Pith. sign in

REVIEW 3 major objections 7 minor 1 cited by

A Novel Zero-Touch, Zero-Trust, AI/ML Enablement Framework for IoT Network Security

T0 review · 3 major / 7 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read The paper proposes that integrating Zero-Trust authentication, Zero-Touch provisioning, and AI/ML threat detection yields a scalable, automated defense against DDoS attacks in 5G/6G IoT networks.

desk verdict A coherent but unvalidated framework proposal bolted onto an unreproducible ML benchmark; the central claim of end-to-end protection never gets tested. read the letter →

arxiv 2502.03614 v1 pith:43XMVLFY submitted 2025-02-05 cs.LG cs.AIcs.CR

Sushil Shakya , Robert Abbas , Sasa Maric This is my paper · ORCID
classification cs.LGcs.AIcs.CR
keywords IoTsecurityDDoSdetectionZeroTrustTouchprovisioningmachinelearningXGBoost5G/6Gnetworksnetworkframework
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 argues that securing the Internet of Things in 5G/6G networks requires combining three mechanisms usually deployed separately: Zero-Touch provisioning to onboard devices automatically, Zero-Trust verification to authenticate every request, and AI/ML threat detection to spot attacks in real time. It proposes an integrated framework built from those three components and reports a comparative evaluation of five machine learning classifiers for DDoS detection. The authors claim the framework would provide end-to-end protection, covering secure onboarding, anomaly detection, and automated response such as blackholing malicious traffic. A sympathetic reader would care because the architecture targets a concrete pain point: the manual configuration and implicit trust that leave large IoT fleets exposed to botnet-driven DDoS attacks.

What carries the argument

The carrying object is the three-layer architecture: a Zero-Trust security layer that authenticates and verifies all traffic, an AI/ML-driven detection layer with XGBoost as the flagship model, and a Zero-Touch automation layer that onboards devices and executes responses without human intervention. These are tied together by an operational workflow that moves from input, to zero-trust analysis, to AI/ML threat detection, to zero-touch response, and finally to output of benign traffic. The ML evaluation uses standard machinery — Min-Max normalization, Gini impurity, Euclidean distance, gradient-boosted objective with regularization, and ROC-AUC — to compare the five models.

What would settle it

Run the same five models on a publicly available current IoT DDoS dataset in an online streaming setting where authentication and automated response are actually enforced, then compare per-class recall and detection latency against the paper's reported values; if XGBoost's accuracy drops materially or the ensemble advantage disappears, the claim that the framework provides end-to-end protection at the reported level is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that the integration of Zero-Trust principles, Zero-Touch provisioning, and AI/ML-based threat detection yields a scalable, automated defense for IoT networks in 5G/6G environments, and specifically that this framework can detect and mitigate DDoS attacks end to end. The evidence offered is a comparison of five classifiers — XGBoost, Random Forest, K-Nearest Neighbors, Stochastic Gradient Descent, and Naive Bayes — on a labeled, anonymized IoT traffic dataset, with XGBoost reporting the highest accuracy, precision, recall, F1 score, and ROC-AUC. The framework itself is described as a workflow: incoming traffic passes zero-trust verification, is scored by AI/ML models, and triggers automated zero-touch responses such as isolation and blackholing when threats are found.

Load-bearing premise

The load-bearing premise is that the unnamed anonymized IoT traffic dataset used to score the classifiers represents real 5G/6G IoT DDoS traffic, so the offline accuracy of 99.82% would carry over to the proposed live, real-time zero-trust and zero-touch pipeline.

Editorial extensions

If this is right

  • If the framework works as described, new IoT devices can be onboarded automatically with secure boot and identity verification, removing a major source of configuration errors.
  • Ensemble classifiers such as XGBoost would be able to distinguish DDoS traffic from benign traffic with the reported 99.82% accuracy, making automated blackholing feasible in real time.
  • The combination of network slicing and micro-segmentation would limit lateral movement, so a compromised device cannot roam freely inside the network.
  • The framework's continuous learning would give it a defense against zero-day attacks, something signature-based intrusion detection cannot do.
  • By reducing manual intervention, the architecture scales to the very large device counts expected in 5G/6G IoT deployments.

Reading between the lines

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

  • The paper leaves implicit that the reported classifier scores are offline and static; nothing in the experiments tests the latency or throughput of the zero-trust and zero-touch stages, so the real-time claim rests on an assumed transfer from offline scoring to live operation.
  • A consequence the authors do not spell out is that in a deployed framework the ML classifier becomes a security control, so adversaries will target it with adversarial traffic; robustness testing and drift monitoring, not just accuracy, would be needed before deployment.
  • The comparative results suggest that once ensemble methods reach roughly 99% accuracy, further gains depend more on data quality and feature engineering than on choosing among the top models, pointing toward dataset curation as the next bottleneck.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. The paper proposes an integrated IoT security framework that combines Zero-Trust principles, Zero-Touch provisioning, and AI/ML-based threat detection, aimed at DDoS detection, mitigation, and prevention in 5G/6G IoT environments. The authors describe a conceptual architecture and workflow (Section III) and report a comparative evaluation of five machine learning classifiers (XGBoost, Random Forest, KNN, SGD, Naïve Bayes) on an unnamed, anonymized labeled IoT traffic dataset (Sections II-E and IV). The results favor ensemble methods, with XGBoost achieving the highest reported accuracy (99.82%) and AUC (0.9997). The paper concludes that the framework offers scalable, adaptive, proactive security for IoT networks and lists future work items such as federated learning and explainability.

Significance. If the integrated framework were validated end-to-end, the idea of combining Zero-Trust, Zero-Touch, and ML-based detection for IoT/5G/6G DDoS defense would be a useful architectural contribution. The paper also provides a straightforward comparison of five standard classifiers, and the reported metrics are internally consistent with each other. However, the significance is severely limited because the central framework claim is not evaluated at all, the dataset is not identified, and the ML experiments are detached from the proposed architecture. The paper does not ship machine-checked proofs, reproducible code, or any end-to-end system measurement. The architectural idea is plausible but remains a conceptual sketch rather than a demonstrated result.

major comments (3)
  1. [Section IV and Table I] The experimental evaluation lacks the information needed to verify or reproduce the reported results. Section II-E identifies the data only as "a labeled dataset of IoT network traffic that had been anonymized," without naming the dataset, the number of samples, the class balance, or the attack types. Section IV reports accuracy, precision, recall, F1, and AUC for each model, but gives no train/test split, cross-validation procedure, hyperparameter settings (e.g., k in KNN, learning rate in SGD, number of trees in Random Forest), or error bars. Consequently, the values in Table I, including the headline XGBoost accuracy of 99.82%, cannot be verified or benchmarked against prior work, and the abstract's claim that ensemble models are best for detecting and mitigating DDoS vectors is not substantiated.
  2. [Section III and Section IV] The central claim of the paper, stated in Section III, is that the integrated Zero-Trust Zero-Touch AI/ML-enabled framework "will ensure the implementation of end-to-end protection" in IoT networks. This claim is not evaluated. Section III-B describes an operational workflow (input, zero-trust analysis, AI/ML threat detection, zero-touch response, output) but provides no algorithms, state machines, interface definitions, or latency characteristics. The ML results in Section IV are produced by standalone classifiers on an offline dataset; no experiment sends traffic through the Section III-B pipeline, measures end-to-end detection or response time, or compares the integrated framework with a baseline without Zero-Trust/Zero-Touch components. The load-bearing link between the proposed integration and the reported classifier metrics is therefore missing.
  3. [Abstract and Section V] The abstract states that the paper compares models for "detecting and mitigating different DDoS vectors," and Section V claims "proactive threat mitigation" through automated responses such as blackholing. However, the reported metrics are only binary classification scores (accuracy, precision, recall, F1, AUC) on a single labeled dataset. There is no measurement of mitigation actions, no response latency, no evaluation of the effect on benign traffic, and no evidence of prevention. The term "mitigating" in the abstract overstates what the experimental results demonstrate; the paper evaluates detection only, and even that is done in isolation from the proposed framework.
minor comments (7)
  1. [Abstract] The model name "Native Bayes" should be "Naïve Bayes" to match the body of the paper.
  2. [Section II-C] The heading "Zero-Trust Security and Network Sclicing" contains a typo: "Sclicing" should be "Slicing."
  3. [Section III] The opening sentence of Section III contains a duplicated "of" ("implementation of of end-to-end protection").
  4. [Section V] In the first bullet under "Key points," "Zerou Trust" should be "Zero Trust."
  5. [Section II-G and Eq. (11)] The definition of AUC as an integral of TPR over FPR is mathematically imprecise for discrete classifiers and is not used to explain how the AUC values in Table I were computed; the authors should state the estimator actually used (e.g., Mann-Whitney U).
  6. [References] Reference [3] contains the placeholder arXiv number "arXiv:2009.00000" and should be corrected or removed; the same applies to other references with incomplete or implausible bibliographic details, such as [8] and [21].
  7. [Figures] Figure references are inconsistent: the text mentions Fig. 3, Fig. 4, and Fig. 5 for the workflow and architecture diagrams, which may confuse readers if the figures are not numbered in the order of appearance.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the classifier metrics are empirical, the framework is architectural, and no prediction reduces by construction to a fitted input.

full rationale

I examined the derivation chain. Section II-E describes an unnamed labeled IoT dataset; Section II-F defines five ML models; Section II-G gives standard evaluation metrics; Section IV reports empirical classifier performance; Section III presents a conceptual framework architecture. No equation defines a predicted quantity in terms of a fitted parameter, and the reported accuracy, precision, recall, F1, and AUC values are measured results, not predictions derived from the framework. The paper's assertion that the integrated framework 'will ensure end-to-end protection' is not derived from the classifier scores by construction; it is an unsupported extrapolation, which is an external-validity or completeness gap rather than a circularity. The only self-citation, reference [21] in Section II-F, supports the choice of five models and is not load-bearing: the model set is conventional and the results are empirically evaluated in this paper, not justified solely by that citation. No uniqueness theorem or ansatz is imported from prior work, and no known result is renamed as a framework. Therefore the derivation chain is self-contained with respect to circularity.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The load-bearing premises are the representativeness of the undisclosed dataset, the feasibility of integrating the described components as specified, and the applicability of standard ML formulas. The paper provides no evidence for the first two. No new entities are posited; zero trust, zero touch, and AI/ML are existing concepts.

free parameters (1)
  • ML model hyperparameters (e.g., k in KNN, learning rate in SGD)
    No hyperparameter values or tuning procedure are given, yet the reported accuracies depend on them.
assumptions (3)
  • domain assumption The anonymized IoT traffic dataset is representative of real IoT DDoS traffic and correctly labeled.
    Section II-E describes the dataset only as a labeled and anonymized IoT traffic dataset, with no name, provenance, or labeling audit.
  • ad hoc to paper The three framework components (zero trust, zero touch, AI/ML detection) can be integrated as described without additional specification.
    Section III describes the architecture conceptually but provides no interface definitions, protocols, or performance analysis.
  • standard math Standard ML equations and evaluation metrics are accepted as background.
    Equations (1) through (11) are textbook definitions; no proof is needed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Novel Zero-Touch, Zero-Trust, AI/ML Enablement Framework for IoT Network Security." pith.science (2026). https://pith.science/paper/43XMVLFY

@misc{pith2026250203614,
  author       = {Pith},
  title        = {Pith review of: A Novel Zero-Touch, Zero-Trust, AI/ML Enablement Framework for IoT Network Security},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/43XMVLFY}},
  note         = {Machine review of arXiv:2502.03614}
}
read the original abstract

The IoT facilitates a connected, intelligent, and sustainable society; therefore, it is imperative to protect the IoT ecosystem. The IoT-based 5G and 6G will leverage the use of machine learning and artificial intelligence (ML/AI) more to pave the way for autonomous and collaborative secure IoT networks. Zero-touch, zero-trust IoT security with AI and machine learning (ML) enablement frameworks offers a powerful approach to securing the expanding landscape of Internet of Things (IoT) devices. This paper presents a novel framework based on the integration of Zero Trust, Zero Touch, and AI/ML powered for the detection, mitigation, and prevention of DDoS attacks in modern IoT ecosystems. The focus will be on the new integrated framework by establishing zero trust for all IoT traffic, fixed and mobile 5G/6G IoT network traffic, and data security (quarantine-zero touch and dynamic policy enforcement). We perform a comparative analysis of five machine learning models, namely, XGBoost, Random Forest, K-Nearest Neighbors, Stochastic Gradient Descent, and Native Bayes, by comparing these models based on accuracy, precision, recall, F1-score, and ROC-AUC. Results show that the best performance in detecting and mitigating different DDoS vectors comes from the ensemble-based approaches.

Figures

Figures reproduced from arXiv: 2502.03614 by the authors.

Figure 1
Figure 1. Comparison of Machine Learning Models 1) XGBoost: The gradient-boosted decision tree model XGBoost [16] is meant to be quick and efficient. It works well with complicated datasets because it in￾troduces regularisation to reduce overfitting. The model [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Metrics used for model evaluation Evaluation Metrices used are as follows: • Accuracy: Accuracy = TP + TN TP + TN + FP + FN (7) • Precision: Precision = TP TP + FP (8) • Recall: Recall = TP TP + FN (9) • F1 Score: F1 = 2 · Precision · Recall Precision + Recall (10) • AUC: AUC = Z 1 0 TPR(FPR) d(FPR) (11) III. PROPOSED FRAMEWORK The new proposed framework, an integration of the Zero-Trust Zero-Touch AI/ML-Enabled IoT… view at source ↗
Figure 3
Figure 3. Operational Workflow A. Framework Architecture The framework consists of three core components: • Zero-Trust Security: Ensures that all traffic will be authenticated and verified of all IoT devices, networks, and workloads, and data and their com￾munications—no more implicit trust, slicing huge margins off the attack surface. The ML-Powered NGFW uses ML-based classifications to intelli￾gently group related IoT devic… view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Zero Trust-Zero Touch Framework [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Proposed IoT Zero Trust-Zero Touch Framework System Design [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 8
Figure 8. Figure 8: ROC Curve for Random Forest C. K-Nearest Neighbors (KNN) K-Nearest Neighbors (KNN) is a simple yet effective machine learning algorithm that classifies data points based on their proximity to the k nearest neighbors in the feature space. In the context of IoT security,…
Figure 7
Figure 7. Figure 7: ROC Curve for XGBoost B. Random Forest Random Forest also performed strongly, at 99.79% for accuracy, precision and recall. In addition, it had a score of 0.9822 for AUC. Although slightly behind XGBoost in AUC, it remains highly robust due to its ensemble of decision …
Figure 12
Figure 12. Figure 12: Performance Comparison of ML Models V. DISCUSSION XGBoost emerged as the top performer, excelling in all metrics. Random Forest is nearly as accurate, while KNN, SGD, and Na¨ıve Bayes bring unique ad￾vantages in simplicity or speed, albeit with some trade￾offs. Ensemb…
Figure 10
Figure 10. Figure 10: ROC Curve for SGD E. Na¨ıve Bayes Na¨ıve Bayes, at 91.09% accuracy and 0.9829 AUC, suffers from its simplified assumption of feature indepen￾dence, which limits its capacity to handle complex IoT traffic patterns. Still, its low overhead makes it attractive for ultra-…
Figure 11
Figure 11. Figure 11: ROC Curve for Naive Bayes [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. System Security Framework for 5G Advanced /6G IoT Integrated Terrestrial Network-Non-Terrestrial Network (TN-NTN) with AI-Enabled Cloud Security

    cs.CR 2025-08 reject novelty 3.0 of 10

    A position paper combines existing ideas (AI cloud security, zero trust, federated learning, network slicing) into a proposed but unspecified and unevaluated security framework for integrated TN-NTN 5G/6G IoT networks.

Reference graph

Works this paper leans on

31 extracted references · 31 canonical work pages · cited by 1 Pith paper

  1. [21]

    S Shakya, R Abbas ”A Comparative Analysis of Machine Learning Models for DDoS Detection in IoT Networks” https://arxiv.org/abs/2411.05890

    work page Pith review arXiv

  2. [3]

    Lam and R

    L. Lam and R. Abbas, ”Anomaly detection in 5G networks,” arXiv preprint arXiv:2009.00000 , 2020

    work page arXiv 2009

  3. [1]

    Xu and F

    X. Xu and F. Wang, ”A survey on IoT security: Application areas, security threats, and solutions,” IEEE Internet Things J. , vol. 7, no. 10, pp. 9531–9544, 2020

    work page 2020

  4. [2]

    Available: https://www

    Cisco, ”Cisco Annual Internet Report (2018–2023) White Paper,” Cisco, 2022, [Online]. Available: https://www. cisco.com/c/en/us/solutions/collateral/executive-perspectives/ annual-internet-report/white-paper-c11-741490.html

    work page 2018

  5. [4]

    Mirkovic and P

    J. Mirkovic and P. Reiher, ”A taxonomy of DDoS attack and DDoS defense mechanisms,” ACM SIGCOMM Comput. Com- mun. Rev., vol. 34, no. 2, pp. 39–53, 2004

    work page 2004

  6. [5]

    S. T. Zargar, J. Joshi, and D. Tipper, ”A survey of defense mechanisms against distributed denial of service (DDoS) flooding attacks,” IEEE Commun. Surveys Tuts., vol. 15, no. 4, pp. 2046– 2069, 2013

    work page 2013

  7. [6]

    Allen, ”Ransomware Threats in the IoT Ecosystem,” Comput- ers & Security , vol

    J. Allen, ”Ransomware Threats in the IoT Ecosystem,” Comput- ers & Security , vol. 130, p. 103249, 2024

    work page 2024

  8. [7]

    Rajatheva et al., ”White paper on broadband connectivity in 6G,” 6G Research Visions, no

    N. Rajatheva et al., ”White paper on broadband connectivity in 6G,” 6G Research Visions, no. 10, 2020

    work page 2020

Show all 31 references
  1. [8]

    Lee, ”ML-based DDoS Detection for 6G-Enabled IoT: Chal- lenges and Opportunities,” IEEE Trans

    M. Lee, ”ML-based DDoS Detection for 6G-Enabled IoT: Chal- lenges and Opportunities,” IEEE Trans. on Industrial Informatics, vol. 19, no. 4, pp. 1234–1245, 2023

    2023

  2. [9]

    Kindervag, ”No more chewy centers: Introducing the zero trust model of information security,” Forrester Research, 2010

    J. Kindervag, ”No more chewy centers: Introducing the zero trust model of information security,” Forrester Research, 2010

    2010

  3. [10]

    Chang, ”A Zero-Trust Architecture for 5G-based IoT Net- works,” IEEE Access, vol

    Y . Chang, ”A Zero-Trust Architecture for 5G-based IoT Net- works,” IEEE Access, vol. 11, pp. 56789–56803, 2023

    2023

  4. [11]

    Khan and R

    U. Khan and R. Smith, ”Enhancing neural-based intrusion detec- tion systems with adversarial training for IoT security,” Ad Hoc Networks, vol. 148, p. 103951, 2023

    2023

  5. [12]

    S. R. Pandya, ”Smart city IoT architecture and challenges: A comprehensive review,” Ad Hoc Networks , vol. 122, p. 102663, 2022

    2022

  6. [13]

    Zhao, ”A factory automation case study in 5G-based IIoT,” IEEE Trans

    L. Zhao, ”A factory automation case study in 5G-based IIoT,” IEEE Trans. Ind. Informat., vol. 19, no. 7, pp. 4587–4600, 2023

    2023

  7. [14]

    S. R. Pokhrel, S. Moh, and J. Park, ”Towards detecting IoT botnets: a survey of machine learning approaches on botnet datasets,” Sensors, vol. 21, no. 1, p. 146, 2021

    2021

  8. [15]

    Sommer and V

    R. Sommer and V . Paxson, ”Outside the closed world: On using machine learning for network intrusion detection,” in Proc. IEEE Symp. Security and Privacy (SP) , 2010, pp. 305–316

    2010

  9. [16]

    Dhaliwal, A

    J. Dhaliwal, A. Nahid, and R. Abbas, ”Effective intrusion detec- tion system using XGBoost,” Electronics, vol. 7, no. 12, p. 345, 2018

    2018

  10. [17]

    Chen and C

    T. Chen and C. Guestrin, ”XGBoost: A scalable tree boosting system,” in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016, pp. 785–794

    2016

  11. [18]

    Dai, ”A Review of XGBoost for Cyber Threat Detection,” IEEE Access, vol

    F. Dai, ”A Review of XGBoost for Cyber Threat Detection,” IEEE Access, vol. 12, pp. 112233–112245, 2024

    2024

  12. [19]

    Bland, ”A Critical Assessment of Zero Trust in 5G Networks,” IEEE Communications Magazine , vol

    G. Bland, ”A Critical Assessment of Zero Trust in 5G Networks,” IEEE Communications Magazine , vol. 60, no. 11, pp. 42–48, 2022

    2022

  13. [20]

    Chan and M

    S. Chan and M. G. Tyson, ”Context-Aware Zero Trust for IoT Ecosystems,” IEEE Internet Things J., vol. 9, no. 13, pp. 10001– 10012, 2022

    2022

  14. [22]

    Freed and D

    A. Freed and D. Holt, ”Designing Zero Trust Architectures for Next-Generation IoT,” IEEE Trans. Netw. Serv. Manag., vol. 20, no. 2, pp. 1333–1348, 2023

    2023

  15. [23]

    Mir and N

    F. Mir and N. K. Noori, ”Micro-segmentation strategies for securing IoT networks,” IEEE Commun. Surveys Tuts. , vol. 23, no. 4, pp. 185–200, 2021

    2021

  16. [24]

    Ma et al., ”Federated learning in edge computing: A survey on frameworks, applications, and challenges,” IEEE Internet Things J., vol

    X. Ma et al., ”Federated learning in edge computing: A survey on frameworks, applications, and challenges,” IEEE Internet Things J., vol. 9, no. 24, pp. 25028–25047, 2022

    2022

  17. [25]

    Blum, ”Adaptive federated anomaly detection in distributed IoT networks,” Computers, vol

    T. Blum, ”Adaptive federated anomaly detection in distributed IoT networks,” Computers, vol. 12, no. 4, p. 87, 2023

    2023

  18. [26]

    Carter and B

    L. Carter and B. Simpson, ”A Forensic Approach to IoT-based DDoS Attacks,” ACM Comput. Surv. , vol. 54, no. 8, pp. 1–28, 2021

    2021

  19. [27]

    Divekar, G

    A. Divekar, G. Parekh, D. Savla, M. S. Das, and S. R. Pandya, ”Benchmarking datasets for anomaly-based network intrusion detection: KDD CUP 99 alternatives,” in Proc. IEEE 3rd Intl. Conf. on Computing, Communication and Security (ICCCS) , 2018, pp. 1–8

    2018

  20. [28]

    Xu and M

    Y . Xu and M. K. Stewart, ”Data Quality and Preprocessing for IoT DDoS Detection,” Future Internet , vol. 13, no. 5, p. 115, 2021

    2021

  21. [29]

    M. S. Ahmed, A. N. Mahmood, and J. Hu, ”Improving network anomaly detection with minority oversampling in big data,” Computers & Security , vol. 114, p. 102595, 2022

    2022

  22. [30]

    L. G. C. Castedo, ”Blockchain for IoT security: A survey,” Internet of Things , vol. 19, p. 100567, 2022

    2022

  23. [31]

    Gadepalli and J

    S. Gadepalli and J. Rao, ”Explainable AI in intrusion detection systems: A survey,” ACM Comput. Surv. , vol. 54, no. 10, pp. 1–32, 2021

    2021

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.