. [37]

    Learning web request patterns

    Levene, M.; Poulovassilis, A.; Davison, B.D. Learning web request patterns. Web Dynamics: Adapting to Change in Content, Size, Topology and Use 2004, pp. 435–459

    2004

  30. [38]

    Text mining: open source tokenization tools-an analysis

    Vijayarani, S.; Janani, R.; et al. Text mining: open source tokenization tools-an analysis. Advanced Computational Intelligence: An International Journal (ACII) 2016, 3, 37–47

    2016

  31. [39]

    Real-Time Bus Arrival Prediction: A Deep Learning Approach for Enhanced Urban Mobility

    Rashvand, N.; Hosseini, S.S.; Azarbayjani, M.; Tabkhi, H. Real-Time Bus Arrival Prediction: A Deep Learning Approach for Enhanced Urban Mobility. arXiv preprint arXiv:2303.15495 2023

    2023 arXiv

  32. [40]

    Attention is all you need

    Vaswani, A. Attention is all you need. Advances in Neural Information Processing Systems 2017

    2017

  33. [41]

    Enhancing automatic modulation recognition for iot applications using transformers

    Rashvand, N.; Witham, K.; Maldonado, G.; Katariya, V .; Marer Prabhu, N.; Schirner, G.; Tabkhi, H. Enhancing automatic modulation recognition for iot applications using transformers. IoT 2024, 5, 212–226

    2024

  34. [42]

    Web application firewall using machine learning and features engineering

    Shaheed, A.; Kurdy, M.B. Web application firewall using machine learning and features engineering. Security and Communication Networks 2022, 2022, 5280158

    2022

  35. [43]

    CNN Web Application Firewall

    DuckDuckBug. CNN Web Application Firewall. https://github.com/DuckDuckBug/cnn_waf, 2023. Accessed: January 29, 2025

    2023

  36. [44]

    Detecting web attacks from HTTP weblogs using variational LSTM autoencoder deviation network

    Jagat, R.R.; Sisodia, D.S.; Singh, P . Detecting web attacks from HTTP weblogs using variational LSTM autoencoder deviation network. IEEE Transactions on Services Computing 2024

    2024

  37. [45]

    LLM-assisted Physical Invariant Extraction for Cyber-Physical Systems Anomaly Detection

    Abshari, D.; Fu, C.; Sridhar, M. LLM-assisted Physical Invariant Extraction for Cyber-Physical Systems Anomaly Detection. arXiv preprint arXiv:2411.10918 2024

    2024 arXiv

  38. [46]

    A comprehensive survey on the security of smart grid: Challenges, mitigations, and future research opportunities

    Zibaeirad, A.; Koleini, F.; Bi, S.; Hou, T.; Wang, T. A comprehensive survey on the security of smart grid: Challenges, mitigations, and future research opportunities. arXiv preprint arXiv:2407.07966 2024

    2024 arXiv

  39. [47]

    A Survey of Anomaly Detection in Cyber-Physical Systems

    Abshari, D.; Sridhar, M. A Survey of Anomaly Detection in Cyber-Physical Systems. arXiv preprint arXiv:2502.13256 2025

    2025 arXiv

  40. [48]

    GenSQLi: A Generative Artificial Intelligence Framework for Automatically Securing Web Application Firewalls Against Structured Query Language Injection Attacks

    Babaey, V .; Ravindran, A. GenSQLi: A Generative Artificial Intelligence Framework for Automatically Securing Web Application Firewalls Against Structured Query Language Injection Attacks. Future Internet 2025, 17, 8

    2025

  41. [49]

    GenXSS: an AI-Driven Framework for Automated Detection of XSS Attacks in WAFs.arXiv preprint arXiv:2504.08176 2025

    Babaey, V .; Ravindran, A. GenXSS: an AI-Driven Framework for Automated Detection of XSS Attacks in WAFs.arXiv preprint arXiv:2504.08176 2025

    2025 arXiv

  42. [50]

    A prompt pattern catalog to enhance prompt engineering with chatgpt

    White, J.; Fu, Q.; Hays, S.; Sandborn, M.; Olea, C.; Gilbert, H.; Elnashar, A.; Spencer-Smith, J.; Schmidt, D.C. A prompt pattern catalog to enhance prompt engineering with chatgpt. arXiv preprint arXiv:2302.11382 2023

    2023 arXiv

  43. [51]

    Hybrid speech recognition with deep bidirectional LSTM

    Graves, A.; Jaitly, N.; Mohamed, A.r. Hybrid speech recognition with deep bidirectional LSTM. In Proceedings of the 2013 IEEE workshop on automatic speech recognition and understanding. IEEE, 2013, pp. 273–278

    2013

  44. [52]

    Reasoning with LLMs for Zero-Shot Vulnerability Detection

    Zibaeirad, A.; Vieira, M. Reasoning with LLMs for Zero-Shot Vulnerability Detection. arXiv preprint arXiv:2503.17885 2025

    2025 arXiv

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.