Pith. sign in

REVIEW 4 major objections 5 minor 2 cited by

Federated Learning-Based Data Collaboration Method for Enhancing Edge Cloud AI System Security Using Large Language Models

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Adding an LLM layer that weights node updates and selectively activates secure aggregation is claimed to give federated learning a 15% improvement in data protection and model robustness over traditional FL.

desk verdict LLM-guided FL weighting is a fresh idea, but the claimed 15% gain has zero reported evidence and the mechanism has an unresolved privacy-scoring contradiction. read the letter →

arxiv 2506.18087 v1 pith:Q5PP7UIQ submitted 2025-06-22 cs.CR cs.AI

Huaiying Luo , Cheng Ji This is my paper · ORCID
classification cs.CRcs.AI
keywords federatedlearninglargelanguagemodelsedgecloudAIsecuremulti-partycomputationadversarialtrainingdataprivacymodelrobustnesscommunicationlatency
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

The paper argues that federated learning for edge-cloud AI can be made both more private and more efficient by putting a large language model inside the aggregation loop. The LLM is supposed to read encrypted metadata about each node's gradient trends, loss landscape, and update history, then assign aggregation weights that favor reliable nodes and downweight or flag suspicious ones. It also decides when the full secure multi-party computation protocol is needed, so heavy cryptographic operations are reserved for high-risk rounds, and it guides adversarial sample generation during training. The reported experiments on an industrial IoT intrusion-detection dataset claim about 15% improvement over vanilla federated learning and over differential-privacy, SMC-based, and homomorphic-encryption alternatives in data protection and model robustness. A reader should care because it points toward a division of labor where a powerful central model handles judgment while local nodes retain data.

What carries the argument

The load-bearing mechanism is the LLM as a controller rather than a classifier. It operates on encrypted metadata summaries of each node's gradient trends, loss landscape, update frequency, and divergence from the global model, and outputs a weight vector used in the softmax aggregation formula \($w^{{t+1}}$_i = \exp(\$\alpha$\,\mathrm{Performance}(i)) / \sum_j \exp(\$\alpha$\,\mathrm{Performance}(j))\). The same controller decides when to run full homomorphic SMC aggregation instead of lightweight masking, and it proposes perturbation directions for adversarial training in the total loss \(L_{\mathrm{total}}(\$\theta$)= \mathbb{E}_{x,y\sim D}[L(f(x,\$\theta$),y)+\$\lambda$ L_{\mathrm{adv}}(f(x,\$\theta$),f(x+\delta,\$\theta$))]\). This lets the system vary its cryptographic and defensive effort round by round instead of applying uniform protection.

What would settle it

Run the proposed pipeline with the LLM's aggregation weights replaced by uniform or random weights while keeping everything else fixed; if robustness and latency are unchanged, the LLM is not doing the work. Alternatively, in a controlled poisoning experiment, compare the LLM's reliability scores against ground-truth labels of which nodes are malicious; systematic disagreement would falsify the mechanism.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that an LLM can act as a semi-centralized coordinator in federated learning: it consumes encrypted metadata, computes a score vector that becomes a softmax weight per node, and triggers SMC-based encrypted aggregation only when its inferred privacy risk crosses a threshold. Combined with adversarial training, the paper claims this selective, LLM-guided aggregation yields stronger privacy and robustness than always-on encryption or differential privacy, because it avoids unnecessary cryptographic computation and suppresses unreliable node updates. The central quantitative claim is a roughly 15% advantage over traditional federated learning in data protection and model robustness.

Load-bearing premise

The entire method depends on the premise that a large language model can derive trustworthy, private signals about which nodes are reliable or adversarial from encrypted metadata and update statistics—and the paper presents no test that this is possible.

Editorial extensions

If this is right

  • If the LLM's trust scores are accurate, federated aggregation can suppress poisoned or unreliable node updates without sacrificing the contributions of healthy nodes.
  • Rounds with low inferred risk can skip full SMC, lowering communication latency and computational cost while preserving most of the privacy benefit.
  • Adversarial training steered by LLM-identified vulnerable feature spaces should keep global-model accuracy more stable under attack than the DP, SMC, or HE baselines.
  • The claimed 15% margin over traditional federated learning, if reproducible, gives edge-cloud operators a concrete reason to test LLM-assisted aggregation as a security upgrade.

Reading between the lines

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

  • The paper never demonstrates that an LLM can actually extract trustworthy reliability signals from encrypted metadata; a natural next test is to replace the LLM's weights with uniform random weights and see whether the claimed gains vanish.
  • Because the LLM sees metadata summaries, an unexamined risk is that its outputs leak information about individual nodes; a membership-inference audit over the LLM's scores would be a useful extension.
  • The selective SMC trigger requires a privacy-risk threshold, but the paper does not say how that threshold is learned; estimating it from data while preserving privacy is an open design question.
  • If the LLM is itself adversarial or compromised, it controls both weighting and encryption scheduling, creating a single point of failure that the paper does not address.
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 / 5 minor

Summary. The manuscript proposes a federated learning (FL) framework for edge-cloud AI in which a large language model (LLM) acts as a control layer: it assigns per-node aggregation weights (Eq. 5), decides when to trigger secure multi-party computation (SMC) encryption, and guides adversarial perturbation generation (Eq. 8). The text states that the method is evaluated on the Edge-IIoTset dataset against Vanilla FL, DP-FL, SMC-FL, and HE-FL, and the abstract and conclusion claim about 15% improvement in data protection and model robustness. However, the manuscript contains no figures, tables, numeric results, or hyperparameter settings; the experimental section consists only of captions and qualitative assertions. The central mechanism is also underspecified: the LLM is said to reason over encrypted metadata while the aggregator receives SMC ciphertexts, and Eq. (7) is mathematically ill-posed.

Significance. If fully realized, the proposed hierarchical LLM-controlled aggregation could address a relevant and practical problem: reducing unnecessary cryptographic overhead and improving robustness in heterogeneous edge-cloud FL. The paper shows awareness of existing work on LLM-based edge security and SMC/FL, and it identifies an interesting design question, namely when SMC should be applied selectively rather than unconditionally. However, the contribution is at present an architecture sketch. There is no formal privacy guarantee, no convergence analysis, no reproducibility artifact, and no empirical evidence; the signature 15% figure is an assertion rather than a measured outcome. The paper is not rejected because it diverges from consensus but because the internal mechanism and the evidence provided do not support the stated claims.

major comments (4)
  1. [Abstract; Section IV-B] The headline claim of a 15% improvement in data protection and model robustness is unsupported. The experimental section contains only captions for Figures 1 and 2, which are absent from the manuscript; no table, curve, error bar, or raw numeric result is given, and 'data protection' is never defined as a measurable quantity. Because this claim is the stated contribution in the abstract and conclusion, the omission is load-bearing and cannot be corrected without a substantial new experimental section.
  2. [Section III-A/B; Eq. (5)] Eq. (5) defines the node weights through Performance(i), which is never defined, and the text says the LLM reasons over 'encrypted metadata' while the aggregator receives SMC ciphertexts. The manuscript does not explain what information the LLM can access or how it converts metadata into Performance(i). If the metadata reveal accuracy trends, update divergence, and security alerts, the SMC privacy guarantee is compromised; if they do not, Eq. (5) cannot be computed as written. A precise access model, a concrete prompt design, and a privacy analysis of the metadata channel are required.
  3. [Section III-B; Eq. (7)] Eq. (7) is ill-posed. The objective E[||θ~^{t+1}_i - θ~^t_i||^2] does not depend on the decision variable f_i, so the argmin is vacuous; the constraint f_i in {0,1} is disconnected from the phrase 'F is an optional set of frequencies.' This optimization cannot be executed as stated and should be reformulated with a participation-dependent utility or removed.
  4. [Section IV-A] The experimental setup omits all reproducibility-critical parameters: α in Eq. (5), λ in Eq. (9), ε in Eq. (8), the model architecture, the number of nodes and data partition, the LLM version and prompts, the SMC implementation, and the training rounds. As a result, the comparison against VFL, DP-FL, SMC-FL, and HE-FL cannot be reproduced or interpreted even if the missing figures were supplied.
minor comments (5)
  1. [Fig. 2 caption] The caption of Figure 2 reads 'SResistance to Adversarial Examples Across Different Methods' and should read 'Resistance to Adversarial Examples Across Different Methods'.
  2. [Eq. (8)] In Eq. (8), the notation 'where ||δ||_p ≤ ε is the limit of the adversarial perturbation' is incorrect: ||δ||_p ≤ ε is a constraint on the argmax, not a limit; the equation should be written as a constrained maximization.
  3. [Section IV-A] The Edge-IIoTset dataset is mentioned without a citation or version; a reference and a description of preprocessing, label distribution, and train/test split should be added.
  4. [Section III-A] Section III-A states that SMC is triggered when 'privacy risk exceeds a learned threshold,' but no learning procedure for this threshold is described; the text should either specify how it is trained or label it as future work.
  5. [Throughout] The paper uses both 'large-scale language models' and 'large language models' (for example, the first paragraph of Section II versus the rest of the paper); the terminology should be unified.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the claimed performance gain is an asserted empirical outcome, not a quantity derived from the model equations.

full rationale

The derivation chain in this paper consists of standard federated learning local updates (Eq. 1), averaging aggregation (Eq. 2), SMC encryption (Eqs. 3-4), LLM-based weighted aggregation (Eqs. 5-6), participation scheduling (Eq. 7), and adversarial training objectives (Eqs. 8-9). Each equation is a conventional definition or optimization objective taken from the literature, and none of the claimed results, such as the 15% improvement in data protection and model robustness, is obtained algebraically from these equations. The paper does not fit a parameter to a subset of data and then predict a closely related quantity; alpha and lambda appear as free hyperparameters, and no numerical result is computed from them in the text. The citations refer to external prior work, and no load-bearing argument reduces to a self-citation: the authors do not invoke their own prior theorem to forbid alternatives or to justify a central premise. Several internal weaknesses exist, including the undefined Performance(i) in Eq. 5, the vacuous objective in Eq. 7 whose expression does not depend on f_i, and the absence of the experimental figures referenced in Section IV, but these are correctness and completeness defects, not circular derivation. No step in the paper defines a quantity in terms of the very quantity it is supposed to predict, nor renames an empirical pattern as a new result. The advertised improvement is an unsupported empirical assertion rather than a derived prediction, so the circularity score is 0.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The paper's contribution rests on a set of unvalidated assumptions about LLM reasoning over encrypted summaries and on unspecified free parameters; no new entities are introduced.

free parameters (4)
  • alpha in Eq. 5
    Sensitivity parameter in the softmax weighting over node performance; no value or tuning method is given.
  • lambda in Eq. 9
    Regularization coefficient for adversarial training; unspecified.
  • epsilon in Eq. 8
    Maximum perturbation size for adversarial samples; unspecified.
  • Privacy risk threshold for SMC trigger
    Learned threshold that decides when full SMC computation is used; the learning procedure and value are not described.
assumptions (3)
  • domain assumption An LLM can reliably infer node quality and adversarial behavior from encrypted metadata or summary statistics.
    The weighting and selective SMC mechanism depends on this capability, which is asserted in Section III-A without experiment or implementation.
  • domain assumption The additive homomorphic encryption scheme supports the weighted sum aggregation in encrypted space as described.
    The protocol is described as inspired by Paillier but no correctness verification or security analysis is provided (Section III-B).
  • domain assumption Adversarial training with bounded perturbations transfers to heterogeneous federated learning without unexpected privacy-utility tradeoffs.
    Standard in centralized settings, but the paper gives no analysis of its behavior under data heterogeneity and encryption (Equation 9).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Federated Learning-Based Data Collaboration Method for Enhancing Edge Cloud AI System Security Using Large Language Models." pith.science (2026). https://pith.science/paper/Q5PP7UIQ

@misc{pith2026250618087,
  author       = {Pith},
  title        = {Pith review of: Federated Learning-Based Data Collaboration Method for Enhancing Edge Cloud AI System Security Using Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q5PP7UIQ}},
  note         = {Machine review of arXiv:2506.18087}
}
read the original abstract

With the widespread application of edge computing and cloud systems in AI-driven applications, how to maintain efficient performance while ensuring data privacy has become an urgent security issue. This paper proposes a federated learning-based data collaboration method to improve the security of edge cloud AI systems, and use large-scale language models (LLMs) to enhance data privacy protection and system robustness. Based on the existing federated learning framework, this method introduces a secure multi-party computation protocol, which optimizes the data aggregation and encryption process between distributed nodes by using LLM to ensure data privacy and improve system efficiency. By combining advanced adversarial training techniques, the model enhances the resistance of edge cloud AI systems to security threats such as data leakage and model poisoning. Experimental results show that the proposed method is 15% better than the traditional federated learning method in terms of data protection and model robustness.

Figures

Figures reproduced from arXiv: 2506.18087 by the authors.

Figure 2
Figure 2. SResistance to Adversarial Examples Across Different [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 1
Figure 1. Structured Communication Latency Comparison with [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Exploring Reasoning-Infused Text Embedding with Large Language Models for Zero-Shot Dense Retrieval

    cs.CL 2025-08 conditional novelty 4.0 of 10

    Reasoning-infused text embedding, which prepends LLM-generated reasoning to queries before embedding, improves zero-shot dense retrieval on BRIGHT.

  2. Collaborative Inference and Learning between Edge SLMs and Cloud LLMs: A Survey of Algorithms, Execution, and Open Challenges

    cs.DC 2025-07 conditional novelty 4.0 of 10

    A survey that builds a taxonomy of edge-cloud LLM-SLM collaboration for inference and training, claiming to be the first to unify both phases.

Reference graph

Works this paper leans on

38 extracted references · 21 canonical work pages · cited by 2 Pith papers

  1. [1]

    Advances in appfl: A comprehensive and extensible federated learning framework,

    Z. Li, S. He, Z. Yang, M. Ryu, K. Kim, and R. Madduri, “Advances in appfl: A comprehensive and extensible federated learning framework,” arXiv preprint arXiv:2409.11585 , 2024

    arXiv 2024

  2. [2]

    Mt2st: Adaptive multi-task to single-task learning,

    D. Liu and Y . Yu, “Mt2st: Adaptive multi-task to single-task learning,” arXiv preprint arXiv:2406.18038 , 2024

    arXiv 2024

  3. [3]

    Hades: Hardware accelerated decoding for efficient speculation in large language models,

    Z. Yang, Y . Jin, and X. Xu, “Hades: Hardware accelerated decoding for efficient speculation in large language models,” arXiv preprint arXiv:2412.19925, 2024

    arXiv 2024

  4. [4]

    Research on large language model cross-cloud privacy protection and collaborative training based on federated learning,

    Z. Yang, Y . Jin, Y . Zhang, J. Liu, and X. Xu, “Research on large language model cross-cloud privacy protection and collaborative training based on federated learning,” arXiv preprint arXiv:2503.12226 , 2025

    arXiv 2025

  5. [5]

    Fine-grained control of generative data aug- mentation in iot sensing,

    T. Wang, Q. Yang, R. Wang, D. Sun, J. Li, Y . Chen, Y . Hu, C. Yang, T. Kimura, D. Kara et al., “Fine-grained control of generative data aug- mentation in iot sensing,” Advances in Neural Information Processing Systems, vol. 37, pp. 32 787–32 812, 2024

    work page 2024

  6. [6]

    Adaptive fault tolerance mechanisms of large language models in cloud computing environ- ments,

    Y . Jin, Z. Yang, X. Xu, Y . Zhang, and S. Ji, “Adaptive fault tolerance mechanisms of large language models in cloud computing environ- ments,” arXiv preprint arXiv:2503.12228 , 2025

    arXiv 2025

  7. [7]

    Enhancing thyroid disease prediction using machine learning: A comparative study of ensemble models and class balancing techniques,

    J. Zhong and Y . Wang, “Enhancing thyroid disease prediction using machine learning: A comparative study of ensemble models and class balancing techniques,” 2025

    work page 2025

  8. [8]

    Stutz, J

    D. Stutz, J. T. Assis, A. A. Laghari, A. A. Khan, N. Andreopoulos, A. Terziev, A. Deshpande, D. Kulkarni, and E. G. Grata, Enhancing Security in Cloud Computing Using Artificial Intelligence (AI) , 2024, pp. 179–220

    work page 2024

Show all 38 references
  1. [9]

    Performance of unscented kalman filter for model updating with experimental data,

    M. Cheng and T. C. Becker, “Performance of unscented kalman filter for model updating with experimental data,” Earthquake Engineering & Structural Dynamics, vol. 50, no. 7, pp. 1948–1966, 2021

    1948

  2. [10]

    Pmfl: Partial meta-federated learning for heterogeneous tasks and its applications on real-world medical records,

    T. Zhang, S. Zhang, Z. Chen, Y . Bengio, and D. Liu, “Pmfl: Partial meta-federated learning for heterogeneous tasks and its applications on real-world medical records,” in 2022 IEEE International Conference on Big Data (Big Data) . IEEE, 2022, pp. 4453–4462

    2022

  3. [11]

    Uncertainty-based extensible codebook for discrete federated learning in heterogeneous data silos,

    T. Zhang, Y . Cao, and D. Liu, “Uncertainty-based extensible codebook for discrete federated learning in heterogeneous data silos,” arXiv preprint arXiv:2402.18888, 2024

    2024 arXiv

  4. [12]

    Machine learning: Trends, perspec- tives, and prospects,

    M. I. Jordan and T. M. Mitchell, “Machine learning: Trends, perspec- tives, and prospects,” Science, vol. 349, no. 6245, pp. 255–260, 2015

    2015

  5. [13]

    RAG-RLRC-LaySum at BioLaySumm: Integrating retrieval-augmented generation and readability control for layman summarization of biomedical texts,

    Y . Ji, Z. Li, R. Meng, S. Sivarajkumar, Y . Wang, Z. Yu, H. Ji, Y . Han, H. Zeng, and D. He, “RAG-RLRC-LaySum at BioLaySumm: Integrating retrieval-augmented generation and readability control for layman summarization of biomedical texts,” in Proceedings of the 23rd Workshop o...

    2024

  6. [14]

    Machine learning algorithms-a review,

    B. Mahesh et al., “Machine learning algorithms-a review,” International Journal of Science and Research (IJSR).[Internet] , vol. 9, no. 1, pp. 381–386, 2020

    2020

  7. [15]

    Regional style and color transfer,

    Z. Ding, P. Li, Q. Yang, S. Li, and Q. Gong, “Regional style and color transfer,” in 2024 5th International Conference on Computer Vision, Image and Deep Learning (CVIDL) . IEEE, 2024, pp. 593–597

    2024

  8. [16]

    Covid-19 localization and recognition on chest radiographs based on yolov5 and efficientnet,

    T. Zhang, B. Zhang, F. Zhao, and S. Zhang, “Covid-19 localization and recognition on chest radiographs based on yolov5 and efficientnet,” in 2022 7th International Conference on Intelligent Computing and Signal Processing (ICSP). IEEE, 2022, pp. 1827–1830

    2022

  9. [17]

    An ensemble approach to stock price prediction using deep learning and time series models,

    M. Sui, C. Zhang, L. Zhou, S. Liao, and C. Wei, “An ensemble approach to stock price prediction using deep learning and time series models,” in 2024 IEEE 6th International Conference on Power, Intelligent Com- puting and Systems (ICPICS) . IEEE, 2024, pp. 793–797

    2024

  10. [18]

    Data augmentation for human activity recognition via condition space interpolation within a generative model,

    T. Wang, Y . Chen, Q. Yang, D. Sun, R. Wang, J. Li, T. Kimura, and T. Abdelzaher, “Data augmentation for human activity recognition via condition space interpolation within a generative model,” in 2024 33rd International Conference on Computer Communications and Networks (ICCC...

    2024

  11. [19]

    Geometric matrix completion for missing data estimation in power distribution systems,

    S. Qiu, Y . Li, Z. Wang, Y . Shen, Z. Li, and F. Shen, “Geometric matrix completion for missing data estimation in power distribution systems,” in 2024 The 9th International Conference on Power and Renewable Energy (ICPRE), 2024, pp. 1605–1609

    2024

  12. [20]

    Gpt-4 technical report,

    J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkat et al. , “Gpt-4 technical report,” arXiv preprint arXiv:2303.08774 , 2023

    2023 arXiv

  13. [21]

    Large language model based multi-agents: A survey of progress and challenges,

    T. Guo, X. Chen, Y . Wang, R. Chang, S. Pei, N. V . Chawla, O. Wiest, and X. Zhang, “Large language model based multi-agents: A survey of progress and challenges,” arXiv preprint arXiv:2402.01680 , 2024

    2024 arXiv

  14. [22]

    Composerx: Multi-agent symbolic music composition with llms,

    Q. Deng, Q. Yang, R. Yuan, Y . Huang, Y . Wang, X. Liu, Z. Tian, J. Pan, G. Zhang, H. Lin et al. , “Composerx: Multi-agent symbolic music composition with llms,” arXiv preprint arXiv:2404.18081 , 2024

    2024 arXiv

  15. [23]

    Exploring diverse methods in visual question answering,

    P. Li, Q. Yang, X. Geng, W. Zhou, Z. Ding, and Y . Nian, “Exploring diverse methods in visual question answering,” in 2024 5th International Conference on Electronic Communication and Artificial Intelligence (ICECAI). IEEE, 2024, pp. 681–685

    2024

  16. [24]

    Assertion detection in clinical natural language processing using large language models,

    Y . Ji, Z. Yu, and Y . Wang, “Assertion detection in clinical natural language processing using large language models,” in 2024 IEEE 12th International Conference on Healthcare Informatics (ICHI) , 2024, pp. 242–247

    2024

  17. [25]

    Score: Story coherence and retrieval enhancement for ai narratives,

    Q. Yi, Y . He, J. Wang, X. Song, S. Qian, X. Yuan, M. Zhang, L. Sun, K. Li, K. Lu et al., “Score: Story coherence and retrieval enhancement for ai narratives,” arXiv preprint arXiv:2503.23512 , 2025

    2025

  18. [26]

    Give: Structured reasoning of large language models with knowledge graph inspired veracity extrapolation,

    J. He, M. D. Ma, J. Fan, D. Roth, W. Wang, and A. Ribeiro, “Give: Structured reasoning of large language models with knowledge graph inspired veracity extrapolation,” 2025. [Online]. Available: https://arxiv.org/abs/2410.08475

    2025 arXiv

  19. [27]

    Self-give: Associative thinking from limited structured knowledge for enhanced large language model reasoning,

    J. He, J. Fan, B. Jiang, I. Houine, D. Roth, and A. Ribeiro, “Self-give: Associative thinking from limited structured knowledge for enhanced large language model reasoning,” 2025. [Online]. Available: https://arxiv.org/abs/2505.15062

    2025 arXiv

  20. [28]

    Can speculative sampling accelerate react without compromising reasoning quality?

    H. Xu, J. Ye, Y . Li, and H. Chen, “Can speculative sampling accelerate react without compromising reasoning quality?” in The Second Tiny Papers Track at ICLR 2024 , 2024. [Online]. Available: https://openreview.net/forum?id=42b9hJrIpX

    2024

  21. [29]

    Generative ai in cybersecurity: A com- prehensive review of llm applications and vulnerabilities,

    M. A. Ferrag, F. Alwahedi, A. Battah, B. Cherif, A. Mechri, N. Tihanyi, T. Bisztray, and M. Debbah, “Generative ai in cybersecurity: A com- prehensive review of llm applications and vulnerabilities,” Internet of Things and Cyber-Physical Systems , 2025

    2025

  22. [30]

    Reinforcement learning-based load balancing with large language models and edge intelligence for dynamic cloud environments,

    B. Desai and K. Patil, “Reinforcement learning-based load balancing with large language models and edge intelligence for dynamic cloud environments,” Journal of Innovative Technologies , vol. 6, no. 1, pp. 1–13, 2023

    2023

  23. [31]

    Scalability optimization in cloud-based ai inference services: Strategies for real-time load balancing and automated scaling,

    Y . Jin and Z. Yang, “Scalability optimization in cloud-based ai inference services: Strategies for real-time load balancing and automated scaling,” arXiv preprint arXiv:2504.15296 , 2025

    2025 arXiv

  24. [32]

    Enhancing the security of edge-ai runtime environments: a fine-tuning method based on large language models,

    D. Tang, P. Xiao, T. Zheng, X. Li, and C. Yang, “Enhancing the security of edge-ai runtime environments: a fine-tuning method based on large language models,” Wireless Networks, vol. 31, no. 2, pp. 1825–1838, 2025

    2025

  25. [33]

    Distributed threat intelligence at the edge devices: A large language model-driven approach,

    S. M. Hasan, A. M. Alotaibi, S. Talukder, and A. R. Shahid, “Distributed threat intelligence at the edge devices: A large language model-driven approach,” in 2024 IEEE 48th Annual Computers, Software, and Appli- cations Conference (COMPSAC) . IEEE, 2024, pp. 1496–1497

    2024

  26. [34]

    Research on cloud platform network traffic monitoring and anomaly detection system based on large language models,

    Z. Yang, Y . Jin, J. Liu, X. Xu, Y . Zhang, and S. Ji, “Research on cloud platform network traffic monitoring and anomaly detection system based on large language models,” arXiv preprint arXiv:2504.17807 , 2025

    2025 arXiv

  27. [35]

    Large language models empowered autonomous edge ai for connected intelligence,

    Y . Shen, J. Shao, X. Zhang, Z. Lin, H. Pan, D. Li, J. Zhang, and K. B. Letaief, “Large language models empowered autonomous edge ai for connected intelligence,” IEEE Communications Magazine , vol. 62, no. 10, pp. 140–146, 2024

    2024

  28. [36]

    A survey on the integration and optimization of large language models in edge computing envi- ronments,

    S. Bhardwaj, P. Singh, and M. K. Pandit, “A survey on the integration and optimization of large language models in edge computing envi- ronments,” in 2024 16th International Conference on Computer and Automation Engineering (ICCAE) . IEEE, 2024, pp. 168–172

    2024

  29. [37]

    Llm-cloud complete: Lever- aging cloud computing for efficient large language model-based code completion,

    M. Zhang, B. Yuan, H. Li, and K. Xu, “Llm-cloud complete: Lever- aging cloud computing for efficient large language model-based code completion,” Journal of Artificial Intelligence General science (JAIGS) ISSN: 3006-4023, vol. 5, no. 1, pp. 295–326, 2024

    2024

  30. [38]

    Cached model-as-a-resource: Provisioning large language model agents for edge intelligence in space-air-ground integrated networks,

    M. Xu, D. Niyato, H. Zhang, J. Kang, Z. Xiong, S. Mao, and Z. Han, “Cached model-as-a-resource: Provisioning large language model agents for edge intelligence in space-air-ground integrated networks,” arXiv preprint arXiv:2403.05826, 2024

    2024 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.