Pith. sign in

REVIEW 4 major objections 4 minor 6 cited by

ToDMA: Large Model-Driven Massive Token Communications for Semantic Multiple Access

T0 review · 4 major / 4 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 ToDMA, a token-domain multiple access scheme, lets many uncoordinated devices share uplink resources by transmitting token indices, while pretrained transformers at the receiver repair collision-lost tokens, cutting latency fourfold.

desk verdict Worth reading, but the headline 4x latency gain does not follow from the paper's own formula, and the core scheme is already in the authors' workshop paper. read the letter →

arxiv 2505.10946 v3 pith:NB6DR3VC submitted 2025-05-16 cs.IT cs.AIcs.LGeess.SPmath.IT

Li Qiao , Mahdi Boloursaz Mashhadi , Zhen Gao , Robert Schober , Deniz Gündüz This is my paper · ORCID
classification cs.ITcs.AIcs.LGeess.SPmath.IT
keywords semanticcommunicationstokenunsourcedrandomaccessmultipleapproximatemessagepassingmaskedpredictioncompressedsensingmassiveMIMO
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

ToDMA asks whether token-based processing can turn the uplink of a massive access network into a single shared token channel. The paper claims yes: when many uncoordinated devices transmit tokenized image or text sources over the same time-frequency resources, a receiver can first detect the active tokens by compressed sensing, then use the consistency of each device's channel across time slots to assign tokens to devices, and finally let a pretrained bidirectional transformer fill in tokens lost to collisions. The payoff, if true, is a grant-free non-orthogonal multiple access scheme whose latency is four times lower than an orthogonal token-communication baseline at comparable reconstruction quality. This matters because massive machine-type traffic in future networks needs low-latency access with minimal coordination, and token-domain redundancy offers a new resource to exploit.

What carries the argument

The load-bearing mechanism is the shared token-modulation codebook $\mathbf{U}\in\mathbb{C}^{L\times Q}$, which turns each of the $Q$ token indices into a fixed length-$L$ codeword so that the received signal is a sparse superposition; the receiver's three-stage pipeline then converts that sparsity into reconstructed sequences: approximate message passing for active-token and CSI estimation, K-means++ clustering of per-token CSI across $N$ time slots to assign tokens to the $K$ devices, and candidate-restricted masked-token prediction by a pretrained bidirectional transformer to fill collision-induced gaps. The paper also names an enabling premise, 'semantic orthogonality': the contextual redundancy of natural text and images must be strong enough that a pretrained model can distinguish and recover a device's token sequence from the mixed and partially missing token stream.

What would settle it

Feed ToDMA a source whose tokens are independent and uniformly distributed over the codebook, for example random strings or randomly shuffled image tokens, and run the default settings ($K=40$, $M=256$, $L=K+1$) at SNR $=25$ dB; if TER and PSNR/LPIPS match the context-unaware Non-Orth Com baseline, the masked-token prediction is doing no work and the semantic-orthogonality premise is falsified.

Watch

Extended reading notes

Core claim

The central discovery is that token-domain semantic orthogonality can be turned into a multiple access dimension. Each token index from a shared codebook is mapped to a shared modulation codeword, so the superposition over the wireless channel is a sparse linear mixture of codewords. The receiver runs an AMP-based estimator to recover the active token set and per-token CSI in each time slot, clusters the CSI to assign tokens to devices, and uses pretrained BERT or MaskGIT to predict masked positions caused by token collisions. Simulations on ImageNet-100 and QUOTES500K show that ToDMA keeps token error rates and perceptual quality close to an error-free orthogonal baseline while cutting latency by a factor of four, and outperforms a context-unaware non-orthogonal baseline that randomly guesses collided tokens.

Load-bearing premise

The recovery pipeline assumes the transmitted token sequences are predictable from their context, so that a pretrained transformer can reliably fill in tokens lost to collisions; if the source tokens carry little contextual redundancy, the whole advantage over a context-unaware baseline disappears.

Editorial extensions

If this is right

  • If ToDMA's claims hold, uplink access for massive IoT can be grant-free and non-orthogonal: devices transmit when they have data, and the receiver separates them using token structure rather than per-device preambles.
  • The fourfold latency reduction over Orth-Com comes with comparable or better distortion and perceptual quality in the tested image and text tasks, so orthogonal token transmission may be unnecessary when sources are contextually redundant.
  • Receiver complexity scales linearly with the number of active devices and receive antennas, decoupling the large tokenizer dimension $Q$ from the number of devices, which is desirable for massive MIMO.
  • As the number of receive antennas $M$ grows, token detection error approaches zero without increasing the codeword length $L$, meaning detection accuracy can be bought with more antennas rather than more communication overhead.
  • ToDMA functions as a joint source-channel code in which the source's contextual redundancy is used to repair channel collisions, so stronger contextual models should further reduce token error.

Reading between the lines

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

  • Inference: the decisive test is the strength of contextual redundancy. On low-entropy or adversarial sources such as random identifiers or shuffled sensor logs, the masked-token predictor will have nothing to condition on, so ToDMA should collapse to the context-unaware baseline; the paper does not quantify how much redundancy is needed.
  • Inference: ToDMA is effectively an unsourced random access code over token sequences, so its throughput and latency could be compared against information-theoretic bounds for unsourced multiple access; the paper stops short of that comparison.
  • Inference: the slow-fading assumption, channel vectors constant across all $N$ token slots, is likely the practical bottleneck; under mobility the CSI clustering step would degrade, so a Doppler-robust variant would be a natural follow-up.
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 / 4 minor

Summary. The paper introduces ToDMA, a token-domain multiple access scheme for massive uncoordinated uplink transmissions. Each active device tokenizes its source signal (text or image) and modulates every token with a codeword from a common Gaussian modulation codebook. At the receiver, an AMP-based algorithm detects active tokens and estimates their associated CSI per time slot; a clustering step assigns tokens to devices; residual token collisions leave masked positions, which are filled by candidate-restricted masked-token prediction using pretrained bidirectional transformers (BERT for text, MaskGIT for images). Simulations on ImageNet-100 and QUOTES500K show that ToDMA reduces token error rate and improves PSNR/LPIPS/BERTScore relative to a context-unaware non-orthogonal baseline, and the paper claims a 4-times latency reduction relative to an orthogonal adaptive-QAM baseline. The receiver complexity is reported to be linear in K and M except for the transformer-based prediction stage.

Significance. The core idea—using contextual redundancy in token sequences as a collision-resolution mechanism for unsourced multiple access—is novel and timely, and the empirical validation with standard pretrained models is a strength. If the latency and quality claims survive scrutiny, ToDMA would be a practical interface between token-based source coding and unsourced random access. The paper also provides a useful AMP-based detection formulation and a clear complexity breakdown. However, the headline latency gain does not follow from the paper's own rate formula, and two structural assumptions (known number of active devices and perfect token detection) are not fully supported; these issues need to be resolved before the paper can be recommended for publication.

major comments (4)
  1. [Section VIII-D, Eq. (35), Fig. 10] The claimed 4-times latency reduction over Orth-Com is not supported by the paper's own equations. Inserting SNR = 25 dB (linear SNR about 316), BER = 1e-3, K = 20, N = 256, Q = 1024, and B = 10 MHz into Eq. (35) gives ROrth ≈ 10 MHz × log2(1 + 1.5·316/5.30) ≈ 65 Mb/s, so the Orth-Com latency is 20·256·10/(65e6) ≈ 0.79 ms, whereas the ToDMA latency is 21·256/(10e6) ≈ 0.54 ms; the ratio is about 1.5, not 4. Even at BER = 1e-9 the ratio is about 2.0. Since the 4-times statement appears as a contribution in Section III and underlies Fig. 10, the latency comparison must be recomputed and the claimed gain revised to a value consistent with Eq. (35), or the comparison setup must be re-specified (for example, by including Orth-Com's signaling overhead) and defended.
  2. [Section VI-A, problem (25)] The token-assignment stage requires the number of active devices K as an input to K-means++ clustering. In the grant-free unsourced scenario of Section IV-B, K is not known at the receiver; the paper neither provides an activity-count estimator nor analyzes the sensitivity of clustering and subsequent token assignment to K mismatch. Since all later masked-token recovery depends on this clustering, the paper should either introduce an estimation procedure for K or clearly state a separate assumption and evaluate the impact of K mismatch on the reported performance.
  3. [Section VI-B.2, after Eq. (28)] The fine-grained assignment is developed under the assumption that token detection is perfect (bPn = Pn), justified by the claim that token detection error approaches zero as the number of antennas M grows. However, the main simulations use M = 256, where Fig. 5(a) shows nonzero TDER; the effect of detection errors on the candidate token sets and on the final TER/PSNR is not quantified. Please either analyze the imperfect-detection case or restrict the corresponding claims to the perfect-detection regime.
  4. [Section I-C and Section VI-C] The notion of 'token-domain semantic orthogonality' is introduced only by example and is never defined formally. The entire advantage of ToDMA over the context-unaware baseline rests on the ability of BERT/MaskGIT to resolve masked positions from context, so the paper should state a quantitative condition (for example, in terms of the conditional entropy of tokens given their context, or a minimum prediction accuracy threshold) under which the proposed recovery is effective, and it should discuss regimes where the contextual redundancy is insufficient.
minor comments (4)
  1. [Section V-B, Eq. (13)] In the update for R^t_{q,m}, the codebook coefficient is written as u^*_{l,m}; it should be u^*_{l,q}.
  2. [Algorithm 1, line 4] The initialization line contains a stray semicolon and comma ('bBk = 0Q×N , ;'); please clean up the pseudocode.
  3. [Eq. (35)] The formula for ROrth is typeset ambiguously: the placement of SNR in the denominator makes it appear that spectral efficiency decreases with SNR at fixed BER; please rewrite it as ROrth = B log2(1 + 1.5·SNR / (−ln(5·BER))).
  4. [Section VIII-D] The statement 'we assume K = KT' contradicts the massive-access premise that only a small fraction of devices are active (K ≪ KT); if all devices are active, the sparsity argument in Eq. (2) and the 'massive' characterization need to be revisited.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: ToDMA's derivation is self-contained and independently benchmarked; cited self-works are not load-bearing.

full rationale

The claimed derivation chain runs from a shared random modulation codebook (Section IV) through AMP-based active-token detection (Section V), CSI-clustering token assignment (Sections VI-A and VI-B), and pretrained-transformer masked-token prediction (Section VI-C). No system parameter is fitted to the reported TER, PSNR, LPIPS, or BERTScore results and then re-labeled as a prediction: the thresholds and EM-updated sparsity ratios are computed from the signal model, while BERT, MaskGIT, and the VQ-GAN tokenizer are external pretrained components with fixed weights. The two self-citations ([1], [15]) merely identify prior TokenCom/workshop context and do not supply a uniqueness theorem, a forbidden alternative, or a constraint that forces the receiver design; the multiple-access construction is described and simulated within the paper. The 'semantic orthogonality' notion is a conceptual label for contextual redundancy rather than a derived quantity, and the comparisons against context-unaware Non-Orth Com and Orth-Com provide an external check. The possible numerical inconsistency in the 4x latency claim (Section VIII-D, Eq. (35)) would be a correctness/consistency concern, not a circularity, because that latency comparison is a benchmark calculation from the authors' own stated formulas and does not rename a fitted parameter as a prediction.

Assumptions & free parameters 2 free parameters · 5 assumptions · 1 invented entities

The scheme rests on standard CS results plus domain assumptions about channel coherence, known K, and source predictability. No new physical entities are introduced; the conceptual notion of token-domain semantic orthogonality is asserted and only indirectly validated by simulation.

free parameters (2)
  • T_r^h = not specified
    Detection threshold in eq. (24) that converts estimated sparsity ratios into the active token set. Its value is never given in Section VIII, so the simulation results are not exactly reproducible and performance may depend on it.
  • T_s^h (score threshold) = 1/(2 * average cluster distance)
    Confidence threshold for identifying incorrectly assigned tokens in Section VI-B. The formula ties it to clustering output but the factor 1/2 is heuristic and not justified.
assumptions (5)
  • standard math In the large system limit, AMP decouples the matrix estimation problem into scalar denoising problems (Proposition 1).
    Used to derive the posterior mean updates (11)-(21); cited to Donoho-Maleki-Montanari and Rangan [60,61].
  • domain assumption Each device's channel vector h_k is time-invariant over N time slots and unique across devices.
    Needed so that CSI clustering can group tokens by device; introduced in Section IV-B and used in Section VI-A.
  • domain assumption The receiver knows the number of active devices K.
    K-means++ clustering in Section VI-A takes K as input; no estimator is provided.
  • domain assumption Source token sequences are sufficiently context-predictable by the pretrained bidirectional transformer that masked tokens can be recovered from surrounding tokens.
    The entire collision-mitigation stage (Section VI-C) rests on this; evaluated only for ImageNet-100 and QUOTES500K.
  • standard math The shared modulation codebook U has i.i.d. complex Gaussian entries, serving as a random measurement matrix.
    Used to justify CS/AMP applicability; random Gaussian matrices are standard in compressed sensing.
invented entities (1)
  • token-domain semantic orthogonality
    purpose: Conceptual basis for using MLLM context to separate mixed token streams and fill collided tokens.
    The paper asserts that sources like natural text/images are 'semantically orthogonal' enough for pretrained models to recover missing tokens. It is only indirectly validated through simulations; no independent measurement or falsifiable prediction outside the paper is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ToDMA: Large Model-Driven Massive Token Communications for Semantic Multiple Access." pith.science (2026). https://pith.science/paper/NB6DR3VC

@misc{pith2026250510946,
  author       = {Pith},
  title        = {Pith review of: ToDMA: Large Model-Driven Massive Token Communications for Semantic Multiple Access},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NB6DR3VC}},
  note         = {Machine review of arXiv:2505.10946}
}
read the original abstract

Token communications (TokenCom) is an emerging generative semantic communication paradigm, where tokens serve as compact representation units across modalities. Their contextual dependencies can be exploited by pretrained large models for semantic recovery. In this paper, we propose token-domain multiple access (ToDMA), a large-model-driven semantic multiple access scheme for massive token communications. ToDMA integrates unsourced random access with context-aware token processing. It enables massive uncoordinated devices to transmit tokenized source representations over common uplink resources. Specifically, each token index is associated with a shared modulation codeword, exposing token-level structure to the receiver for context-aware recovery. At the receiver, compressed sensing is first employed to jointly detect active tokens and estimate their corresponding channel state information (CSI) from the superposed signals. The source token sequences are then reconstructed by exploiting the consistency of token-associated CSI across multiple token positions. In the presence of token collisions, some active tokens may remain unassigned, leading to missing entries in the reconstructed token sequences. To recover these tokens, candidate-restricted masked-token prediction is performed using pretrained contextual models, thereby leveraging token-level context to mitigate collision effects. Simulation results on both image and text transmission tasks demonstrate that ToDMA reduces access latency while maintaining favorable token recovery and semantic reconstruction quality, showing its scalability for semantic multiple access.

Figures

Figures reproduced from arXiv: 2505.10946 by the authors.

Figure 1
Figure 1. A conceptual illustration of Semantic Orthogonality: Due to the semantic orthogonality of the two text sources, LLMs can reconstruct them from a mixed signal. codebook, based on the bidirectional context information. Masked image modeling has also been adopted in vision pre￾training, e.g., MaskGIT [13], yielding impressive results for various vision tasks. The next-token prediction approach is another effective pre-… view at source ↗
Figure 2
Figure 2. Proposed token-domain multiple access (ToDMA) framework. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Illustration of token assignment and masked token predictions. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Token detection performance versus length of modulation codeword [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Token detection performance versus the number of antennas [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Performance comparison of the proposed ToDMA-based image transmission with benchmark schemes, validated on the ImageNet-100 dataset, SNR [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Visual quality illustration for K = 40. Due to space limitations, only 11 randomly chosen images out of the 40 devices/images are shown. From top to bottom: Original images; reconstructed images using “Orth-Com”; reconstructed images using “ToDMA”; reconstructed images…
Figure 8
Figure 8. Figure 8: Visual quality illustration for K = {20, 60, 80}. From top to bottom: Reconstructed images using “ToDMA”; reconstructed images using “Non-Orth Com”; token error visualization for “ToDMA”; and token error visualization for “Non-Orth Com”. Green blocks represent correct …
Figure 9
Figure 9. Figure 9: Comparison of TER and BERTScore performance at [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 6
Figure 6. Figure 6: In this setting, we assume K = KT. As can be observed, the latency of the proposed ToDMA is 4 times lower than that of the “Orth-Com” scheme, even without accounting for the signaling overhead of Orth-Com, while maintaining comparable performance. In contrast, the “Non…
Figure 10
Figure 10. Figure 10: Latency compared to adaptive QAM for various target BERs. [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 6 Pith papers

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

  1. Loss-Resilient Wireless Video Token Communication over Block Fading Channels

    cs.LG 2026-08 conditional novelty 6.0 of 10

    Shuffling video tokens by importance and scheduling them over fading channel blocks, then reconstructing from surviving tokens, improves perceptual video quality under packet loss compared with DCVC, VVC, and H.265.

  2. Test-Time Scalable AI-RAN: Inference Time Allocation for Cell-Free MIMO

    eess.SP 2026-08 conditional novelty 5.0 of 10

    A new framework allocates inference time between test-time-scalable precoding and quantization AI modules in cell-free MIMO, with the optimal split depending on the temporal correlation of the channel.

  3. Geometric Cross-Modal Token Selection for Latency-Constrained Multimodal Token Communication

    eess.SP 2026-08 conditional novelty 5.0 of 10

    Selecting tokens that lie inside multiple anchor-centric semantic grain regions improves multimodal VQA/AVQA accuracy under latency and erasure constraints compared with pairwise attention-based selection.

  4. Wireless TokenCom: RL-Based Tokenizer Agreement for Multi-User Wireless Token Communications

    cs.LG 2026-02 conditional novelty 5.0 of 10

    Joint tokenizer/codebook selection, subchannel assignment, and beamforming for multi-user video TokenCom is posed as an MDP and solved by DQN for discrete choices and DDPG for beamforming, with simulated gains over H.265.

  5. Text-Guided Token Communication for Wireless Image Transmission

    cs.IT 2025-07 reject novelty 5.0 of 10

    A text-guided token transmission system using pre-trained image and text models outperforms a deep JSCC baseline on perceptual and semantic metrics, but relies on an assumption that text is available at the receiver.

  6. Token Communication in the Era of Large Models: An Information Bottleneck-Based Approach

    eess.SP 2025-07 reject novelty 4.0 of 10

    A unified token-based wireless communication framework combines an information-bottleneck-style tokenizer with a causal multimodal language model for joint understanding and generation.

Reference graph

Works this paper leans on

70 extracted references · 49 canonical work pages · cited by 6 Pith papers

  1. [1]

    Token-domain mul- tiple access: Exploiting semantic orthogonality for collision mitigation,

    L. Qiao, M. B. Mashhadi, Z. Gao, and D. G ¨und¨uz, “Token-domain mul- tiple access: Exploiting semantic orthogonality for collision mitigation,” arXiv preprint arXiv:2502.06118 , Feb. 2025

    arXiv 2025

  2. [2]

    MM-LLMs: Recent advances in multimodal large language models,

    D. Zhang et al. , “MM-LLMs: Recent advances in multimodal large language models,” in Findings of the Association for Computational Linguistics ACL 2024 , pp. 12401–12430, 2024

    work page 2024

  3. [3]

    Hello GPT-4o,

    OpenAI, “Hello GPT-4o,” 2024. Accessed: 2024-09-29

    2024

  4. [4]

    Blip-2: Bootstrapping language- image pre-training with frozen image encoders and large language models,

    J. Li, D. Li, S. Savarese, and S. Hoi, “Blip-2: Bootstrapping language- image pre-training with frozen image encoders and large language models,” in Proc. Int. Conf. Mach. Learn. (ICML) , pp. 19730–19742, PMLR, 2023

    work page 2023

  5. [5]

    Visual instruction tuning,

    H. Liu, C. Li, Q. Wu, and Y . J. Lee, “Visual instruction tuning,” Proc. Adv. Neural Inf. Process. Syst. (NeurIPS) , vol. 36, 2024

    work page 2024

  6. [6]

    LAMBO: Large AI model empowered edge intelli- gence,

    L. Dong et al. , “LAMBO: Large AI model empowered edge intelli- gence,” IEEE Commun. Mag. , vol. 63, no. 4, pp. 88–94, 2025

    work page 2025

  7. [7]

    A comprehensive survey of large language models and multimodal large language models in medicine,

    H. Xiao et al., “A comprehensive survey of large language models and multimodal large language models in medicine,” Information Fusion , p. 102888, 2024

    work page 2024

  8. [8]

    Generative AI-enhanced multi-modal semantic communi- cation in Internet of Vehicles: System design and methodologies,

    J. Lu et al., “Generative AI-enhanced multi-modal semantic communi- cation in Internet of Vehicles: System design and methodologies,” arXiv preprint arXiv:2409.15642, 2024

    arXiv 2024

Show all 70 references
  1. [9]

    Attention is all you need,

    A. Vaswani, “Attention is all you need,” Proc. Adv. Neural Inf. Process. Syst. (NeurIPS), 2017

    2017

  2. [10]

    BERT: Pre-training of deep bidirectional transformers for language understanding,

    J. Devlin, “BERT: Pre-training of deep bidirectional transformers for language understanding,” arXiv preprint arXiv:1810.04805 , 2018

    2018 arXiv

  3. [11]

    Neural discrete representation learning,

    A. Van den Oord, O. Vinyals, and K. Kavukcuoglu, “Neural discrete representation learning,” Proc. Adv. Neural Inf. Process. Syst. (NeurIPS), vol. 30, 2017

    2017

  4. [12]

    What makes for good tokenizers in vision transformer?,

    S. Qian, Y . Zhu, W. Li, M. Li, and J. Jia, “What makes for good tokenizers in vision transformer?,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 45, no. 11, pp. 13011–13023, 2022

    2022

  5. [13]

    MaskGIT: Masked generative image transformer,

    H. Chang et al., “MaskGIT: Masked generative image transformer,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR) , pp. 11315– 11325, 2022. 13

    2022

  6. [14]

    Taming transformers for high- resolution image synthesis,

    P. Esser, R. Rombach, and B. Ommer, “Taming transformers for high- resolution image synthesis,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), pp. 12873–12883, 2021

    2021

  7. [15]

    Token communications: A unified framework for cross-modal context-aware semantic communications,

    L. Qiao, M. B. Mashhadi, Z. Gao, R. Tafazolli, M. Bennis, and D. Niyato, “Token communications: A unified framework for cross-modal context-aware semantic communications,” arXiv preprint arXiv:2502.12096, Feb. 2025

    2025 arXiv

  8. [16]

    Goal-oriented and semantic communication in 6G AI-native networks: The 6G-GOALS approach,

    E. C. Strinati et al. , “Goal-oriented and semantic communication in 6G AI-native networks: The 6G-GOALS approach,” in Proc. 2024 Joint Eur. Conf. Netw. Commun. & 6G Summit (EuCNC/6G Summit), pp. 1–6, IEEE, 2024

    2024

  9. [17]

    Beyond transmitting bits: Context, semantics, and task-oriented communications,

    D. G ¨und¨uz et al. , “Beyond transmitting bits: Context, semantics, and task-oriented communications,” IEEE J. Select. Areas Commun., vol. 41, no. 1, pp. 5–41, 2023

    2023

  10. [18]

    Joint source–channel coding: Fundamentals and recent progress in practical designs,

    D. G ¨und¨uz, M. A. Wigger, T.-Y . Tung, P. Zhang, and Y . Xiao, “Joint source–channel coding: Fundamentals and recent progress in practical designs,” Proc. IEEE, 2024

    2024

  11. [19]

    Deep learning enabled semantic communication systems,

    H. Xie, Z. Qin, G. Y . Li, and B.-H. Juang, “Deep learning enabled semantic communication systems,” IEEE Trans. Signal Processing , vol. 69, pp. 2663–2675, 2021

    2021

  12. [20]

    Deep joint source- channel coding for wireless image transmission,

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

    2019

  13. [21]

    Large AI model empowered multimodal semantic communications,

    F. Jiang et al. , “Large AI model empowered multimodal semantic communications,” IEEE Commun. Mag., vol. 63, no. 1, pp. 76–82, 2025

    2025

  14. [22]

    Generative AI-driven semantic communication networks: Architecture, technologies and applications,

    C. Liang, H. Du, Y . Sun, D. Niyato, J. Kang, D. Zhao, and M. A. Imran, “Generative AI-driven semantic communication networks: Architecture, technologies and applications,” IEEE Trans. Cogn. Commun. Netw. , vol. 11, no. 1, pp. 27–47, 2025

    2025

  15. [23]

    Large AI model-based semantic communications,

    F. Jiang et al., “Large AI model-based semantic communications,” IEEE Wireless Commun., vol. 31, no. 3, pp. 68–75, 2024

    2024

  16. [24]

    Deep generative modeling reshapes compression and transmission: From efficiency to resiliency,

    J. Dai, X. Qin, S. Wang, L. Xu, K. Niu, and P. Zhang, “Deep generative modeling reshapes compression and transmission: From efficiency to resiliency,” IEEE Wireless Commun., vol. 31, no. 4, pp. 48–56, 2024

    2024

  17. [25]

    Information compression in the AI era: Recent advances and future challenges,

    J. Chen, Y . Fang, A. Khisti, A. ¨Ozg¨ur, and N. Shlezinger, “Information compression in the AI era: Recent advances and future challenges,”IEEE J. Select. Areas Commun. , pp. 1–1, 2025

    2025

  18. [26]

    LLMZip: Lossless text compression using large language models,

    C. S. K. Valmeekam, K. Narayanan, D. Kalathil, et al. , “LLMZip: Lossless text compression using large language models,” arXiv preprint arXiv:2306.04050, 2023

    2023 arXiv

  19. [27]

    Language modeling is compression,

    G. Del ´etang et al., “Language modeling is compression,” arXiv preprint arXiv:2309.10668, 2023

    2023 arXiv

  20. [28]

    Latency-aware generative semantic communications with pre-trained diffusion models,

    L. Qiao et al., “Latency-aware generative semantic communications with pre-trained diffusion models,” IEEE Wireless Commun. Lett. , vol. 13, no. 10, pp. 2652–2656, 2024

    2024

  21. [29]

    Extreme video compression with prediction using pre-trained diffusion models,

    B. Li, Y . Liu, X. Niu, et al., “Extreme video compression with prediction using pre-trained diffusion models,” in Proc. 2024 16th Int. Conf. Wireless Commun. Signal Process. (WCSP) , pp. 1449–1455, 2024

    2024

  22. [30]

    Language-oriented semantic latent representation for image transmission,

    G. Cicchetti et al., “Language-oriented semantic latent representation for image transmission,” in Proc. 2024 IEEE Int. Workshop Mach. Learn. Signal Process. (MLSP) , pp. 1–6, IEEE, 2024

    2024

  23. [31]

    Toward intelligent communications: Large model empowered semantic communications,

    H. Xie, Z. Qin, X. Tao, and Z. Han, “Toward intelligent communications: Large model empowered semantic communications,” IEEE Commun. Mag., vol. 63, no. 1, pp. 69–75, 2025

    2025

  24. [32]

    Goal-oriented semantic communica- tion for wireless video transmission via generative ai,

    N. Li, Y . Deng, and D. Niyato, “Goal-oriented semantic communica- tion for wireless video transmission via generative ai,” arXiv preprint arXiv:2502.20927, 2025

    2025 arXiv

  25. [33]

    Diffusion-based generative multicasting with intent-aware semantic decomposition,

    X. Liu, M. B. Mashhadi, L. Qiao, Y . Ma, R. Tafazolli, and M. Bennis, “Diffusion-based generative multicasting with intent-aware semantic decomposition,” arXiv preprint arXiv:2411.02334 , 2024

    2024

  26. [34]

    Accelerating mobile edge generation (MEG) by constrained learning,

    X. Xu, Y . Liu, X. Mu, H. Xing, and A. Nallanathan, “Accelerating mobile edge generation (MEG) by constrained learning,” IEEE Trans. Cogn. Commun. Netw., pp. 1–1, 2025

    2025

  27. [35]

    Generative semantic communication via textual prompts: Latency performance tradeoffs,

    M. Ren et al., “Generative semantic communication via textual prompts: Latency performance tradeoffs,” IEEE Trans. Veh. Technol. , pp. 1–6, 2025

    2025

  28. [36]

    High perceptual quality wireless image delivery with denoising diffusion models,

    S. F. Yilmaz et al., “High perceptual quality wireless image delivery with denoising diffusion models,” in Proc. IEEE Conf. Comput. Commun. Workshops (INFOCOM WKSHPS), pp. 1–5, 2024

    2024

  29. [37]

    Generative joint source-channel coding for semantic image transmission,

    E. Erdemir et al., “Generative joint source-channel coding for semantic image transmission,” IEEE J. Select. Areas Commun. , vol. 41, no. 8, pp. 2645–2657, 2023

    2023

  30. [38]

    Semantics-guided diffusion for deep joint source- channel coding in wireless image transmission,

    M. Zhang et al. , “Semantics-guided diffusion for deep joint source- channel coding in wireless image transmission,” arXiv preprint arXiv:2501.01138, 2025

    2025 arXiv

  31. [39]

    SING: Semantic image communications using null-space and INN-guided dif- fusion models,

    J. Chen, S. F. Yilmaz, D. You, P. L. Dragotti, and D. G ¨und¨uz, “SING: Semantic image communications using null-space and INN-guided dif- fusion models,” arXiv preprint arXiv:2503.12484 , 2025

    2025 arXiv

  32. [40]

    Adaptive semantic token selection for AI-native goal-oriented commu- nications,

    A. Devoto, S. Petruzzi, J. Pomponi, P. Di Lorenzo, and S. Scardapane, “Adaptive semantic token selection for AI-native goal-oriented commu- nications,” arXiv preprint arXiv:2405.02330 , 2024

    2024 arXiv

  33. [41]

    Next generation multiple access for IMT towards 2030 and beyond,

    Z. Ding, R. Schober, P. Fan, and H. V . Poor, “Next generation multiple access for IMT towards 2030 and beyond,” Sci. China Inf. Sci. , vol. 67, no. 6, p. 166301, 2024

    2024

  34. [42]

    Multiple access techniques for intelligent and multifunctional 6G: Tutorial, survey, and outlook,

    B. Clerckx et al. , “Multiple access techniques for intelligent and multifunctional 6G: Tutorial, survey, and outlook,”Proc. IEEE, vol. 112, no. 7, pp. 832–879, 2024

    2024

  35. [43]

    Massive connectivity with massive MIMO—Part I: Device activity detection and channel estimation,

    L. Liu and W. Yu, “Massive connectivity with massive MIMO—Part I: Device activity detection and channel estimation,” IEEE Trans. Signal Process., vol. 66, pp. 2933–2946, June 2018

    2018

  36. [44]

    Compressive sensing- based adaptive active user detection and channel estimation: Massive access meets massive MIMO,

    M. Ke, Z. Gao, Y . Wu, X. Gao, and R. Schober, “Compressive sensing- based adaptive active user detection and channel estimation: Massive access meets massive MIMO,” IEEE Trans. Signal Processing , vol. 68, pp. 764–779, 2020

    2020

  37. [45]

    A perspective on massive random-access,

    Y . Polyanskiy, “A perspective on massive random-access,” inProc. IEEE Int. Symp. Inf. Theory (ISIT) , pp. 2523–2527, 2017

    2017

  38. [46]

    Coopera- tive activity detection: Sourced and unsourced massive random access paradigms,

    X. Shao, X. Chen, D. W. K. Ng, C. Zhong, and Z. Zhang, “Coopera- tive activity detection: Sourced and unsourced massive random access paradigms,” IEEE Trans. Signal Process., vol. 68, pp. 6578–6593, 2020

    2020

  39. [47]

    Massive unsourced random access based on uncou- pled compressive sensing: Another blessing of massive MIMO,

    V . Shyianov et al., “Massive unsourced random access based on uncou- pled compressive sensing: Another blessing of massive MIMO,” IEEE J. Select. Areas Commun. , vol. 39, no. 3, pp. 820–834, 2020

    2020

  40. [48]

    Massive unsourced random access: Exploiting angular domain sparsity,

    X. Xie et al. , “Massive unsourced random access: Exploiting angular domain sparsity,” IEEE Trans. Commun., vol. 70, no. 4, pp. 2480–2498, 2022

    2022

  41. [49]

    Unsourced multiple access with common alarm messages: Network slicing for massive and critical IoT,

    K. H. Ngo et al. , “Unsourced multiple access with common alarm messages: Network slicing for massive and critical IoT,” IEEE Trans. Wireless Commun., vol. 72, pp. 907–923, Feb. 2024

    2024

  42. [50]

    Design and analysis of massive uncoupled unsourced ran- dom access with Bayesian joint decoding,

    F. Tian et al., “Design and analysis of massive uncoupled unsourced ran- dom access with Bayesian joint decoding,” IEEE Trans. Veh. Technol., vol. 73, no. 7, pp. 10350–10364, 2024

    2024

  43. [51]

    Unsourced multiple access: A coding paradigm for massive random access,

    G. Liva and Y . Polyanskiy, “Unsourced multiple access: A coding paradigm for massive random access,” Proc. IEEE , vol. 112, no. 9, pp. 1214–1229, 2024

    2024

  44. [52]

    Compressive sensing-based grant-free massive access for 6G massive communication,

    Z. Gao et al., “Compressive sensing-based grant-free massive access for 6G massive communication,” IEEE Internet Things J. , vol. 11, no. 5, pp. 7411–7435, 2024

    2024

  45. [53]

    Massive digital over-the-air computation for communication-efficient federated edge learning,

    L. Qiao et al. , “Massive digital over-the-air computation for communication-efficient federated edge learning,” IEEE J. Select. Areas Commun., vol. 42, no. 11, pp. 3078–3094, 2024

    2024

  46. [54]

    Distributed deep joint source-channel coding over a multiple access channel,

    S. F. Yilmaz, C. Karamanlı, and D. G ¨und¨uz, “Distributed deep joint source-channel coding over a multiple access channel,” in Proc. IEEE Int. Conf. Commun. (ICC) , pp. 1400–1405, 2023

    2023

  47. [55]

    Orthogonal model division multiple access,

    H. Liang et al. , “Orthogonal model division multiple access,” IEEE Trans. Wireless Commun., vol. 23, no. 9, pp. 11693–11707, 2024

    2024

  48. [56]

    DeepMA: End-to-end deep multiple access for wireless image transmission in semantic communication,

    W. Zhang et al., “DeepMA: End-to-end deep multiple access for wireless image transmission in semantic communication,” IEEE Trans. Cogn. Commun. Netw., vol. 10, no. 2, pp. 387–402, 2024

    2024

  49. [57]

    Exploiting semantic communication for non- orthogonal multiple access,

    X. Mu and Y . Liu, “Exploiting semantic communication for non- orthogonal multiple access,” IEEE J. Select. Areas Commun. , vol. 41, no. 8, pp. 2563–2576, 2023

    2023

  50. [58]

    Generative AI empowered semantic feature multiple access (SFMA) over wireless networks,

    J. Wang et al. , “Generative AI empowered semantic feature multiple access (SFMA) over wireless networks,” IEEE Trans. Cogn. Commun. Netw., 2025

    2025

  51. [59]

    Learning to interfere in non-orthogonal multiple-access joint source-channel coding,

    S. F. Yilmaz, C. Karamanli, and D. Gunduz, “Learning to interfere in non-orthogonal multiple-access joint source-channel coding,” arXiv preprint arXiv:2504.03690, 2025

    2025 arXiv

  52. [60]

    Message passing algo- rithms for compressed sensing: I. motivation and construction,

    D. L. Donoho, A. Maleki, and A. Montanari, “Message passing algo- rithms for compressed sensing: I. motivation and construction,” in Proc. IEEE Inf. Theory Workshop (ITW) , pp. 1–5, 2010

    2010

  53. [61]

    Generalized approximate message passing for estimation with random linear mixing,

    S. Rangan, “Generalized approximate message passing for estimation with random linear mixing,” in Proc. IEEE Int. Symp. Inf. Theory (ISIT), pp. 2168–2172, 2011

    2011

  54. [62]

    The expectation-maximization algorithm,

    T. K. Moon, “The expectation-maximization algorithm,” IEEE Signal Process. Mag., vol. 13, no. 6, pp. 47–60, 1996

    1996

  55. [63]

    k-means++: the advantages of careful seeding,

    D. Arthur and S. Vassilvitskii, “k-means++: the advantages of careful seeding,” in Proceedings of the Eighteenth Annual ACM-SIAM Sympo- sium on Discrete Algorithms , SODA ’07, p. 1027–1035, 2007

    2007

  56. [64]

    Variable-rate variable-power MQAM for fading channels,

    A. J. Goldsmith and S.-G. Chua, “Variable-rate variable-power MQAM for fading channels,” IEEE Trans. Commun., vol. 45, no. 10, pp. 1218– 1230, 1997. 14

    1997

  57. [65]

    Rank awareness in joint sparse recovery,

    M. E. Davies and Y . C. Eldar, “Rank awareness in joint sparse recovery,” IEEE Trans. Inform. Theory , vol. 58, no. 2, pp. 1135–1146, 2012

    2012

  58. [66]

    Imagenet large scale visual recognition chal- lenge,

    O. Russakovsky et al. , “Imagenet large scale visual recognition chal- lenge,” Int. J. Comput. Vis. , vol. 115, pp. 211–252, 2015

    2015

  59. [67]

    The unreasonable effectiveness of deep features as a perceptual metric,

    R. Zhang et al. , “The unreasonable effectiveness of deep features as a perceptual metric,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), pp. 586–595, 2018

    2018

  60. [68]

    Scaling laws for neural language models,

    J. Kaplan et al. , “Scaling laws for neural language models,” arXiv preprint arXiv:2001.08361, 2020

    2001 arXiv

  61. [69]

    Proposing contextually relevant quotes for images,

    S. Goel, R. Madhok, and S. Garg, “Proposing contextually relevant quotes for images,” in Advances in Information Retrieval: 40th European Conference on IR Research , pp. 591–597, Springer, 2018

    2018

  62. [70]

    BERTScore: Evaluating text generation with BERT,

    T. Zhang et al. , “BERTScore: Evaluating text generation with BERT,” arXiv preprint arXiv:1904.09675 , 2019

    1904 arXiv

Pith tools

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