Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

A New Pathway to Integrated Learning and Communication (ILAC): Large AI Model and Hyperdimensional Computing for Communication

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read This paper proposes a unified integrated learning and communication (ILAC) framework pairing large AI models with hyperdimensional computing, and argues through a cost-to-performance case study that jointly optimizing task assignment…

desk verdict Solid ILAC survey, but the case study's joint-optimization claim is unsupported: model size is decoupled from accuracy and the baselines are set up to lose. read the letter →

arxiv 2506.18432 v2 pith:6CKCVER2 submitted 2025-06-23 eess.SP

classification eess.SP
keywords integratedlearningandcommunicationhyperdimensionalcomputinglargeAImodels6Gnetworksresourceallocationcost-to-performanceratiomodelcompressiondistributed
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 learning and communication in future wireless networks should be designed as one optimization problem, not two separate ones. It organizes the field into two architectures, model distribution and data distribution, and proposes adding large AI models for generality plus hyperdimensional computing (HDC) for lightweight, noise-tolerant model representations. The central quantitative claim comes from a case study: jointly choosing which client runs which task, how much to compress each model, and how to split bandwidth and power among clients minimizes a cost-to-performance ratio. On simulated networks, this dual-function design reports lower cost-to-performance ratios than designs that optimize communication alone or learning alone across all tested bandwidth budgets. A sympathetic reader would care because this offers a concrete route toward 6G systems that run learning tasks under tight bandwidth, power, and latency constraints.

What carries the argument

The load-bearing object is the HDC model itself: each class is represented as a single high-dimensional random vector formed by summing per-sample hypervectors, so a trained model is just a small collection of hypervectors whose size equals the number of classes times the hypervector dimension D. Addition bundles items into sets, multiplication binds features to values, and cosine similarity or Hamming distance performs classification. Around this object the paper builds a compression stage, where a foundation model maps each HDC model to a smaller representation depending on channel state, and an aggregation stage, where a server-side foundation model fuses received models per task. The optimization machinery is the CPR objective, minimized by Dinkelbach's transform and alternating optimization, with per-iteration complexity O(MN + $N^{3}$) and overall complexity O(I1 I2 (MN + $N^{3}$)).

What would settle it

Run the proposed CPR optimization on a real HDC classification task with a public dataset, measuring classification accuracy after foundation-model compression and after noisy transmission. If the measured accuracy of compressed models is substantially lower than the assumed formula, or if the joint design no longer beats the communication-only baseline on measured CPR, the central claim fails.

Watch

Extended reading notes

Core claim

The paper claims that ILAC can be instantiated as a single optimization problem whose objective is the Cost-to-Performance Ratio (CPR): the total computational cost of compressing hyperdimensional models divided by a product of communication rate and HDC inference accuracy. Each client trains an HDC classifier on its assigned task, a foundation model compresses the resulting hypervectors before wireless transmission, and the central server aggregates the received models with another foundation model. The authors solve the CPR minimization problem with the Dinkelbach transformation plus alternating optimization, alternating between task assignment and model size on one side and bandwidth and power allocation on the other. Their reported result, across bandwidth budgets from 1 to 10 MHz, is that the joint design converges to a stationary point and consistently achieves lower CPR than both a communication-oriented baseline and a learning-oriented baseline.

Load-bearing premise

The optimization conclusions rest on assuming the classification accuracy of a compressed, wirelessly received hyperdimensional model follows a known formula and that a foundation model can shrink these models without meaningfully degrading that accuracy; the paper does not specify or validate either function.

Editorial extensions

If this is right

  • If the case study is right, network designers should allocate bandwidth, power, task assignment, and model compression jointly rather than treating communication and learning as separate subsystems.
  • Larger total bandwidth lowers the cost-to-performance ratio in the proposed design, so the framework gives a quantitative way to trade spectrum for learning quality.
  • Because HDC models can be trained in a single pass without gradient backpropagation, the framework points to distributed learning systems that skip iterative fine-tuning and still adapt to new tasks over the air.
  • The convergence result, a stationary point obtained in roughly 90 alternating-optimization iterations, suggests the joint problem is tractable at practical network scales with polynomial complexity in the number of clients and tasks.

Reading between the lines

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

  • Extension: The paper treats HDC accuracy after compression and noisy transmission as a known function phi, but does not measure it; a natural next step is to fit phi empirically on real datasets and re-run the optimization, which would show how much of the reported CPR gain survives realistic compression loss.
  • Extension: Since the objective is a ratio of cost to a product of rate and accuracy, the same framework could be extended to trace a full Pareto frontier between communication rate and inference accuracy instead of a single weighted ratio, giving operators a menu of operating points.
  • Extension: The same task-assignment, compression, bandwidth, and power formulation could be applied to over-the-air aggregation or semantic communication, where multiple clients' hypervectors are combined in the channel itself, rather than only to digital transmission of compressed models.
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 / 5 minor

Summary. The paper is an invited survey of Integrated Learning and Communication (ILAC) for 6G networks, organized around a taxonomy of model distribution and data distribution architectures. It reviews split learning, aggregation learning, intelligent transceivers, federated learning, multi-agent reinforcement learning, CSI prediction, and resource allocation, and then proposes an enhanced framework combining large AI models with hyperdimensional computing (HDC). The paper's quantitative contribution is a case study in Section VI-C in which task assignment, compressed model size, bandwidth, and power are jointly optimized to minimize a Cost-to-Performance Ratio (CPR) defined in Eq. (21). The authors use Dinkelbach's transformation and alternating optimization, and report in Fig. 14 that the proposed joint design achieves lower CPR than communication-oriented and learning-oriented baselines. The abstract claims that this constitutes a 'practical and effective solution' to balance learning performance and communication constraints.

Significance. If the case study were fully specified and validated, it would provide a useful demonstration of joint learning-communication optimization in an ILAC setting, and the survey portion has clear value as a structured synthesis of recent work on model/data distribution, large AI models, and HDC. The algorithmic description of Dinkelbach and alternating optimization, including the complexity analysis O(I1 I2 (MN + N^3)), is a useful contribution. The paper does not, however, ship code, data, or concrete instances of the key functions that drive the case study; its central quantitative claim is therefore currently unsupported. The survey parts and the proposed framework are likely of interest to the journal's readership, but the case-study conclusions need substantial grounding before the abstract's 'practical and effective' claim can be accepted.

major comments (3)
  1. [Section VI-C.2, Eqs. (14), (19), (20)] The case study's quantitative results depend on three functions that are never specified: the HDC accuracy function phi(.), the compression-cost function eta(.), and the foundation-model preprocessing function f_p(.). The text only states that phi(.) 'can be determined by the accuracy on the validation set' and that eta(.) 'characterizes the processing capability of the client-side foundation model.' Because no concrete HDC encoder, dataset, foundation model, or functional form is given, the CPR curves in Fig. 14 cannot be reproduced, and the abstract's claim of a 'practical and effective solution' is not supported. The authors should either instantiate these functions from real HDC models and measurements, or explicitly present the case study as an illustrative abstraction with a sensitivity analysis over these functions.
  2. [Section VI-C.2, Eqs. (19)-(20) and constraints (21d)-(21e)] The compressed model size sc_j appears only in the cost function eta(s0_j/sc_j) in Eq. (19) and in constraints (21d)-(21e); it does not appear as an argument of the accuracy function phi(Pc_j) in Eq. (20). Consequently, the optimizer can reduce sc_j to lower the numerator of the CPR without any corresponding penalty to the HDC performance term Q. This means the formulation does not actually model the inference-accuracy-versus-communication-cost trade-off that is central to the ILAC concept; the claimed joint optimization of model size and accuracy is therefore not realized by the stated problem.
  3. [Section VI-C.4, Fig. 14(c)] The baselines are described as optimizing 'only the transmission rate' or 'only ... HDC performance,' while the proposed design minimizes the full CPR. Since CPR is the ratio of the total cost to the product in Eq. (20), optimizing a single factor of that product is unlikely to minimize the ratio, so the proposed design's lower CPR is largely a property of the comparison protocol rather than evidence of a superior joint trade-off. A meaningful comparison should optimize the same aggregate metric under the same functional forms for all designs, or report the full Pareto frontier between communication performance and learning performance.
minor comments (5)
  1. [Section VI-C.4] The simulation text says the 'channel fading coefficient' is '128.1 + 37.6 log10 dj', which is a path-loss expression in dB, whereas h_j in Eqs. (15) and (17) is used as a complex channel coefficient in |h_j|^2. The relation between the simulated value and the rate formula should be clarified.
  2. [Section VI-C.1] The notation T_c^j = sum_i A_ij T_i is an abuse of notation, since it sums task labels rather than defining a set of tasks; similar remarks apply to K_c^j. Using index sets would improve readability.
  3. [Abstract and Section VI-C.3] The abstract promises 'a practical and effective solution,' but Section VI-C.3 describes the case study as a 'preliminary investigation' serving as a 'foundational reference for future work.' These statements should be aligned to avoid overclaiming.
  4. [Section VI-C.2, Eq. (20)] The communication term in Q is a sum over all clients while the HDC term is an average over tasks; the asymmetry affects the scaling of the CPR and should be justified, since it influences the comparison across designs.
  5. [Throughout] There are minor copy-editing issues, including 'UA V' in Table II and 'the the number of iterations' in Section VI-C.3.

Circularity Check

2 steps flagged · score 8.0 of 10

The case-study 'dual-function wins' claim is built into the CPR objective: the joint design minimizes the very metric on which it is compared, and the model-size variable is decoupled from the accuracy term.

  1. self definitional [Section VI-C4, Fig. 14(c)]
    "Two baseline designs are considered: a communication-oriented baseline that optimizes only the transmission rate in (20), and a learning-oriented baseline that focuses solely on enhancing HDC performance. ... Moreover, the proposed dual function design, which jointly optimizes communication and learning objectives, achieves lower CPR values for all bandwidths. It consistently outperforms both the communication-oriented and learning-oriented designs, thereby demonstrating the effectiveness of the proposed joint optimization strategy."

    The proposed design solves problem (21), whose objective is exactly the CPR being reported, while the baselines maximize only one multiplicative factor inside the denominator Q of (20). Any feasible baseline point is feasible for (21), so a solver for (21) must return a CPR no larger than that baseline point by definition of minimization. The 'outperformance' in Fig. 14(c) therefore restates the objective rather than providing independent evidence about HDC, large-AI-model compression, or ILAC. The baselines are also not specified as solving constrained versions of (21), so the comparison is uncontrolled.

  2. other [Section VI-C2, Eqs. (19), (20), (21d)]
    "the computational cost for each client Cj is related to the compression ratio, expressed as ec_j = η(s0_j/sc_j) ... Q = ... where ϕ(·) measures the accuracy of the HDC model and can be determined by the accuracy on the validation set. ... sc_j ≤ Σ_i A_ij KiD, ∀j."

    In the formalized objective, the optimized compressed-model size sc_j enters only the cost function η(·) and the constraints (21d)-(21e); the accuracy term ϕ(˜Pc_j) in (20) is not written as a function of sc_j. As written, the optimizer can change sc_j without changing the HDC performance term, so the claimed joint 'optimal balance between learning performance and communication constraints' is not actually derived from a modeled trade-off. Neither η(·), ϕ(·), nor the compression map f_p(·) in Eq. (14) is specified or validated, so the quantitative CPR curves cannot support the conclusion that the dual-function design wins in practice.

full rationale

The survey portions of the paper (taxonomy of model/data distribution, reviews of SL/FL/MARL, HDC background) are not circular; they are expository. Self-citations such as [43] and [107] are used as survey references and are not load-bearing for the central case-study claim. The circularity is confined to Section VI-C. The abstract's claim of 'a practical and effective solution to achieve an optimal balance' rests on the case study, and that case study's quantitative conclusion is forced by construction: the joint design minimizes the CPR metric and is then compared on that same metric against baselines that optimize only one factor of it. Additionally, the model-size variable is decoupled from the accuracy function, so the optimization does not actually implement a learning-communication trade-off; the 'balance' is an artifact of an unspecified/decoupled formulation. The paper's own caveat that AO only guarantees a stationary point does not address this definitional issue. For these reasons, the central validation claim reduces by definition rather than by measurement, warranting a score of 8. If the case study had been framed only as a symbolic illustration without claiming quantitative superiority, the score would be lower.

Assumptions & free parameters 4 free parameters · 4 assumptions · 2 invented entities

The case study's quantitative claims rest on several unspecified functions and a toy simulation, so the ledger is dominated by free functions rather than fitted constants.

free parameters (4)
  • accuracy function phi(.) = unspecified
    Used in Eq. (20) to measure HDC model accuracy; stated to be 'determined by the accuracy on the validation set' but no functional form or data is given.
  • compression cost function eta(.) = unspecified
    Used in Eq. (19) to model the computational cost of foundation-model compression; no formula or calibration is provided.
  • foundation-model preprocessor fp(.) = unspecified
    Used in Eq. (14) to compress HDC models before transmission; treated as a black box without specification.
  • hypervector dimension D = not stated in simulation
    Appears in Eq. (13) for HDC model size; the simulation does not report the value of D.
assumptions (4)
  • standard math Dinkelbach and alternating optimization converge to a stationary point
    Section VI-C3 states that global optimality is only guaranteed if each subproblem is solved exactly, and convergence is to a stationary point.
  • domain assumption Random hypervectors are nearly orthogonal in high dimension
    Used in Section VI-B to justify HDC similarity and robustness, citing references [132]-[134].
  • domain assumption Validation-set accuracy phi(.) represents true HDC inference accuracy
    Eq. (20) uses this accuracy in the objective without quantifying generalization error.
  • ad hoc to paper Foundation model can compress HDC models without significant loss
    Eqs. (14) and (19) assume compression reduces transmission cost while retaining accuracy, but no compression mechanism or loss analysis is provided.
invented entities (2)
  • Cost-to-Performance Ratio (CPR)
    purpose: Unified metric that balances computational cost and communication/learning performance in problem (21).
    Defined ad hoc in Section VI-C2; no external benchmark or independent support is given.
  • Foundation-model-based preprocessing function fp(.)
    purpose: Black-box compression/fragmentation of HDC models prior to transmission in Eq. (14).
    No concrete algorithm, model, or data supports its existence or performance.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A New Pathway to Integrated Learning and Communication (ILAC): Large AI Model and Hyperdimensional Computing for Communication." pith.science (2026). https://pith.science/paper/6CKCVER2

@misc{pith2026250618432,
  author       = {Pith},
  title        = {Pith review of: A New Pathway to Integrated Learning and Communication (ILAC): Large AI Model and Hyperdimensional Computing for Communication},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6CKCVER2}},
  note         = {Machine review of arXiv:2506.18432}
}
read the original abstract

The rapid evolution of forthcoming sixth-generation (6G) wireless networks necessitates the seamless integration of artificial intelligence (AI) with wireless communications to support emerging intelligent applications that demand both efficient communication and robust learning performance. This dual requirement calls for a unified framework of integrated learning and communication (ILAC), where AI enhances communication through intelligent signal processing and adaptive resource management, while wireless networks support AI model deployment by enabling efficient and reliable data exchanges. However, achieving this integration presents significant challenges in practice. Communication constraints, such as limited bandwidth and fluctuating channels, hinder learning accuracy and convergence. Simultaneously, AI-driven learning dynamics, including model updates and task-driven inference, introduce excessive burdens on communication systems, necessitating flexible, context-aware transmission strategies. Finally, we present a case study on a cost-to-performance optimization problem, where task assignments, model size selection, bandwidth allocation, and transmission power control are jointly optimized, considering computational cost, communication efficiency, and inference accuracy. Leveraging the Dinkelbach and alternating optimization algorithms, we offer a practical and effective solution to achieve an optimal balance between learning performance and communication constraints.

Figures

Figures reproduced from arXiv: 2506.18432 by the authors.

Figure 1
Figure 1. Illustration of ILAC. TABLE I: Comparison of Key Surveys and Our Contribution Year Ref. Core Contribution Technical Focus 2017 [32] Pioneering survey on mobile edge computing (MEC) architectures and computation offloading strategies MEC infrastructure; computation offloading 2018 [33] Foundational taxonomy for contextual intelligence and data analytics in IoT Context-aware computing in IoT ecosystems 2019 [34] Syste… view at source ↗
Figure 2
Figure 2. Structure of the paper: An overview of the main topics. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Illustration of the unified ILAC framework. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: Illustration of the relationship between communica [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Overall performance versus complexity with different [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Architecture of SL for heterogeneous wireless networks. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Three main paradigms of aggregation learning in practical wireless communication scenarios. [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: Structures of a conventional wireless communication system and intelligent transceiver: (a) Symmetrical intelligent [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Workflow of low-rank adaptation fine-tuning for large [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Architecture and workflow for split federated learning [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]
Figure 11
Figure 11. Figure 11: Adaptive resource allocation via MARL for wireless [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 12
Figure 12. Figure 12: A workflow of HDC encoding, training, and inferenc [PITH_FULL_IMAGE:figures/full_fig_p017_12.png]
Figure 13
Figure 13. Figure 13: The novel framework for effective ILAC encompasses HDC encoding, local training, and compression via a foundation [PITH_FULL_IMAGE:figures/full_fig_p019_13.png]
Figure 14
Figure 14. Figure 14: (a) Transmission rate vs. number of iterations for different [PITH_FULL_IMAGE:figures/full_fig_p021_14.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. Energy Efficient Federated Learning with Hyperdimensional Computing over Wireless Communication Networks

    cs.DC 2026-02 conditional novelty 5.0 of 10

    Jointly optimizing the hypervector dimension together with bandwidth, power, transmission time, and CPU frequency can cut simulated federated-learning energy by up to 83.3% versus fixed allocations, at comparable accu...

Reference graph

Works this paper leans on

145 extracted references · 71 canonical work pages · cited by 1 Pith paper

  1. [1]

    A vision of 6G wireless systems: Applications, trends, technologies, and open research problems,

    W. Saad, M. Bennis, and M. Chen, “A vision of 6G wireless systems: Applications, trends, technologies, and open research problems,” IEEE Network, vol. 34, no. 3, pp. 134–142, May 2020

  2. [2]

    Robust beamforming for RIS-aided communications: Gradient-based manifold meta learning,

    F. Zhu, X. Wang, C. Huang, Z. Yang, X. Chen, A. Alhammadi et al., “Robust beamforming for RIS-aided communications: Gradient-based manifold meta learning,” IEEE Trans. Wireless Commun. , vol. 23, no. 11, pp. 15 945–15 956, Nov. 2024

  3. [3]

    A joint communication and computation design for semantic wireless communication with probability graph,

    Z. Zhao, Z. Yang, X. Gan, Q.-V . Pham, C. Huang, W. Xu et al. , “A joint communication and computation design for semantic wireless communication with probability graph,” J. Franklin Inst. , vol. 361, no. 13, p. 107055, Sep. 2024

  4. [4]

    Robust deep learning-based physical layer communications: Strategies and approaches,

    F. Zhu, X. Wang, C. Zhu, T. Gong, Z. Yang, C. Huang et al., “Robust deep learning-based physical layer communications: Strategies and approaches,” IEEE Network, early access, 2025

  5. [5]

    Multi- sources fusion learning for multi-points NLOS localization in OFDM system,

    B. Wang, Z. Shuai, C. Huang, Q. Yang, Z. Yang, R. Jin et al., “Multi- sources fusion learning for multi-points NLOS localization in OFDM system,” IEEE J. Sel. Topics Signal Process. , vol. 18, no. 7, pp. 1339– 1350, Sep. 2024. 24

  6. [6]

    Robust beamforming with gradient-based liquid neural network,

    X. Wang, F. Zhu, C. Huang, A. Alhammadi, F. Bader, Z. Zhang et al., “Robust beamforming with gradient-based liquid neural network,” IEEE Wireless Commun. Lett. , vol. 13, no. 11, pp. 3020–3024, Nov. 2024

  7. [7]

    Robust continuous-time beam tracking with liquid neural network,

    F. Zhu, X. Wang, C. Huang, R. Jin, Q. Yang, A. Alhammadi et al. , “Robust continuous-time beam tracking with liquid neural network,” in Proc. IEEE Global Commu. Conf. , Cape Town, South Africa, 2024, pp. 4878–4883

  8. [8]

    Technique for recovering wavefront phase bad points by deep learning,

    J. Wu, J. Liang, S. Fei, and X. Zhong, “Technique for recovering wavefront phase bad points by deep learning,” Chin. J. Electron. , vol. 32, no. 2, pp. 303–312, Mar. 2023

Show all 145 references
  1. [9]

    A joint learning and communications framework for federated learning over wireless networks,

    M. Chen, Z. Yang, W. Saad, C. Yin, H. V . Poor, and S. Cui, “A joint learning and communications framework for federated learning over wireless networks,” IEEE Trans. Wireless Commun., vol. 20, no. 1, pp. 269–283, Jan. 2021

  2. [10]

    Com- pression ratio allocation for probabilistic semantic communication with RSMA,

    Z. Zhao, Z. Yang, Y . Hu, C. Zhu, M. Shikh-Bahaei, W. Xuet al., “Com- pression ratio allocation for probabilistic semantic communication with RSMA,” IEEE Trans. Commun. , early access, 2025

  3. [11]

    Beamforming inferring by conditional WGAN-GP for holographic antenna arrays,

    F. Zhu, X. Wang, C. Huang, A. Alhammadi, H. Chen, Z. Zhang et al., “Beamforming inferring by conditional WGAN-GP for holographic antenna arrays,” IEEE Wireless Commun. Lett. , vol. 13, no. 7, pp. 2023–2027, Jul. 2024

  4. [12]

    Energy- efficient probabilistic semantic communication over space-air-ground integrated networks,

    Z. Zhao, Z. Yang, M. Chen, C. Zhu, W. Xu, Z. Zhang et al., “Energy- efficient probabilistic semantic communication over space-air-ground integrated networks,” IEEE Trans. Wireless Commun. , early access, 2025

  5. [13]

    Data-importance-aware power allocation for adaptive real-time communication in computer vision applications,

    C. Xu, Y . Ma, R. Tafazolli, and J. Wang, “Data-importance-aware power allocation for adaptive real-time communication in computer vision applications,” arXiv preprint arXiv:2504.08922, 2025. [Online]. Available: https://arxiv.org/abs/2504.08922

  6. [14]

    Radio resource allocation techniques for efficient spectrum access in cognitive radio networks,

    G. I. Tsiropoulos, O. A. Dobre, M. H. Ahmed, and K. E. Baddour, “Radio resource allocation techniques for efficient spectrum access in cognitive radio networks,” IEEE Commun. Surveys Tuts., vol. 18, no. 1, pp. 824–847, 1st Quart. 2016

  7. [15]

    Imperfect CSI: A key factor of uncertainty to over-the-air federated learning,

    J. Yao, Z. Yang, W. Xu, D. Niyato, and X. You, “Imperfect CSI: A key factor of uncertainty to over-the-air federated learning,” IEEE Wireless Commun. Lett., vol. 12, no. 12, pp. 2273–2277, Dec. 2023

  8. [16]

    An edge-enabled wireless split learning testbed: Design and implementation,

    Z. Wang, L. Boccardo, and Y . Deng, “An edge-enabled wireless split learning testbed: Design and implementation,” IEEE Commun. Lett. , vol. 28, no. 6, pp. 1337–1341, Jun. 2024

  9. [17]

    Similarity-based label inference attack against training and inference of split learning,

    J. Liu, X. Lyu, Q. Cui, and X. Tao, “Similarity-based label inference attack against training and inference of split learning,” IEEE Trans. Inf. Forensics Secur., vol. 19, pp. 2881–2895, Jan. 2024

  10. [18]

    State aggregation learning from markov transition data,

    Y . Duan, T. Ke, and M. Wang, “State aggregation learning from markov transition data,” in Adv. Neural Inf. Process. Syst. , vol. 32, Vancouver, BC, Canada, 2019

  11. [19]

    Software-defined transceivers in dy- namic access networks,

    D. Hillerkuss and J. Leuthold, “Software-defined transceivers in dy- namic access networks,” J. Lightwave Technol. , vol. 34, no. 2, pp. 792–797, Jan. 2016

  12. [20]

    Advanced modulation techniques for flexible optical transceivers: The rate/reach tradeoff,

    G. Bosco, “Advanced modulation techniques for flexible optical transceivers: The rate/reach tradeoff,” J. Lightwave Technol. , vol. 37, no. 1, pp. 36–49, Jan. 2019

  13. [21]

    Intelligent reflection enabling technologies for integrated and green Internet-of-Everything beyond 5G: Communication, sensing, and security,

    W. Shi, W. Xu, X. You, C. Zhao, and K. Wei, “Intelligent reflection enabling technologies for integrated and green Internet-of-Everything beyond 5G: Communication, sensing, and security,” IEEE Wireless Commun., vol. 30, no. 2, pp. 147–154, Apr. 2023

  14. [22]

    Big AI models for 6G wireless networks: Opportunities, challenges, and research directions,

    Z. Chen, Z. Zhang, and Z. Yang, “Big AI models for 6G wireless networks: Opportunities, challenges, and research directions,” IEEE Wireless Commun., vol. 31, no. 5, pp. 164–172, Oct. 2024

  15. [23]

    A survey on hyperdimensional computing aka vector symbolic architectures, part I: Models and data transformations,

    D. Kleyko, D. A. Rachkovskij, E. Osipov, and A. Rahimi, “A survey on hyperdimensional computing aka vector symbolic architectures, part I: Models and data transformations,” ACM Comput. Surv., vol. 55, no. 6, pp. 1–40, Dec. 2022

  16. [24]

    A survey on hyperdimensional computing aka vector symbolic architectures, part II: Applications, cognitive models, and challenges,

    D. Kleyko, D. Rachkovskij, E. Osipov, and A. Rahimi, “A survey on hyperdimensional computing aka vector symbolic architectures, part II: Applications, cognitive models, and challenges,” ACM Comput. Surv., vol. 55, no. 9, pp. 1–52, Jan. 2023

  17. [25]

    A joint communication and computation design for distributed RISs assisted probabilistic semantic communication in IIoT,

    Z. Zhao, Z. Yang, C. Huang, L. Wei, Q. Yang, C. Zhong et al., “A joint communication and computation design for distributed RISs assisted probabilistic semantic communication in IIoT,” IEEE Internet Things J., vol. 11, no. 16, pp. 26 568–26 579, Aug. 2024

  18. [26]

    Wireless federated learning over resource-constrained networks: Digital versus analog transmissions,

    J. Yao, W. Xu, Z. Yang, X. You, M. Bennis, and H. V . Poor, “Wireless federated learning over resource-constrained networks: Digital versus analog transmissions,” IEEE Trans. Wireless Commun., vol. 23, no. 10, pp. 14 020–14 036, Oct. 2024

  19. [27]

    Multi-hop RIS-empowered terahertz communications: A DRL- based hybrid beamforming design,

    C. Huang, Z. Yang, G. C. Alexandropoulos, K. Xiong, L. Wei, C. Yuen et al., “Multi-hop RIS-empowered terahertz communications: A DRL- based hybrid beamforming design,” IEEE J. Sel. Areas Commun. , vol. 39, no. 6, pp. 1663–1677, Jun. 2021

  20. [28]

    Split learning in 6G edge networks,

    Z. Lin, G. Qu, X. Chen, and K. Huang, “Split learning in 6G edge networks,” IEEE Wireless Commun., vol. 31, no. 4, pp. 170–176, Aug. 2024

  21. [29]

    GoMORE: Global model reuse for resource-constrained wireless federated learning,

    J. Yao, Z. Yang, W. Xu, M. Chen, and D. Niyato, “GoMORE: Global model reuse for resource-constrained wireless federated learning,” IEEE Wireless Commun. Lett. , vol. 12, no. 9, pp. 1543–1547, Sep. 2023

  22. [30]

    Empowering over- the-air personalized federated learning via RIS,

    W. Shi, J. Yao, J. Xu, W. Xu, L. Xu, and C. Zhao, “Empowering over- the-air personalized federated learning via RIS,” Sci. China Inf. Sci. , vol. 67, no. 11, pp. 219 302:1–2, Nov. 2024

  23. [31]

    Learning to communicate with deep multi-agent reinforcement learning,

    J. Foerster, I. A. Assael, N. De Freitas, and S. Whiteson, “Learning to communicate with deep multi-agent reinforcement learning,” in Adv. Neural Inf. Process. Syst. , vol. 29, Barcelona, Spain, 2016

  24. [32]

    Mobile edge computing: A survey on ar- chitecture and computation offloading,

    P. Mach and Z. Becvar, “Mobile edge computing: A survey on ar- chitecture and computation offloading,” IEEE Commun. Surveys Tuts. , vol. 19, no. 3, pp. 1628–1656, 3rd Quart. 2017

  25. [33]

    Context-aware com- puting, learning, and big data in internet of things: A survey,

    O. B. Sezer, E. Dogdu, and A. M. Ozbayoglu, “Context-aware com- puting, learning, and big data in internet of things: A survey,” IEEE Internet Things J. , vol. 5, no. 1, pp. 1–27, Feb. 2018

  26. [34]

    Application of machine learning in wireless networks: Key techniques and open issues,

    Y . Sun, M. Peng, Y . Zhou, Y . Huang, and S. Mao, “Application of machine learning in wireless networks: Key techniques and open issues,” IEEE Commun. Surveys Tuts. , vol. 21, no. 4, pp. 3072–3108, 4th Quart. 2019

  27. [35]

    Mobile edge intelligence and computing for the internet of vehicles,

    J. Zhang and K. B. Letaief, “Mobile edge intelligence and computing for the internet of vehicles,” Proc. IEEE, vol. 108, no. 2, pp. 246–261, Feb. 2020

  28. [36]

    Federated learning in mobile edge networks: A comprehensive survey,

    W. Y . B. Lim, N. C. Luong, D. T. Hoang, Y . Jiao, Y .-C. Liang, Q. Yang et al., “Federated learning in mobile edge networks: A comprehensive survey,” IEEE Commun. Surveys Tuts. , vol. 22, no. 3, pp. 2031–2063, 3rd Quart. 2020

  29. [37]

    Federated machine learning: Survey, multi-level classification, desirable criteria and future directions in communication and networking systems,

    O. A. Wahab, A. Mourad, H. Otrok, and T. Taleb, “Federated machine learning: Survey, multi-level classification, desirable criteria and future directions in communication and networking systems,” IEEE Commun. Surveys Tuts., vol. 23, no. 2, pp. 1342–1397, 2nd Quart. 2021

  30. [38]

    Communication-efficient and distributed learning over wireless networks: Principles and applications,

    J. Park, S. Samarakoon, A. Elgabli, J. Kim, M. Bennis, S.-L. Kim et al., “Communication-efficient and distributed learning over wireless networks: Principles and applications,” Proc. IEEE, vol. 109, no. 5, pp. 796–819, May 2021

  31. [39]

    Resource scheduling in edge computing: A survey,

    Q. Luo, S. Hu, C. Li, G. Li, and W. Shi, “Resource scheduling in edge computing: A survey,” IEEE Commun. Surveys Tuts. , vol. 23, no. 4, pp. 2131–2165, 4th Quart. 2021

  32. [40]

    Distributed learning in wireless networks: Recent progress and future challenges,

    M. Chen, D. G ¨und¨uz, K. Huang, W. Saad, M. Bennis, A. V . Feljan et al., “Distributed learning in wireless networks: Recent progress and future challenges,” IEEE J. Sel. Areas Commun. , vol. 39, no. 12, pp. 3579–3605, Dec. 2021

  33. [41]

    A survey on federated learning for resource-constrained IoT devices,

    A. Imteaj, U. Thakker, S. Wang, J. Li, and M. H. Amini, “A survey on federated learning for resource-constrained IoT devices,” IEEE Internet Things J., vol. 9, no. 1, pp. 1–24, Jan. 2022

  34. [42]

    Edge artificial intelligence for 6G: Vision, enabling technologies, and applications,

    K. B. Letaief, Y . Shi, J. Lu, and J. Lu, “Edge artificial intelligence for 6G: Vision, enabling technologies, and applications,” IEEE J. Sel. Areas Commun., vol. 40, no. 1, pp. 5–36, Jan. 2022

  35. [43]

    Edge learning for B5G networks with distributed signal processing: Semantic communication, edge computing, and wireless sensing,

    W. Xu, Z. Yang, D. W. K. Ng, M. Levorato, Y . C. Eldar, and M. Debbah, “Edge learning for B5G networks with distributed signal processing: Semantic communication, edge computing, and wireless sensing,” IEEE J. Sel. Topics Signal Process. , vol. 17, no. 1, pp. 9–39, Jan. 2023

  36. [44]

    Energy-efficient edge inference in integrated sensing, communication, and computation networks,

    J. Yao, W. Xu, G. Zhu, K. Huang, and S. Cui, “Energy-efficient edge inference in integrated sensing, communication, and computation networks,” IEEE J. Sel. Areas Commun. , early access, May 2025

  37. [45]

    Sum rate maximization for distributed RISs assisted probabilistic semantic communication,

    Z. Zhao, Z. Yang, Q. Yang, C. Huang, M. Shikh-Bahaei, and Z. Zhang, “Sum rate maximization for distributed RISs assisted probabilistic semantic communication,” in Proc. IEEE Int. Workshop Mach. Learn. Signal Process., London, United Kingdom, 2024, pp. 1–6

  38. [46]

    A joint communication and computation design for probabilistic semantic communications,

    Z. Zhao, Z. Yang, M. Chen, Z. Zhang, and H. V . Poor, “A joint communication and computation design for probabilistic semantic communications,” Entropy, vol. 26, no. 5, p. 394, Apr. 2024

  39. [47]

    Power- efficient transmission for user-centric networks with limited fronthaul capacity and computation resource,

    J. Shi, X. Chen, N. Huang, H. Jiang, Z. Yang, and M. Chen, “Power- efficient transmission for user-centric networks with limited fronthaul capacity and computation resource,” IEEE Trans. on Commun., vol. 68, no. 9, pp. 5649–5660, 2020

  40. [48]

    Bi-objective optimiza- tion of information rate and harvested power in RIS-aided SWIPT 25 systems,

    A. Mohamed, A. Zappone, and M. Di Renzo, “Bi-objective optimiza- tion of information rate and harvested power in RIS-aided SWIPT 25 systems,” IEEE Wireless Commun. Lett. , vol. 11, no. 10, pp. 2195– 2199, Jan. 2022

  41. [49]

    Robust secure beamforming in MISO full-duplex two-way secure communications,

    R. Feng, Q. Li, Q. Zhang, and J. Qin, “Robust secure beamforming in MISO full-duplex two-way secure communications,” IEEE Trans. Veh. Technol., vol. 65, no. 1, pp. 408–414, Jan. 2016

  42. [50]

    Overview of multi-objective optimization methods,

    L. Xiujuan and S. Zhongke, “Overview of multi-objective optimization methods,” J. Syst. Eng. Electron., vol. 15, no. 2, pp. 142–146, Jun. 2004

  43. [51]

    Weighted sum power maximization for STAR-RIS assisted SWIPT systems,

    Y . Li, J. Wang, Y . Zou, W. Xie, and Y . Liu, “Weighted sum power maximization for STAR-RIS assisted SWIPT systems,” IEEE Trans. Wireless Commun., vol. 23, no. 12, pp. 18 394–18 408, 2024

  44. [52]

    Fair resource allo- cation schemes for cooperative dynamic free-space optical networks,

    A. S. Ghazy, H. A. I. Selmy, and H. M. H. Shalaby, “Fair resource allo- cation schemes for cooperative dynamic free-space optical networks,” J. Opt. Commun. Netw. , vol. 8, no. 11, pp. 822–834, Nov. 2016

  45. [53]

    Privacy and efficiency of communications in federated split learning,

    Z. Zhang, A. Pinto, V . Turina, F. Esposito, and I. Matta, “Privacy and efficiency of communications in federated split learning,” IEEE Trans. Big Data, vol. 9, no. 5, pp. 1380–1391, Oct. 2023

  46. [54]

    Optimizing split points for error-resilient splitfed learning,

    C. Shiranthika, P. Saeedi, and I. V . Baji ´c, “Optimizing split points for error-resilient splitfed learning,” 2024. [Online]. Available: https://arxiv.org/abs/2405.19453

  47. [55]

    Accelerating split federated learning over wireless communication networks,

    C. Xu, J. Li, Y . Liu, Y . Ling, and M. Wen, “Accelerating split federated learning over wireless communication networks,” IEEE Trans. Wireless Commun., vol. 23, no. 6, pp. 5587–5599, Jun. 2023

  48. [56]

    Splitfed: When federated learning meets split learning,

    C. Thapa, P. C. M. Arachchige, S. Camtepe, and L. Sun, “Splitfed: When federated learning meets split learning,” in Proc. AAAI Conf. Artif. Intell., vol. 36, no. 8, virtual event, 2022, pp. 8485–8493

  49. [57]

    Wireless distributed learning: A new hybrid split and federated learning approach,

    X. Liu, Y . Deng, and T. Mahmoodi, “Wireless distributed learning: A new hybrid split and federated learning approach,” IEEE Trans. Wireless Commun., vol. 22, no. 4, pp. 2650–2665, Apr. 2022

  50. [58]

    Joint opti- mization of model partition and resource allocation for split federated learning over vehicular edge networks,

    M. Wu, R. Yang, X. Huang, Y . Wu, J. Kang, and S. Xie, “Joint opti- mization of model partition and resource allocation for split federated learning over vehicular edge networks,” IEEE Trans. Veh. Technol. , vol. 73, no. 10, pp. 15 860–15 865, Oct. 2024

  51. [59]

    Federated split learning with joint personalization-generalization for inference-stage optimization in wireless edge networks,

    D.-J. Han, D.-Y . Kim, M. Choi, D. Nickel, J. Moon, M. Chiang et al., “Federated split learning with joint personalization-generalization for inference-stage optimization in wireless edge networks,” IEEE Trans. Mob. Comput., vol. 23, no. 6, pp. 7048–7065, Jun. 2024

  52. [60]

    Ef- ficient parallel split learning over resource-constrained wireless edge networks,

    Z. Lin, G. Zhu, Y . Deng, X. Chen, Y . Gao, K. Huang et al. , “Ef- ficient parallel split learning over resource-constrained wireless edge networks,” IEEE Trans. Mob. Comput., vol. 23, no. 10, pp. 9224–9239, Oct. 2024

  53. [61]

    Over- the-air split machine learning in wireless MIMO networks,

    Y . Yang, Z. Zhang, Y . Tian, Z. Yang, C. Huang, C. Zhonget al., “Over- the-air split machine learning in wireless MIMO networks,” IEEE J. Sel. Areas Commun. , vol. 41, no. 4, pp. 1007–1022, Apr. 2023

  54. [62]

    Adaptive split learning over energy-constrained wireless edge networks,

    Z. Li, W. Wu, S. Wu, and W. Wang, “Adaptive split learning over energy-constrained wireless edge networks,” in Proc. IEEE Conf. Comput. Commun. Workshops, Vancouver, BC, Canada, 2024, pp. 1–6

  55. [63]

    Adaptive layer splitting for wireless llm inference in edge computing: A model-based reinforcement learning approach,

    Y . Chen, R. Li, X. Yu, Z. Zhao, and H. Zhang, “Adaptive layer splitting for wireless llm inference in edge computing: A model-based reinforcement learning approach,” Front. Inform. Technol. Electron. Eng., vol. 26, pp. 278–292, Mar. 2025

  56. [64]

    Splitllm: Hierarchical split learning for large language model over wireless network,

    S. Zhang, G. Cheng, Z. Li, and W. Wu, “Splitllm: Hierarchical split learning for large language model over wireless network,” arXiv preprint arXiv:2501.13318 , 2025. [Online]. Available: https: //arxiv.org/abs/2501.13318

  57. [65]

    Instance-conditional timescales of decay for non-stationary learning,

    N. Jain and P. Shenoy, “Instance-conditional timescales of decay for non-stationary learning,” in Proc. AAAI Conf. Artif. Intell. , vol. 38, no. 11, Vancouver, BC, Canada, 2024, pp. 12 773–12 781

  58. [66]

    Deep learning based mod- ulation recognition with multi-cue fusion,

    T. Wang, Y . Hou, H. Zhang, and Z. Guo, “Deep learning based mod- ulation recognition with multi-cue fusion,” IEEE Wireless Commun. Lett., vol. 10, no. 8, pp. 1757–1760, Aug. 2021

  59. [67]

    On ensemble learning- based secure fusion strategy for robust cooperative sensing in full- duplex cognitive radio networks,

    Y . Zhang, Q. Wu, and M. R. Shikh-Bahaei, “On ensemble learning- based secure fusion strategy for robust cooperative sensing in full- duplex cognitive radio networks,” IEEE Trans. Commun. , vol. 68, no. 10, pp. 6086–6100, Oct. 2020

  60. [68]

    Deep-ensemble- learning-based gps spoofing detection for cellular-connected uavs,

    Y . Dang, C. Benza¨ıd, B. Yang, T. Taleb, and Y . Shen, “Deep-ensemble- learning-based gps spoofing detection for cellular-connected uavs,” IEEE Internet Things J. , vol. 9, no. 24, pp. 25 068–25 085, Dec. 2022

  61. [69]

    A survey on neural network interpretability,

    Y . Zhang, P. Ti ˇno, A. Leonardis, and K. Tang, “A survey on neural network interpretability,” IEEE Trans. Emerg. Top. Comput. Intell. , vol. 5, no. 5, pp. 726–742, Oct. 2021

  62. [70]

    Deep learning- enabled semantic communication systems with task-unaware transmit- ter and dynamic data,

    H. Zhang, S. Shao, M. Tao, X. Bi, and K. B. Letaief, “Deep learning- enabled semantic communication systems with task-unaware transmit- ter and dynamic data,” IEEE J. Sel. Areas Commun. , vol. 41, no. 1, pp. 170–185, Jan. 2023

  63. [71]

    Deep learning-based end-to-end wireless communication systems with conditional gans as unknown channels,

    H. Ye, L. Liang, G. Y . Li, and B.-H. Juang, “Deep learning-based end-to-end wireless communication systems with conditional gans as unknown channels,” IEEE Trans. Wireless Commun. , vol. 19, no. 5, pp. 3133–3143, May 2020

  64. [72]

    End-to-end learning of communications systems without a channel model,

    F. A. Aoudia and J. Hoydis, “End-to-end learning of communications systems without a channel model,” in Proc. Asilomar Conf. Signals Syst. Comput., Pacific Grove, CA, USA, Oct. 2018, pp. 298–303

  65. [73]

    Adaptive spatial modulation MIMO based on machine learning,

    P. Yang, Y . Xiao, M. Xiao, Y . L. Guan, S. Li, and W. Xiang, “Adaptive spatial modulation MIMO based on machine learning,” IEEE J. Sel. Areas Commun., vol. 37, no. 9, pp. 2117–2131, Sep. 2019

  66. [74]

    Distributed machine learning based downlink channel estimation for RIS assisted wireless communications,

    L. Dai and X. Wei, “Distributed machine learning based downlink channel estimation for RIS assisted wireless communications,” IEEE Trans. Commun., vol. 70, no. 7, pp. 4900–4909, Jul. 2022

  67. [75]

    Learning-based massive beamforming,

    S. Lu, S. Zhao, and Q. Shi, “Learning-based massive beamforming,” in Proc. IEEE Global Commun. Conf. , Taipei, China, 2020, pp. 1–6

  68. [76]

    Power of deep learning for channel estimation and signal detection in OFDM systems,

    H. Ye, G. Y . Li, and B.-H. Juang, “Power of deep learning for channel estimation and signal detection in OFDM systems,” IEEE Wireless Commun. Lett., vol. 7, no. 1, pp. 114–117, Feb. 2018

  69. [77]

    Task-oriented low- label semantic communication with self-supervised learning,

    R. Gu, W. Xu, Z. Yang, D. Niyato, and A. Yener, “Task-oriented low- label semantic communication with self-supervised learning,” IEEE Trans. Wireless Commun., pp. 1–1, early access, 2025

  70. [78]

    On the doppler squint effect in otfs systems over doubly-dispersive channels: Modeling and evaluation,

    X. Wang, X. Shi, J. Wang, and J. Song, “On the doppler squint effect in otfs systems over doubly-dispersive channels: Modeling and evaluation,” IEEE Trans. Wireless Commun., vol. 22, no. 12, pp. 8781– 8796, Dec. 2023

  71. [79]

    Flexible delay-doppler domain multiple access for massive connectivity with high mobility,

    X. Wang, H. Zhang, J. Wang, Z. Yang, H. Lin, and J. Song, “Flexible delay-doppler domain multiple access for massive connectivity with high mobility,” IEEE Trans. Wireless Commun. , early access, 2025

  72. [80]

    Closer twins model: Consistent design of modem scheme and channel estimation under high-mobility scenarios,

    H. Zhang, X. Wang, J. Tan, J. Wang, Z. Yang, and B. Ai, “Closer twins model: Consistent design of modem scheme and channel estimation under high-mobility scenarios,” IEEE Trans. Wireless Commun. , early access, 2025

  73. [81]

    STAR-RIS integrated nonorthogonal multiple access and over-the-air federated learning: Framework, analysis, and optimization,

    W. Ni, Y . Liu, Y . C. Eldar, Z. Yang, and H. Tian, “STAR-RIS integrated nonorthogonal multiple access and over-the-air federated learning: Framework, analysis, and optimization,”IEEE Internet Things J., vol. 9, no. 18, pp. 17 136–17 156, Sep. 2022

  74. [82]

    Semi-federated learning for collabo- rative intelligence in massive IoT networks,

    W. Ni, J. Zheng, and H. Tian, “Semi-federated learning for collabo- rative intelligence in massive IoT networks,” IEEE Internet Things J. , vol. 10, no. 13, pp. 11 942–11 943, Jul. 2023

  75. [83]

    Byzantine-resilient over-the-air federated learning under zero-trust architecture,

    J. Yao, W. Shi, W. Xu, Z. Yang, A. L. Swindlehurst, and D. Niyato, “Byzantine-resilient over-the-air federated learning under zero-trust architecture,” IEEE J. Sel. Areas Commun. , vol. 43, no. 6, pp. 1954– 1969, Apr. 2025

  76. [84]

    Low-parameter federated learning with large language models,

    J. Jiang, X. Liu, and C. Fan, “Low-parameter federated learning with large language models,” arXiv preprint arXiv:2307.13896 , 2023. [Online]. Available: https://arxiv.org/abs/2307.13896

  77. [85]

    pFedLoRA: Model-heterogeneous personalized federated learning with LoRA tuning,

    L. Yi, H. Yu, G. Wang, X. Liu, and X. Li, “pFedLoRA: Model-heterogeneous personalized federated learning with LoRA tuning,” arXiv preprint arXiv:2310.13283 , 2023. [Online]. Available: https://arxiv.org/abs/2310.13283

  78. [86]

    Splitlora: A split parameter-efficient fine-tuning framework for large language models,

    Z. Lin, X. Hu, Y . Zhang, Z. Chen, Z. Fang, X. Chen et al., “Splitlora: A split parameter-efficient fine-tuning framework for large language models,” arXiv preprint arXiv:2407.00952 , 2024. [Online]. Available: https://arxiv.org/abs/2407.00952

  79. [87]

    Fedsllm: Federated split learning for large language models over communication networks,

    K. Zhao, Z. Yang, C. Huang, X. Chen, and Z. Zhang, “Fedsllm: Federated split learning for large language models over communication networks,” in Proc. Int. Conf. Ubiquitous Commun. , Xi’an, China, 2024, pp. 438–443

  80. [88]

    Lyapunov-based optimization of edge resources for energy-efficient adaptive federated learning,

    C. Battiloro, P. Di Lorenzo, M. Merluzzi, and S. Barbarossa, “Lyapunov-based optimization of edge resources for energy-efficient adaptive federated learning,” IEEE Trans. Green Commun. Netw. , vol. 7, no. 1, pp. 265–280, Jun. 2022

  81. [89]

    Multi-agent reinforcement learning: Independent vs. cooper- ative agents,

    M. Tan, “Multi-agent reinforcement learning: Independent vs. cooper- ative agents,” in Mach. Learn. Proc. , 1993, pp. 330–337

  82. [90]

    Markov games as a framework for multi-agent reinforcement learning,

    M. L. Littman, “Markov games as a framework for multi-agent reinforcement learning,” in Mach. Learn. Proc. , 1994, pp. 157–163

  83. [91]

    Communication-efficient decentralized multi-agent reinforcement learning for cooperative adaptive cruise control,

    D. Chen, K. Zhang, Y . Wang, X. Yin, Z. Li, and D. Filev, “Communication-efficient decentralized multi-agent reinforcement learning for cooperative adaptive cruise control,” IEEE Trans. Intell. Vehicles, early access, 2024

  84. [92]

    Communication-efficient MARL for platoon stability and energy- efficiency co-optimization in cooperative adaptive cruise control of CA Vs,

    M. Hua, D. Chen, K. Jiang, F. Zhang, J. Wang, B. Wang et al. , “Communication-efficient MARL for platoon stability and energy- efficiency co-optimization in cooperative adaptive cruise control of CA Vs,”arXiv preprint arXiv:2406.11653 , 2024. [Online]. Available: https://arxiv...

  85. [93]

    Hypercomm: Hypergraph- based communication in multi-agent reinforcement learning,

    T. Zhu, X. Shi, X. Xu, J. Gui, and J. Cao, “Hypercomm: Hypergraph- based communication in multi-agent reinforcement learning,” Neural Netw., p. 106432, Oct. 2024. 26

  86. [94]

    Federated learning for wireless communications: Motivation, opportunities, and challenges,

    S. Niknam, H. S. Dhillon, and J. H. Reed, “Federated learning for wireless communications: Motivation, opportunities, and challenges,” IEEE Commun. Mag. , vol. 58, no. 6, pp. 46–51, Jun. 2020

  87. [95]

    Emergent communication in multi-agent reinforcement learning for future wireless networks,

    M. Chafii, S. Naoumi, R. Alami, E. Almazrouei, M. Bennis, and M. Debbah, “Emergent communication in multi-agent reinforcement learning for future wireless networks,” IEEE Internet Things Mag. , vol. 6, no. 4, pp. 18–24, Dec. 2023

  88. [96]

    Interacting federated and transfer learning-aided CSI prediction for intelligent cellular networks,

    J. Sun, Y . Zhang, G. Gui, H. Zhao, H. Gacanin, and H. Sari, “Interacting federated and transfer learning-aided CSI prediction for intelligent cellular networks,” IEEE Trans. Veh. Technol. , vol. 72, no. 12, pp. 15 776–15 787, Dec. 2023

  89. [97]

    Multi- agent reinforcement learning resources allocation method using dueling double deep q-network in vehicular networks,

    Y . Ji, Y . Wang, H. Zhao, G. Gui, H. Gacanin, H. Sari et al. , “Multi- agent reinforcement learning resources allocation method using dueling double deep q-network in vehicular networks,” IEEE Trans. Veh. Technol., vol. 72, no. 10, pp. 13 447–13 460, Oct. 2023

  90. [98]

    Multiagent deep-reinforcement- learning-based resource allocation for heterogeneous qos guarantees for vehicular networks,

    J. Tian, Q. Liu, H. Zhang, and D. Wu, “Multiagent deep-reinforcement- learning-based resource allocation for heterogeneous qos guarantees for vehicular networks,” IEEE Internet Things J. , vol. 9, no. 3, pp. 1683– 1695, Feb. 2021

  91. [99]

    Scheduling pattern of time triggered ethernet based on reinforcement learning,

    F. He, L. Xiong, X. Zhou, H. Li, and H. Xiong, “Scheduling pattern of time triggered ethernet based on reinforcement learning,” Chin. J. Electron., vol. 32, no. 5, pp. 1022–1035, Sep. 2023

  92. [100]

    Reconfig- urable intelligent surface for Internet of Robotic Things,

    W. Ni, R. Luo, X. Zhang, P. Wang, W. Wang, and H. Tian, “Reconfig- urable intelligent surface for Internet of Robotic Things,” IEEE Internet Things Mag., vol. 8, no. 2, pp. 78–86, Mar. 2025

  93. [101]

    Secure outage analysis of RIS-assisted communications with discrete phase control,

    W. Shi, J. Xu, W. Xu, M. Di Renzo, and C. Zhao, “Secure outage analysis of RIS-assisted communications with discrete phase control,” IEEE Trans. Veh. Technol., vol. 72, no. 4, pp. 5435–5440, Apr. 2023

  94. [102]

    Integrating over-the- air federated learning and non-orthogonal multiple access: What role can RIS play?

    W. Ni, Y . Liu, Z. Yang, H. Tian, and X. Shen, “Integrating over-the- air federated learning and non-orthogonal multiple access: What role can RIS play?” IEEE Trans. Wireless Commun. , vol. 21, no. 12, pp. 10 083–10 099, Dec. 2022

  95. [103]

    On secrecy performance of RIS-assisted MISO systems over Rician channels with spatially random eavesdroppers,

    W. Shi, J. Xu, W. Xu, C. Yuen, A. L. Swindlehurst, and C. Zhao, “On secrecy performance of RIS-assisted MISO systems over Rician channels with spatially random eavesdroppers,” IEEE Trans. Wireless Commun., vol. 23, no. 8, pp. 8357–8371, Aug. 2024

  96. [104]

    Federated learning in multi-RIS-aided systems,

    W. Ni, Y . Liu, Z. Yang, H. Tian, and X. Shen, “Federated learning in multi-RIS-aided systems,” IEEE Internet Things J. , vol. 9, no. 12, pp. 9608–9624, Jun. 2021

  97. [105]

    Combating interference for over-the-air federated learning: A statistical approach via RIS,

    W. Shi, J. Yao, W. Xu, J. Xu, X. You, Y . C. Eldar et al., “Combating interference for over-the-air federated learning: A statistical approach via RIS,” IEEE Trans. Signal Process., vol. 73, pp. 936–953, Jan. 2025

  98. [106]

    Stratified federated learning-based beamforming design for heterogeneous IRSs-aided net- works,

    H. Min, T. Zhou, T. Xu, Y . Huang, and H. Hu, “Stratified federated learning-based beamforming design for heterogeneous IRSs-aided net- works,” IEEE Wireless Commun. Lett. , vol. 12, no. 1, pp. 1–5, Jan. 2023

  99. [107]

    Edge learning for B5G networks with distributed signal processing: Semantic communication, edge computing, and wireless sensing,

    W. Xu, Z. Yang, D. W. K. Ng, M. Levorato, Y . C. Eldar, and M. Debbah, “Edge learning for B5G networks with distributed signal processing: Semantic communication, edge computing, and wireless sensing,” IEEE J. Sel. Top. Signal Process. , vol. 17, no. 1, pp. 9–39, Jan. 2023

  100. [108]

    Energy efficient federated learning over wireless communication networks,

    Z. Yang, M. Chen, W. Saad, C. S. Hong, and M. Shikh-Bahaei, “Energy efficient federated learning over wireless communication networks,” IEEE Trans. Wireless Commun. , vol. 20, no. 3, pp. 1935–1949, Mar. 2021

  101. [109]

    Rcnet: Incorporating structural information into deep rnn for online MIMO-OFDM sym- bol detection with limited training,

    Z. Zhou, L. Liu, S. Jere, J. Zhang, and Y . Yi, “Rcnet: Incorporating structural information into deep rnn for online MIMO-OFDM sym- bol detection with limited training,” IEEE Trans. Wireless Commun. , vol. 20, no. 6, pp. 3524–3537, Jan. 2021

  102. [110]

    Distributed auto-learning gnn for multi-cell cluster-free noma communications,

    X. Xu, Y . Liu, Q. Chen, X. Mu, and Z. Ding, “Distributed auto-learning gnn for multi-cell cluster-free noma communications,” IEEE J. Sel. Areas Commun., vol. 41, no. 4, pp. 1243–1258, Feb. 2023

  103. [111]

    CSI-GPT: Integrating generative pre-trained transformer with federated-tuning to acquire downlink massive mimo channels,

    Y . Zeng, L. Qiao, Z. Gao, T. Qin, Z. Wu, E. Khalaf et al., “CSI-GPT: Integrating generative pre-trained transformer with federated-tuning to acquire downlink massive mimo channels,” IEEE Trans. Veh. Technol., vol. 74, no. 3, pp. 5187–5192, 2025

  104. [112]

    Edge-LLM: Enabling efficient large language model adaptation on edge devices via unified compression and adaptive layer voting,

    Z. Yu, Z. Wang, Y . Li, R. Gao, X. Zhou, S. R. Bommu et al. , “Edge-LLM: Enabling efficient large language model adaptation on edge devices via unified compression and adaptive layer voting,” in ACM/IEEE Des. Autom. Conf. , San Francisco, CA, USA, 2024, pp. 1–6

  105. [113]

    Federated learning for 6G: Applications, challenges, and opportunities,

    Z. Yang, M. Chen, K.-K. Wong, H. V . Poor, and S. Cui, “Federated learning for 6G: Applications, challenges, and opportunities,” Engi- neering, vol. 8, pp. 33–41, Jan. 2022

  106. [114]

    Adaptive federated learning in resource constrained edge computing systems,

    S. Wang, T. Tuor, T. Salonidis, K. K. Leung, C. Makaya, T. He et al., “Adaptive federated learning in resource constrained edge computing systems,” IEEE J. Sel. Areas Commun. , vol. 37, no. 6, pp. 1205–1221, Jun. 2019

  107. [115]

    A lite distributed semantic communication system for internet of things,

    H. Xie and Z. Qin, “A lite distributed semantic communication system for internet of things,” IEEE J. Sel. Areas Commun. , vol. 39, no. 1, pp. 142–153, Jan. 2021

  108. [116]

    Joint privacy enhancement and quantization in federated learning,

    N. Lang, E. Sofer, T. Shaked, and N. Shlezinger, “Joint privacy enhancement and quantization in federated learning,” IEEE Trans. Signal Process., vol. 71, pp. 295–310, Feb. 2023

  109. [117]

    Joint gradient sparsification and device scheduling for federated learning,

    X. Lin, Y . Liu, F. Chen, X. Ge, and Y . Huang, “Joint gradient sparsification and device scheduling for federated learning,” IEEE Trans. Green Commun. Netw., vol. 7, no. 3, pp. 1407–1419, Sep. 2023

  110. [118]

    Deep joint source- channel coding for wireless image transmission,

    E. Bourtsoulatze, D. Burth Kurka, and D. G ¨und¨uz, “Deep joint source- channel coding for wireless image transmission,” IEEE Trans. on Cogn. Commun. Netw., vol. 5, no. 3, pp. 567–579, Sep. 2019

  111. [119]

    Large language models empowered autonomous edge AI for connected intelligence,

    Y . Shen, J. Shao, X. Zhang, Z. Lin, H. Pan, D. Liet al., “Large language models empowered autonomous edge AI for connected intelligence,” IEEE Commun. Mag. , vol. 62, no. 10, pp. 140–146, Oct. 2024

  112. [120]

    Large ai model empowered multimodal semantic communications,

    F. Jiang, L. Dong, Y . Peng, K. Wang, K. Yang, C. Pan et al., “Large ai model empowered multimodal semantic communications,” IEEE Commun. Mag., pp. 1–7, Sep. 2024

  113. [121]

    Large generative AI models for telecom: The next big thing?

    L. Bariah, Q. Zhao, H. Zou, Y . Tian, F. Bader, and M. Debbah, “Large generative AI models for telecom: The next big thing?” IEEE Commun. Mag., vol. 62, no. 11, pp. 84–90, Nov. 2024

  114. [123]

    Unleashing the power of edge-cloud generative AI in mobile networks: A survey of AIGC services,

    M. Xu, H. Du, D. Niyato, J. Kang, Z. Xiong, S. Mao et al., “Unleashing the power of edge-cloud generative AI in mobile networks: A survey of AIGC services,” IEEE Commun. Surveys Tuts. , vol. 26, no. 2, pp. 1127–1170, 2nd Quart. 2024

  115. [124]

    Semantic communications: Overview, open issues, and future research directions,

    X. Luo, H.-H. Chen, and Q. Guo, “Semantic communications: Overview, open issues, and future research directions,” IEEE Wireless Commun., vol. 29, no. 1, pp. 210–219, Feb. 2022

  116. [125]

    WirelessLLM: Empowering large language models towards wireless intelligence,

    J. Shao, J. Tong, Q. Wu, W. Guo, Z. Li, Z. Lin et al. , “WirelessLLM: Empowering large language models towards wireless intelligence,” arXiv preprint arXiv:2405.17053 , 2024. [Online]. Available: https://arxiv.org/abs/2405.17053

  117. [126]

    Generative semantic communications with foundation models: Perception-error analysis and semantic-aware power allocation,

    C. Xu, M. B. Mashhadi, Y . Ma, R. Tafazolli, and J. Wang, “Generative semantic communications with foundation models: Perception-error analysis and semantic-aware power allocation,” IEEE J. Sel. Areas Commun., pp. 1–1, early access, 2025

  118. [127]

    Llama: Open and efficient foundation language models,

    H. Touvron, T. Lavril, G. Izacard, X. Martinet, M.-A. Lachaux, T. Lacroix et al. , “Llama: Open and efficient foundation language models,” arXiv preprint arXiv:2302.13971 , 2023

  119. [128]

    6G for vehicle-to-everything (V2X) communications: Enabling technologies, challenges, and opportunities,

    M. Noor-A-Rahim, Z. Liu, H. Lee, M. O. Khyam, J. He, D. Pesch et al., “6G for vehicle-to-everything (V2X) communications: Enabling technologies, challenges, and opportunities,” Proc. IEEE , vol. 110, no. 6, pp. 712–734, Jun. 2022

  120. [129]

    Integrating pre-trained language model with physical layer communications,

    J.-H. Lee, D.-H. Lee, J. Lee, and J. Pujara, “Integrating pre-trained language model with physical layer communications,” IEEE Trans. Wireless Commun., vol. 23, no. 11, pp. 17 266–17 278, 2024

  121. [130]

    Large language model enhanced multi-agent systems for 6g communications,

    F. Jiang, Y . Peng, L. Dong, K. Wang, K. Yang, C. Pan et al., “Large language model enhanced multi-agent systems for 6g communications,” IEEE Wireless Commun., vol. 31, no. 6, pp. 48–55, 2024

  122. [131]

    LLM4CP: Adapting large language models for channel prediction,

    B. Liu, X. Liu, S. Gao, X. Cheng, and L. Yang, “LLM4CP: Adapting large language models for channel prediction,” J. Commun. Inf. Netw., vol. 9, no. 2, pp. 113–125, Jun. 2024

  123. [132]

    Recent progress and development of hyperdimensional computing (HDC) for edge intelligence,

    C.-Y . Chang, Y .-C. Chuang, C.-T. Huang, and A.-Y . Wu, “Recent progress and development of hyperdimensional computing (HDC) for edge intelligence,” IEEE J. Emerg. Sel. Topics Circuits Syst. , p. 119–136, Mar 2023

  124. [133]

    Hypervector design for efficient hyperdimensional computing on edge devices,

    T. Basaklar, Y . Tuncel, S. Y . Narayana, S. Gumussoy, and U. Y . Ogras, “Hypervector design for efficient hyperdimensional computing on edge devices,” in Proc. Int. Res. Symp. Tiny Mach. Learn. , Burlingame, CA, Mar. 2021, pp. 1–9

  125. [134]

    Hyperdimensional computing: An introduction to com- puting in distributed representation with high-dimensional random vectors,

    P. Kanerva, “Hyperdimensional computing: An introduction to com- puting in distributed representation with high-dimensional random vectors,” Cogn. Comput., vol. 1, no. 2, pp. 139–159, Jan. 2009

  126. [135]

    A framework for collaborative learning in secure high-dimensional space,

    M. Imani, Y . Kim, S. Riazi, J. Messerly, P. Liu, F. Koushanfar et al., “A framework for collaborative learning in secure high-dimensional space,” in Proc. IEEE Int. Conf. Cloud Comput. , Milan, Italy, 2019, pp. 435–446

  127. [136]

    Collective communication for dense sensing environments,

    P. Jakimovski, H. R. Schmidtke, S. Sigg, L. W. F. Chaves, and M. Beigl, “Collective communication for dense sensing environments,” J. Ambient Intell. Smart Environ., vol. 4, no. 2, pp. 123–134, May 2012. 27

  128. [137]

    Dependable MAC layer architecture based on holographic data representation using hyper- dimensional binary spatter codes,

    D. Kleyko, N. Lyamin, E. Osipov, and L. Riliskis, “Dependable MAC layer architecture based on holographic data representation using hyper- dimensional binary spatter codes,” in Proc. Multiple Access Commun. , Berlin, Heidelberg, 2012, pp. 134–145

  129. [138]

    HDM: Hyper-dimensional modulation for robust low- power communications,

    H.-S. Kim, “HDM: Hyper-dimensional modulation for robust low- power communications,” in Proc. IEEE Int. Conf. Commun. , Kansas City, MO, USA, May 2018, pp. 1–6

  130. [139]

    Near- channel classifier: Symbiotic communication and classification in high- dimensional space,

    M. Hersche, S. Lippuner, M. Korb, L. Benini, and A. Rahimi, “Near- channel classifier: Symbiotic communication and classification in high- dimensional space,” Brain Inform., vol. 8, no. 1, p. 16, Aug. 2021

  131. [140]

    Hyper-dimensional modulation for robust short packets in massive machine-type communications,

    C.-W. Hsu and H.-S. Kim, “Hyper-dimensional modulation for robust short packets in massive machine-type communications,” IEEE Trans. Commun., vol. 71, no. 3, pp. 1388–1402, Mar. 2023

  132. [141]

    MulTa- HDC: A multi-task learning framework for hyperdimensional comput- ing,

    C.-Y . Chang, Y .-C. Chuang, E.-J. Chang, and A.-Y . A. Wu, “MulTa- HDC: A multi-task learning framework for hyperdimensional comput- ing,” IEEE Trans. Comput., vol. 70, no. 8, pp. 1269–1284, Aug. 2021

  133. [142]

    FL-HDC: Hyperdimen- sional computing design for the application of federated learning,

    C.-Y . Hsieh, Y .-C. Chuang, and A.-Y . A. Wu, “FL-HDC: Hyperdimen- sional computing design for the application of federated learning,” in Proc. IEEE Int. Conf. Artif. Intell. Circuits Syst., Washington DC, USA, Jun. 2021, pp. 1–5

  134. [143]

    FedHD: federated learning with hyperdimensional computing,

    Q. Zhao, K. Lee, J. Liu, M. Huzaifa, X. Yu, and T. Rosing, “FedHD: federated learning with hyperdimensional computing,” in Proc. Annual Int. Conf. Mobile Comput. Netw., New York, USA, 2022, pp. 791–793

  135. [144]

    HDFL: Private and robust federated learning using hyperdimensional computing,

    H. Kasyap, S. Tripathy, and M. Conti, “HDFL: Private and robust federated learning using hyperdimensional computing,” in Proc. IEEE Int. Conf. Trust Secur. Privacy Comput. Commun. , Exeter, United Kingdom, 2023, pp. 214–221

  136. [145]

    User association and resource allocation in large language model based mobile edge computing system over 6G wireless communications,

    L. Qian and J. Zhao, “User association and resource allocation in large language model based mobile edge computing system over 6G wireless communications,” Singapore, Singapore, 2024, pp. 1–7

  137. [2023]

    Available: https://arxiv.org/abs/2303.18223

    [Online]. Available: https://arxiv.org/abs/2303.18223

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.