Pith. sign in

REVIEW 3 major objections 4 minor 61 references

Rateless Joint Source-Channel Coding, and a Blueprint for 6G Semantic Communications System Design

T0 review · 3 major / 4 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read Rateless joint source-channel coding hands the network the job of adapting the transmission rate, so the application never needs channel state information.

desk verdict A useful architectural idea for JSCC over a binary network interface, but the continuum claim is not backed by the construction. read the letter →

arxiv 2502.06095 v1 pith:PPA7VE5S submitted 2025-02-10 cs.IT cs.AImath.IT

Saeed R. Khosravirad This is my paper · ORCID
classification cs.ITcs.AImath.IT MSC 94A2494A2968T07
keywords ratelessjointsource-channelcodingsemanticcommunicationsrate-adaptivelinkstabilizedbitpipeSVBSCRLACScodeautoencoderJSCC6Gnetworking
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 attempts to resolve the main practical obstacle to joint source-channel coding: the application usually does not know the channel state, while the network does. The proposed fix is a rateless JSCC code that is optimized for a continuum of coding rates, together with a network-side operation that punctures bits out of the codeword to fit the channel capacity and stabilizes the bit-flipping ratio around a target value. From the application's point of view the link becomes a controlled bit pipe, and the network absorbs all rate adaptation. The paper demonstrates the idea with an autoencoder code family called RLACS and reports that, on CIFAR10 image reconstruction, the fine-grained rateless code outperforms a non-rateless benchmark by more than 5 dB in PSNR at low SNR while degrading gracefully as the channel worsens. If successful, the framework offers a way to realize semantic communication services over conventional networks without redesigning the network from scratch.

What carries the argument

The central object is the pair formed by the rateless JSCC code definition and the SVBSC link abstraction, with RLACS as the concrete realization. A rateless code is defined as a tuple $(f,r,g)$ such that for every puncturing length $L$ in $\{0,\dots,\bar L\}$ the expected distortion stays at or below a promised value $d[L]$; the network chooses the modulation order $M$ through Eq. (1) so that the probability the bit-flip ratio exceeds $q_0$ is at most $\varepsilon$. The RLACS machinery is an autoencoder ladder trained sequentially: at iteration $i$, all previous encoder-decoder pairs are frozen and only pair $i$ is optimized against the reconstruction loss over a training channel with $q_0$ and no instability, while decoder $i$ consumes the output of all pairs $1,\dots,i$. The residual attention architecture provides the learned feature hierarchy, and a VIMCO estimator supplies gradients for the binary latent space.

What would settle it

Take the trained RLACS Code 3 and test it over a channel with the same average bit-flip rate $q_0=0.05$ but with errors generated in correlated bursts, such as a Gilbert-Elliott model or a block-fading channel with the same marginal BER; if the PSNR at low SNR drops materially relative to the i.i.d. SVBSC training channel, then the SVBSC abstraction's sufficiency for RLACS is refuted.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that the performance advantage of joint source-channel coding can be obtained while preserving the modular separation of application and network: the application only needs to provide a code that maintains a target distortion for every puncturing length the network might choose, and the network only needs to turn the wireless channel into a stabilized vector binary symmetric channel with bit-flip probability around $q_0$ and stability parameter $\varepsilon$. The formal definition is a $(K,N,\bar L,d)$ rateless code satisfying a distortion bound for every $L$, with null bits inserted at the decoder where the network punctured. The RLACS implementation realizes this with an autoencoder ladder in which earlier trained modules are frozen while a new module is trained for each additional segment of code length; the first bits therefore learn to carry the most reconstruction-critical information. The reported experiments show Code 3, whose segment sizes match the modulation grid, staying near the non-rateless Code 1 at high SNR and beating it by more than 5 dB in PSNR at low SNR, with graceful rather than cliff-like degradation as SNR falls.

Load-bearing premise

The load-bearing premise is that the end-to-end bit channel is well described as a memoryless vector binary symmetric channel whose bit flips are independent with probability around $q_0$; under the stated Rician block-fading channel with uncoded QAM, a deep fade can corrupt many bits in a frame at once, so the training and the graceful-degradation claims rest on an independence assumption that Eq. (1) does not directly guarantee.

Editorial extensions

If this is right

  • Applications can be made channel-agnostic: the network performs rate adaptation by puncturing, so no CSI routing back to the application is needed.
  • Lossy media should escape the cliff and leveling-off effects of packet-ized, error-free networking, because damaged and punctured bits become controlled erasures and errors rather than whole packet drops.
  • Finely matched puncturing granularity matters: Code 3, whose segment sizes align with the modulation constellation grid, clearly beats the coarser Code 2 at low SNR.
  • The code family is extensible: adding modules for higher modulation orders requires no retraining of existing modules, which suits incremental deployment.
  • The link abstraction introduces new QoS metrics, bit-flip ratio, stability $1-\varepsilon$, maximum puncturing $\bar L$, and average puncturing, that replace or supplement block-error-rate targets.

Reading between the lines

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

  • The paper trains RLACS on i.i.d. bit flips; a likely untested limit is that correlated burst errors from deep fades violate the SVBSC independence premise, and decoding quality may drop more than the paper's $\varepsilon$ analysis suggests.
  • The same bit-pipe abstraction should extend to multi-hop links: each hop could puncture up to a pre-agreed limit, with the bottleneck hop setting the end-to-end rate; this is implied by the blueprint but not experimentally tested.
  • Because rateless training assigns higher importance to early bits, the same architecture could be trained with semantic or task-specific losses, not only MSE, making it a candidate for effectiveness-oriented 6G services.
  • A hybrid that uses classic LDPC or Polar codes inside the network to stabilize the BER before RLACS decoding could combine the strengths of both; the paper lists this as future work, so it remains an open testable direction.
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 / 4 minor

Summary. The paper introduces rateless joint source-channel coding (rateless JSCC), a framework in which the application encodes a source into a binary codeword and the network adapts the coding rate by puncturing bits from the end of the codeword, without exchanging channel state information with the application. The network operation is formalized as a stabilized vector binary symmetric channel (SVBSC) that keeps the bit-flipping ratio around a target. The authors then propose a deep autoencoder realization, the RLACS code, built from a ladder of encoder/decoder pairs trained sequentially, and evaluate it on CIFAR10 image transmission over a Rician block-fading channel. They report that the rateless code (Code 3) outperforms a non-rateless DNN benchmark (Code 1) by more than 5 dB in PSNR at low SNR, and that the proposed approach degrades gracefully. The second half of the paper is a blueprint for 6G semantic communication system design, including new QoS metrics and a list of open research problems.

Significance. The central idea of commissioning the network with rate control through puncturing, while the application runs a channel-agnostic JSCC code, is a useful conceptual contribution that could help reconcile JSCC with existing network architectures. If the claimed performance holds, the RLACS code would provide a practical way to achieve graceful degradation without CSI at the application, and the progressive-coding interpretation (Fig. 8) is a nice demonstration. Strengths of the paper include the formal definitions of the rateless code and SVBSC, the detailed description of the sequential training procedure, the reproducible training setup, and the breadth of the system-level discussion. The framework also makes a falsifiable prediction: a code optimized over a finer set of puncturing lengths should outperform a coarser code at off-grid rates, which is testable with the proposed architecture.

major comments (3)
  1. [Definition 2, Sec. IV-A, Algorithm 2] The paper's abstract claims that the code is 'designed and optimized for a continuum of coding rates,' and Definition 2 requires inequality (3) to hold for every L in {0,...,Lbar}. However, the RLACS implementation trains only F discrete code lengths: training iteration i optimizes encoder/decoder i for input length C_i with no nulls (SVBSC(qo,epsilon,C_i,0)), so the trained decoders have never seen the null-padded inputs produced by Algorithm 2 when K-L falls between two consecutive C_i values. Code 3 avoids the problem only by choosing the C_i to coincide exactly with the finitely many modulation-induced rates (Sec. IV-E), and the experiments in Figs. 6 and 7 only test those on-grid rates. The construction and the evaluation therefore do not support the continuum claim; they support a claim about finitely many rates. Please either train the decoders on the actual distribution of null-padded inputs for all L, or revise the abstract and Definition 2 to describe a finite set of supported puncturing lengths.
  2. [Sec. IV-E, Figs. 6 and 7] The reported results come from a single training run with no error bars, confidence intervals, or multiple seeds. The central empirical claims—that Code 3 outperforms Code 1 by more than 5 dB in PSNR at low SNR and that fine granularity is critical (Code 2 vs. Code 3)—are therefore not shown to be statistically robust. Given that the training procedure includes stochastic optimization and random channel realizations, I ask that the authors report mean and variance over at least several training seeds, or provide error bars on the PSNR curves.
  3. [Definition 1 and Sec. IV-D] The SVBSC abstraction models the channel as a set of parallel memoryless binary symmetric channels, and the RLACS decoders are trained on i.i.d. bit flips. However, the stated test channel is block-fading Rician with uncoded QAM: during a deep fade, a large fraction of the bits in a frame can be corrupted simultaneously, producing burst errors rather than independent flips. Equation (1) only bounds the probability that the average BER exceeds qo, not the spatiotemporal correlation of the errors. The paper does not test whether the decoder trained on i.i.d. flips remains robust under correlated errors. I recommend adding an experiment with burst-error channels (e.g., block-fading realizations without the averaging implied by the SVBSC model) to validate the bit-pipe abstraction.
minor comments (4)
  1. [Sec. III-B] The effective rate expression 'R = (K-L)/M' appears to contain a typo: M is not defined in this context and should probably be N, the image dimension, to match the earlier definition R = K/N.
  2. [Sec. IV-E] The statement that the negligible gap between Code 3 and Code 1 at high SNR 'empirically demonstrates optimality' is too strong: Code 1 is another DNN, not a theoretical rate-distortion upper bound. I suggest rephrasing to say that the rateless code does not sacrifice performance at the maximum rate relative to this particular benchmark.
  3. [Throughout] There are several typographical errors that should be corrected, e.g., 'opern' (Sec. V), 'ralteless' (Sec. IV-F), 'espectral' (Sec. IV-E), 'ratelss' (abstract), and the notation-table row 'F; P; Pr' is unclear. A careful proofread is needed.
  4. [Sec. IV-D] The hand-picked values qo = 0.05 and epsilon = 0.05/0.1/0.01 are acknowledged as not optimized, which is reasonable; however, the paper would be strengthened by a brief sensitivity analysis around these values, since the QoS-metric interpretation in Sec. V-B2 depends on them.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular reduction found; RLACS performance claims are simulated outputs, with minor self-citations and a finite-training versus continuum overclaim that is a coverage gap, not circularity.

full rationale

The paper's derivation chain is not circular. The rateless JSCC objective (Definition 2, Eq. (3)) and the sequential training objective (Eq. (4)) are stated as design goals; the RLACS code is then trained on SVBSC(qo, epsilon, C_i, 0) channels (Sec. IV-A) and its PSNR behavior in Figs. 6-7 is measured from that trained system, not from a parameter fitted to the claimed outcome. The design values qo = 0.05 and epsilon are explicitly hand-picked with optimization deferred ('we do not exercise that optimization'), and Code 1 is an ablation, so no fitted parameter is relabeled as a prediction. The only self-citations ([48] for the channel-estimation noise model, and [52] for app/network interfacing) are peripheral and non-load-bearing; [48]'s model is externally standard and falsifiable, and [52] is a pointer, so neither forces the result. The most serious issue is an overclaim, not circularity: the abstract promises a 'continuum of coding rates,' but training covers only F discrete lengths, and Algorithm 2 pads off-grid puncturing with 0.5s that were never seen in training. The paper itself limits the claim in Sec. IV-A ('Ideally, the Cis are picked to cover all possible puncturing length L') and Sec. IV-E ('Cis adjusted such that the puncturing lengths ... aligns perfectly'), so Code 3 is only evaluated at the trained lengths. That missing generalization from finite C_i to all L is a coverage or proof gap, not an equation reducing to its own input. No step in the paper equates a prediction to a fitted parameter or imports a uniqueness conclusion from the author's prior work.

Assumptions & free parameters 6 free parameters · 5 assumptions · 2 invented entities

The central claim rests on several design choices: the SVBSC channel abstraction, the feasibility of training a binary-latent autoencoder with VIMCO, the availability of a reliable puncturing indicator to the decoder, and the hand-picked operating points (qo, epsilon, Ci). These are stated but not independently justified.

free parameters (6)
  • qo (target bit-flipping ratio) = 0.05
    Sets the BER level for the SVBSC training channel and the stability constraint in Eq. (1); the paper notes a tradeoff and defers its optimization.
  • epsilon (link instability) = 0.01, 0.05, 0.1 (test), 0 (training)
    Controls the probability that BER exceeds qo; the paper selects values that happen to give good spectral efficiency, post hoc.
  • Ci (code length granularity) = Code 3: Ci-Ci-1=128 for i=1..10; Code 2: 640; Code 1: C1=1280
    Determines which puncturing lengths are explicitly optimized; the paper notes granularity is critical for low SNR.
  • number of pilot symbols np = 10
    Sets the channel estimation noise variance in the imperfect CSI model (Sec. IV-E).
  • compression ratio = 1/24 (N=3072 pixels, W=128 channel uses)
    Design choice; matches the comparison to [32].
  • Rician K-factor = 20 dB
    Block fading test channel model in Sec. IV-E.
assumptions (5)
  • domain assumption The network can estimate the conditional distribution F_gamma|gamma_tilde and choose M via Eq. (1) so that Pr[q>qo]<=epsilon.
    Assumes the empirical Q(snr,M) mapping is accurate and the modulation selection stabilizes BER; Sec. III-A.
  • domain assumption The end-to-end channel is a memoryless binary symmetric channel after stabilization (SVBSC).
    Definition 1; ignores burst-error correlation from block fading, which is the weakest link in the model.
  • ad hoc to paper The AE with binary latent space can be trained end-to-end with VIMCO estimator to a good local optimum.
    No convergence guarantee; relies on 50 Monte Carlo samples and specific hyperparameters; Sec. IV-D.
  • domain assumption The decoder knows the puncturing length L and the punctured positions are the tail bits.
    Algorithm 2; requires a control signal from network to application, stated in footnote 4.
  • standard math Standard information-theoretic results used for motivation (e.g., suboptimality of separation in finite blocklength [21]).
    Background, not directly used in the construction.
invented entities (2)
  • Stabilized Vector Binary Symmetric Channel (SVBSC)
    purpose: Models the network-provided end-to-end link as a set of parallel BSCs with controlled BER and puncturing.
    A conceptual abstraction introduced in Def. 1; no empirical validation beyond the paper's own simulation.
  • Rateless JSCC code (and RLACS code family)
    purpose: A code whose distortion is controlled across a range of puncturing lengths, enabling network-side rate adaptation.
    The code is the paper's construction; its claims are supported only by the paper's own CIFAR10 experiments.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Rateless Joint Source-Channel Coding, and a Blueprint for 6G Semantic Communications System Design." pith.science (2026). https://pith.science/paper/PPA7VE5S

@misc{pith2026250206095,
  author       = {Pith},
  title        = {Pith review of: Rateless Joint Source-Channel Coding, and a Blueprint for 6G Semantic Communications System Design},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PPA7VE5S}},
  note         = {Machine review of arXiv:2502.06095}
}
read the original abstract

This paper introduces rateless joint source-channel coding (rateless JSCC). The code is rateless in that it is designed and optimized for a continuum of coding rates such that it achieves a desired distortion for any rate in that continuum. We further introduce rate-adaptive and stable communication link operation to accommodate rateless JSCCs. The link operation resembles a ``bit pipe'' that is identified by its rate in bits per frame, and, by the rate of bits that are flipped in each frame. Thus, the link operation is rate-adaptive such that it punctures the rateless JSCC codeword to adapt its length (and coding rate) to the underlying channel capacity, and is stable in maintaining the bit flipping ratio across time frames. Next, a new family of autoencoder rateless JSCC codes are introduced. The code family is dubbed RLACS code (read as relax code, standing for ratelss and lossy autoencoder channel and source code). The code is tested for reconstruction loss of image signals and demonstrates powerful performance that is resilient to variation of channel quality. RLACS code is readily applicable to the case of semantic distortion suited to variety of semantic and effectiveness communications use cases. In the second part of the paper, we dive into the practical concerns around semantic communication and provide a blueprint for semantic networking system design relying on updating the existing network systems with some essential modifications. We further outline a comprehensive list of open research problems and development challenges towards a practical 6G communications system design that enables semantic networking.

Figures

Figures reproduced from arXiv: 2502.06095 by the authors.

Figure 1
Figure 1. Rateless JSCC and rate-adaptive link for semantic co [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Training iteration i for RLACS code, where the trained encoder and decoder modules from previous iterations are frozen (blue color). Dashed arrow lines point to optimization of fi and gi based on the loss calculation. B. Application’s Operation: Rateless JSCC The code is designed to minimize the distortion over a binary channel where BER is bounded “around” qo, and for all the rates in the range [Rlo,Rhi], where Rlo… view at source ↗
Figure 3
Figure 3. Architecture of the DNN used for fθi and gφi are shown on the right and the left sides, respectively. The dimensions of the encoder and decoder layers adapts to Ci −Ci−1 and Ci , to ensure a sufficient capacity in the network architecture. Additionally, in the decoder structure, layer dimensions first expand through the middle of the network and then contract, using a factor z = min(1,256/Ci). To simplify the illust… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Average bit error rate experienced over the testing c [PITH_FULL_IMAGE:figures/full_fig_p021_4.png]
Figure 5
Figure 5. Figure 5: Average spectral efficincy of the curves in Fig. 4. Wit [PITH_FULL_IMAGE:figures/full_fig_p022_5.png]
Figure 6
Figure 6. Figure 6: Image reconstruction in PSNR for the three codes, tes [PITH_FULL_IMAGE:figures/full_fig_p023_6.png]
Figure 7
Figure 7. Figure 7: Image reconstruction of the testing process in PSNR fo [PITH_FULL_IMAGE:figures/full_fig_p024_7.png]
Figure 8
Figure 8. Figure 8: Progressive coding effect of the proposed rateless J [PITH_FULL_IMAGE:figures/full_fig_p026_8.png]
Figure 9
Figure 9. Figure 9: Illustration of the typical constituents of an end-t [PITH_FULL_IMAGE:figures/full_fig_p028_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

61 extracted references · 57 canonical work pages

  1. [1]

    Shannon and W

    C. Shannon and W. Weaver, The Mathematical Theory of Communication . University of Illinois Press, 1998

    work page 1998

  2. [2]

    Outline of a theory of strongly semantic inf ormation,

    L. Floridi, “Outline of a theory of strongly semantic inf ormation,” Minds and machines , vol. 14, pp. 197–221, 2004

    work page 2004

  3. [3]

    The semantics latent in shannon informatio n,

    A. M. Isaac, “The semantics latent in shannon informatio n,” The British Journal for the Philosophy of Science , 2019

    work page 2019

  4. [4]

    A theory of semantic information,

    Y . Zhong, “A theory of semantic information,” China communications , vol. 14, no. 1, pp. 1–17, 2017

    work page 2017

  5. [5]

    A theory of semantic com munication,

    Y . Shao, Q. Cao, and D. G¨ und¨ uz, “A theory of semantic com munication,” IEEE Transactions on Mobile Computing , 2024

    work page 2024

  6. [6]

    Model-t heoretic logic for mathematical theory of semantic informa tion and communication,

    A. F. Saz, S. Xiong, Y . M. Saidutta, and F. Fekri, “Model-t heoretic logic for mathematical theory of semantic informa tion and communication,” arXiv preprint arXiv:2401.17556 , 2024

    work page arXiv 2024

  7. [7]

    Deep hierarchi cal semantic segmentation,

    L. Li, T. Zhou, W. Wang, J. Li, and Y . Yang, “Deep hierarchi cal semantic segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2022, pp. 1246–1257

    work page 2022

  8. [8]

    Deep learning-based i mage semantic coding for semantic communications,

    D. Huang, X. Tao, F. Gao, and J. Lu, “Deep learning-based i mage semantic coding for semantic communications,” in 2021 IEEE Global Communications Conference (GLOBECOM) . IEEE, 2021, pp. 1–6

    work page 2021

Show all 61 references
  1. [9]

    Visual cod ing in a semantic hierarchy,

    Y . Yang, H. Zhang, M. Zhang, F. Shen, and X. Li, “Visual cod ing in a semantic hierarchy,” in Proceedings of the 23rd ACM international conference on Multimedia , 2015, pp. 59–68

    2015

  2. [10]

    Semantic image segm entation with contextual hierarchical models,

    M. Seyedhosseini and T. Tasdizen, “Semantic image segm entation with contextual hierarchical models,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 38, no. 5, pp. 951–964, 2016

    2016

  3. [11]

    Hierarchica l video content description and summarization using unified semantic and visual similarity,

    X. Zhu, J. Fan, A. K. Elmagarmid, and X. Wu, “Hierarchica l video content description and summarization using unified semantic and visual similarity,” Multimedia Systems , vol. 9, pp. 31–53, 2003

    2003

  4. [12]

    Joint sou rce-channel coding for video communications,

    F. Zhai, Y . Eisenberg, and A. K. Katsaggelos, “Joint sou rce-channel coding for video communications,” Handbook of Image and Video Processing , pp. 1065–1082, 2005

    2005

  5. [13]

    Auto-encoding variationa l bayes,

    D. P . Kingma and M. Welling, “Auto-encoding variationa l bayes,” 2022

    2022

  6. [14]

    Generative v isual compression: A review,

    B. Chen, S. Yin, P . Chen, S. Wang, and Y . Ye, “Generative v isual compression: A review,” arXiv preprint arXiv:2402.02140, 2024. February 11, 2025 DRAFT 37

    2024 arXiv

  7. [15]

    Evolving se mantic communication with generative model,

    S. Tang, Q. Yang, D. G¨ und¨ uz, and Z. Zhang, “Evolving se mantic communication with generative model,” arXiv preprint arXiv:2403.20237, 2024

    2024 arXiv

  8. [16]

    Efficient estimation of word representati ons in vector space,

    T. Mikolov, “Efficient estimation of word representati ons in vector space,” arXiv preprint arXiv:1301.3781 , vol. 3781, 2013

    2013 arXiv

  9. [17]

    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 . Xi ao, “Joint source–channel coding: Fundamentals and recent progress in practical designs,” Proceedings of the IEEE , 2024

    2024

  10. [18]

    Temporal and spatial error concea lment techniques for hierarchical MPEG-2 video codec,

    S. Aign and K. Fazel, “Temporal and spatial error concea lment techniques for hierarchical MPEG-2 video codec,” in Proceedings IEEE International Conference on Communicati ons ICC’95 , vol. 3. IEEE, 1995, pp. 1778–1783

    1995

  11. [19]

    What is semantic communication? a view on conveying meaning in the era of machine intelligence,

    Q. Lan, D. Wen, Z. Zhang, Q. Zeng, X. Chen, P . Popovski, an d K. Huang, “What is semantic communication? a view on conveying meaning in the era of machine intelligence,” Journal of Communications and Information Networks , vol. 6, no. 4, pp. 336–371, 2021

    2021

  12. [20]

    Coding theorems for a discrete source with a fidelity crite rion,

    C. E. Shannon et al. , “Coding theorems for a discrete source with a fidelity crite rion,” IRE Nat. Conv. Rec , vol. 4, no. 142-163, p. 1, 1959

    1959

  13. [21]

    Lossy joint source-channel c oding in the finite blocklength regime,

    V . Kostina and S. V erd´ u, “Lossy joint source-channel c oding in the finite blocklength regime,” IEEE Transactions on Information Theory , vol. 59, no. 5, pp. 2545–2575, 2013

    2013

  14. [22]

    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 . Xi ao, “Joint source–channel coding: Fundamentals and recent progress in practical designs,” Proceedings of the IEEE , pp. 1–32, 2024

    2024

  15. [23]

    Semantic communications with variable-length coding for extended reality,

    B. Zhang, Z. Qin, and G. Y . Li, “Semantic communications with variable-length coding for extended reality,” IEEE Journal of Selected Topics in Signal Processing , vol. 17, no. 5, pp. 1038–1051, 2023

    2023

  16. [24]

    V ariable length jo int source-channel coding of text using deep neural network s,

    M. Rao, N. Farsad, and A. Goldsmith, “V ariable length jo int source-channel coding of text using deep neural network s,” in 2018 IEEE 19th International W orkshop on Signal Processing Advances in Wireless Communications (SPAWC) , 2018, pp. 1–5

    2018

  17. [25]

    L ossy joint source-channel coding using Raptor codes,

    O. Y . Bursalioglu, M. Fresia, G. Caire, and H. V . Poor, “L ossy joint source-channel coding using Raptor codes,” International Journal of Digital Multimedia Broadcasting , vol. 2008, no. 1, p. 124685, 2008

    2008

  18. [26]

    Joint sou rce-channel coding for deep-space image transmission usin g rateless codes,

    O. Y . Bursalioglu, G. Caire, and D. Divsalar, “Joint sou rce-channel coding for deep-space image transmission usin g rateless codes,” IEEE Transactions on Communications , vol. 61, no. 8, pp. 3448–3461, 2013

    2013

  19. [27]

    Operational rate-distortion performance for joint source and channel coding of images,

    M. Ruf and J. Modestino, “Operational rate-distortion performance for joint source and channel coding of images,” IEEE Transactions on Image Processing , vol. 8, no. 3, pp. 305–320, 1999

    1999

  20. [28]

    Self-synchronizing variable-l ength codes for image transmission,

    A. Reibman and W. Lam, “Self-synchronizing variable-l ength codes for image transmission,” in Acoustics, Speech, and Signal Processing, IEEE International Conference on , vol. 3. Los Alamitos, CA, USA: IEEE Computer Society, mar 1992, pp. 477–480

    1992

  21. [29]

    Subband image coding using entropy-coded quantization over noisy channels,

    N. Tanabe and N. Farvardin, “Subband image coding using entropy-coded quantization over noisy channels,” IEEE Journal on Selected Areas in Communications , vol. 10, no. 5, pp. 926–943, 1992

    1992

  22. [30]

    Combined source -channel coding of images using the block cosine transform,

    J. Modestino, D. Daut, and A. Vickers, “Combined source -channel coding of images using the block cosine transform, ” IEEE Transactions on Communications , vol. 29, no. 9, pp. 1261–1274, 1981

    1981

  23. [31]

    Joint source and channel coding applied to the p yramid vector quantizer,

    M. Ruf, “Joint source and channel coding applied to the p yramid vector quantizer,” in Proceedings. IEEE International Symposium on Information Theory , 1993, pp. 391–391

    1993

  24. [32]

    Multi-level reliability interface for semantic communi cations over wireless networks,

    T.-Y . Tung, H. Esfahanizadeh, J. Du, and H. Viswanathan , “Multi-level reliability interface for semantic communi cations over wireless networks,” IEEE Transactions on Communications , pp. 1–1, 2025

    2025

  25. [33]

    Deep learning e nabled semantic communication systems,

    H. Xie, Z. Qin, G. Y . Li, and B.-H. Juang, “Deep learning e nabled semantic communication systems,” IEEE Transactions on Signal Processing , vol. 69, p. 2663–2675, 2021. February 11, 2025 DRAFT 38

    2021

  26. [34]

    Wireless image retrieval at the edge,

    M. Jankowski, D. Gunduz, and K. Mikolajczyk, “Wireless image retrieval at the edge,” IEEE Journal on Selected Areas in Communications , vol. 39, no. 1, p. 89–100, Jan. 2021

    2021

  27. [35]

    Deep j oint source-channel coding for wireless image transmissio n,

    E. Bourtsoulatze, D. Burth Kurka, and D. Gunduz, “Deep j oint source-channel coding for wireless image transmissio n,” IEEE Transactions on Cognitive Communications and Network ing, vol. 5, no. 3, p. 567–579, Sep. 2019

    2019

  28. [36]

    Deep joint source-channel coding for semantic communications,

    J. Xu, T.-Y . Tung, B. Ai, W. Chen, Y . Sun, and D. G¨ und¨ uz, “Deep joint source-channel coding for semantic communications,” IEEE communications Magazine , vol. 61, no. 11, pp. 42–48, 2023

    2023

  29. [37]

    Secure Deep-JSCC against multiple eavesdroppers,

    S. A. A. Kalkhoran, M. Letafati, E. Erdemir, B. H. Khalaj , H. Behroozi, and D. G¨ und¨ uz, “Secure Deep-JSCC against multiple eavesdroppers,” in IEEE Global Communications Conference . arXiv, 2023

    2023

  30. [38]

    Nonlinear prin cipal component analysis: neural network models and applic ations,

    M. Scholz, M. Fraunholz, and J. Selbig, “Nonlinear prin cipal component analysis: neural network models and applic ations,” in Principal manifolds for data visualization and dimension r eduction. Springer, 2008, pp. 44–67

    2008

  31. [39]

    Sparse autoencoder,

    A. Ng et al. , “Sparse autoencoder,” CS294A Lecture notes , vol. 72, no. 2011, pp. 1–19, 2011

    2011

  32. [40]

    Stochastic bottleneck: Ra teless auto-encoder for flexible dimensionality reduction ,

    T. Koike-Akino and Y . Wang, “Stochastic bottleneck: Ra teless auto-encoder for flexible dimensionality reduction ,” 2020

    2020

  33. [41]

    Rateless deep graph joint source channel codi ng for holographic-type communication,

    S. Ibuki, T. Okamoto, T. Fujihashi, T. Koike-Akino, and T. Watanabe, “Rateless deep graph joint source channel codi ng for holographic-type communication,” in GLOBECOM 2023 - 2023 IEEE Global Communications Conference , 2023, pp. 3330–3335

    2023

  34. [42]

    Rateless coding for gaussian channels,

    U. Erez, M. D. Trott, and G. W. Wornell, “Rateless coding for gaussian channels,” IEEE Transactions on Information Theory, vol. 58, no. 2, pp. 530–547, 2012

    2012

  35. [43]

    Residual attention network for image classification,

    F. Wang, M. Jiang, C. Qian, S. Yang, C. Li, H. Zhang, X. Wan g, and X. Tang, “Residual attention network for image classification,” in Proceedings of the IEEE conference on computer vision and pa ttern recognition, 2017, pp. 3156–3164

    2017

  36. [44]

    V ariational inference for Mo nte Carlo objectives,

    A. Mnih and D. J. Rezende, “V ariational inference for Mo nte Carlo objectives,” in International Conference on Machine Learning. arXiv, 2016

    2016

  37. [45]

    PyTorch: An imperative style, high-performance deep lear ning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. C hanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga et al. , “PyTorch: An imperative style, high-performance deep lear ning library,” 2019

    2019

  38. [46]

    Learning multiple layers of features f rom tiny images,

    A. Krizhevsky, “Learning multiple layers of features f rom tiny images,” University of Toronto, Tech. Rep., 2009

    2009

  39. [47]

    How much training is needed in multiple-antenna wireless links?

    B. Hassibi and B. Hochwald, “How much training is needed in multiple-antenna wireless links?” IEEE Transactions on Information Theory , vol. 49, no. 4, pp. 951–963, 2003

    2003

  40. [48]

    Exploiti ng diversity for ultra-reliable and low-latency wireless c ontrol,

    S. R. Khosravirad, H. Viswanathan, and W. Y u, “Exploiti ng diversity for ultra-reliable and low-latency wireless c ontrol,” IEEE Transactions on Wireless Communications , vol. 20, no. 1, pp. 316–331, 2021

    2021

  41. [49]

    Progressive image coding o n noisy channels,

    P . G. Sherwood and K. Zeger, “Progressive image coding o n noisy channels,” in Proceedings DCC’97. Data Compression Conference. IEEE, 1997, pp. 72–81

    1997

  42. [50]

    W. Saad, C. Chaccour, C. K. Thomas, and M. Debbah, F oundations of Semantic Communication Networks . John Wiley & Sons, 2024

    2024

  43. [51]

    A lighter UDP,

    L.- ˚A. Larzon, “A lighter UDP,” Ph.D. dissertation, Lule˚ a tekn iska universitet, 1999

    1999

  44. [52]

    Com munication and control interfacing for co-design of wirele ss control systems,

    J. Li, S. R. Khosravirad, J. Du, W. Liu, and U. Mitra, “Com munication and control interfacing for co-design of wirele ss control systems,” in 2023 IEEE 97th V ehicular Technology Conference (VTC2023-S pring), 2023, pp. 1–5

    2023

  45. [53]

    DeepRx: fully convolutional deep learning receiver,

    M. Honkala, D. Korpi, and J. M. Huttunen, “DeepRx: fully convolutional deep learning receiver,” IEEE Transactions on Wireless Communications, vol. 20, no. 6, pp. 3925–3940, 2021

    2021

  46. [54]

    An overview of neural network compression ,

    J. O. Neill, “An overview of neural network compression ,” arXiv preprint arXiv:2006.03669 , 2020

    2006 arXiv

  47. [55]

    Beyond transmitting bits: Context, semantics, and task-oriented communicatio ns,

    D. G¨ und¨ uz, Z. Qin, I. E. Aguerri, H. S. Dhillon, Z. Yang , A. Yener, K. K. Wong, and C.-B. Chae, “Beyond transmitting bits: Context, semantics, and task-oriented communicatio ns,” IEEE Journal on Selected Areas in Communications , vol. 41, no. 1, pp. 5–41, 2022. February 1...

    2022

  48. [56]

    S emantic communication for the internet of sounds: Architecture, design principles, and challenges,

    C. Liang, Y . Sun, C. K. Thomas, L. Mohjazi, and W. Saad, “S emantic communication for the internet of sounds: Architecture, design principles, and challenges,” arXiv preprint arXiv:2407.12203 , 2024

    2024 arXiv

  49. [57]

    Semantic comm unications: Principles and challenges,

    Z. Qin, X. Tao, J. Lu, W. Tong, and G. Y . Li, “Semantic comm unications: Principles and challenges,” arXiv preprint arXiv:2201.01389, 2021

    2021 arXiv

  50. [58]

    Semantic communication s: Overview, open issues, and future research directions,

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

    2022

  51. [59]

    From semantic communi cation to semantic-aware networking: Model, architecture , and open problems,

    G. Shi, Y . Xiao, Y . Li, and X. Xie, “From semantic communi cation to semantic-aware networking: Model, architecture , and open problems,” IEEE Communications Magazine , vol. 59, no. 8, pp. 44–50, 2021

    2021

  52. [60]

    Next generation advanced transceiver technologies for 6G,

    C. Y ou, Y . Cai, Y . Liu, M. Di Renzo, T. M. Duman, A. Yener, a nd A. L. Swindlehurst, “Next generation advanced transceiver technologies for 6G,” arXiv preprint arXiv:2403.16458 , 2024

    2024 arXiv

  53. [61]

    A survey on goal-o riented semantic communication: Techniques, challenges, and future directions,

    T. M. Getu, G. Kaddoum, and M. Bennis, “A survey on goal-o riented semantic communication: Techniques, challenges, and future directions,” IEEE Access , 2024. February 11, 2025 DRAFT

    2024

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.