Towards Compatible Semantic Communication: A Perspective on Digital Coding and Modulation

Guanding Yu; Guangyi Zhang; Kequan Zhou; Qiyu Hu; Yunlong Cai

REVIEW 4 major objections 4 minor 15 references

Towards Compatible Semantic Communication: A Perspective on Digital Coding and Modulation

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Digital semantic communication can work by pre-training in analog and fine-tuning for digital transmission.

desk verdict A readable perspective on digital semantic communication with a sensible analog-pretrain/digital-fine-tune recipe, but the supporting case study is too thin to back the strong empirical claims. read the letter →

arxiv 2412.18876 v1 pith:IY6GYQKX submitted 2024-12-25 eess.SP

Guangyi Zhang , Kequan Zhou , Yunlong Cai , Qiyu Hu , Guanding Yu This is my paper · ORCID

classification eess.SP

keywords semanticcommunicationdigitalmodulationjointsource-channelcodingvectorquantizationconstellationdesignanalogpre-trainingprobabilisticanddeterministic6Gnetworks

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

Semantic communication aims to transmit only task-relevant information, but most deep-learning designs send continuous-valued analog symbols that do not fit today's digital radios. This paper argues that digital semantic communication can be made effective by treating digitization as a downstream task: first pre-train the semantic encoder in the analog domain, then fine-tune it for digital transmission. The paper organizes digital designs into probabilistic methods, which sample symbols from a learned distribution, and deterministic methods, which quantize the encoder output, and it identifies constellation design as a key lever. On CIFAR-10, an analog-pretrained model fine-tuned with an irregular constellation learned by K-means clustering outperforms all other digital schemes, suggesting that compatibility with digital hardware need not sacrifice accuracy.

What carries the argument

The load-bearing mechanism is the multistage design strategy: unsupervised or self-supervised pre-training of the semantic encoder in the analog domain, followed by task-specific fine-tuning in the digital domain with a chosen modulator. Around this, the paper organizes the space of digital designs into probabilistic modulation (a stochastic encoder that samples channel symbols from a learned distribution, trained with variational inference and gradient-estimation tricks) and deterministic modulation (quantizer-dequantizer pairs using scalar, symbol, or vector quantization with differentiable approximations such as the straight-through estimator or additive uniform noise). The third piece of machinery is constellation design, in which the non-uniform statistics of the analog representation motivate non-uniform decision boundaries or learnable irregular constellations, as in the K-means-derived constellation of the case study.

What would settle it

Run the same architecture from random initialization directly through digital modulation (e.g., STE training) and compare with an analog-pre-trained, digital-fine-tuned model across several datasets and channel conditions; if the randomly initialized digital model matches or beats the transferred model, the claimed benefit of analog pre-training is not load-bearing.

Watch

Extended reading notes

Core claim

The paper's central claim is that the transition from analog to digital semantic communication should be treated as a downstream task rather than as a separate training problem. The proposed multistage strategy first pre-trains the joint source-channel encoder on continuous analog signals, then transfers the learned parameters to the digital domain for fine-tuning with a concrete modulation scheme. For the fine-tuning stage, the paper proposes three constellation designs: standard constellations with uniform mapping, standard constellations with non-uniform mapping adjusted to the non-uniform statistics of the representation, and irregular constellations obtained by clustering the continuous representation (e.g., K-means) and optionally refining the cluster centers as learnable parameters. The case study reports that the scheme fine-tuned with irregular constellations outperforms all other digital schemes, and that pre-training in the analog domain markedly improves the expressive capability of the digital model.

Load-bearing premise

The recipe assumes that a semantic representation learned on continuous analog symbols retains its useful structure after being quantized to a finite set of digital symbols, so that fine-tuning from an analog pre-trained encoder converges to a good digital solution.

Editorial extensions

If this is right

Higher modulation order improves digital semantic communication performance, in contrast to conventional systems where lower orders typically achieve better bit error rates.
Pre-training in the analog domain substantially improves the expressive capability of the digital fine-tuned model compared to training directly in the digital domain.
Fine-tuning with an irregular constellation learned from the representation's statistics outperforms standard-constellation digital schemes, making data-driven constellation design a first-order lever.
Digital semantic communication systems are expected to resist distortion accumulation in multi-round transmission better than analog systems, and vector quantization offers inherent robustness against semantic attacks.
A hybrid scalar-plus-vector quantization scheme could support scalable generative transmission at low bit rates.

Reading between the lines

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

If analog-to-digital transfer holds beyond CIFAR-10, the same pretrain-then-quantize recipe should apply to other modalities and other discretization schemes, so an natural test is whether it extends to audio, video, or higher-resolution images.
The case study's irregular constellation is effectively a learned codebook, so the boundary between constellation design and vector quantization likely dissolves as modulation orders grow.
The security argument suggests digital semantic communication could inherit classical encryption and codebook secrecy; a testable extension is whether vector-quantization-based digital semantic communication resists model-inversion attacks better than analog semantic communication at equal task accuracy.
The claim that higher modulation order is always better in digital semantic communication may be an artifact of the AWGN channel and modest constellation sizes; verifying it under fading and with practical bit-mapping would be a strong stress test.

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.

Desk Editor's Note

read the letter

You should know this paper for the taxonomy and the recipe, not for the numbers. The authors split digital SC into probabilistic and deterministic methods, which is a fair organizing frame, and they give a clear rundown of scalar, symbol, and vector quantization with the usual gradient tricks. The multistage proposal — pretrain in analog, fine-tune in digital, optionally with an irregular constellation — is the real takeaway, and it is actionable. That part deserves a thoughtful referee.

What is genuinely new is the framing of the digitization step as a downstream task, with the analog-to-digital transfer as a pretraining strategy. Prior work already used multi-phase training, so the novelty is modest, but the paper says it clearly and connects it to representation learning. I also like the explicit discussion of non-uniform representations and why standard constellations may be suboptimal.

The soft spots are exactly where the reader points. The entire empirical case is one CIFAR-10 experiment, with no architecture, no optimizer details, no epoch counts, and no error bars. The comparison between direct STE and STE with fine-tuning cannot separate the benefit of analog pretraining from the extra gradient steps. The irregular constellation is built from the analog model's own outputs, so a matched codebook could explain part of the gain regardless of transfer. The claim that higher modulation order is generally preferable is not rate-matched; higher order at fixed symbol count simply sends more bits. And the reliability advantage in Fig. 5 is hand-wavy, citing the authors' own prior work without quantitative data here.

None of this kills the paper as a perspective. The taxonomy and the design principles are plausible and useful. But the conclusions as written overstate the evidence. If this goes to review, the authors should either add proper experimental controls — rate matching, seed statistics, equal training budgets, hyperparameters — or soften the "generally preferable" and "clearly demonstrates" language.

Verdict: worth a serious referee, especially if the venue treats it as a tutorial/survey piece. I would not block publication on the thin experiment if the claims are hedged, but I would insist on the missing details before the design principles are presented as established. I'd cite this for the taxonomy and the recipe, not for the empirical results.

Referee Report

4 major / 4 minor

Summary. This paper argues that semantic communication (SC) can be made compatible with existing digital infrastructure by separating the modulation/coding stage from the joint source-channel coding, and it proposes two design paradigms: probabilistic and deterministic digitization. It then presents a multistage design recipe: pre-train the semantic encoder in the analog domain, transfer the parameters to the digital domain, and fine-tune with either standard or data-driven irregular constellations. The paper also discusses inherent advantages of digital SC (multiround reliability, security, scalability) and future research directions. The central evidence for the proposed recipe is a CIFAR-10 image transmission case study in Section IV.C, where analog-pretrained and fine-tuned digital schemes are reported to outperform digital schemes trained from scratch, with irregular constellations performing best.

Significance. If the empirical claims are correct, the paper offers a genuinely useful perspective: a concrete, hardware-compatible design recipe for digital SC, with a clear taxonomy and falsifiable design principles. The paper's strengths are its systematic organization, the explicit statement of the analog-pretraining/digital-fine-tuning principle, and the inclusion of a simple case study that directly tests the principle. However, the experimental evidence is currently too thin to support the strength of the claims: the case study is limited to one dataset, one channel model, and no statistical characterization, and the reported comparisons are confounded by bit rate and training budget. The conceptual contribution is worth publishing if the evidence is strengthened or the claims are appropriately moderated.

major comments (4)

[Section IV.C, Fig. 4] The central empirical claim—that analog pretraining followed by digital fine-tuning outperforms training the digital system from scratch—rests on a single CIFAR-10 experiment with no architecture description, optimizer, number of epochs, learning-rate schedule, batch size, number of seeds, or error bars. The paper states that all benchmark models share the same architecture and training settings, but this is not sufficient to rule out that 'STE (Fine-tune)' receives more total gradient steps than 'STE' or that the reported difference is within seed-to-seed variability. Please provide the full experimental setup and report results over multiple seeds with error bars, or explicitly recast the case study as illustrative rather than evidential.
[Section IV.C, observation (i)] The statement that 'higher modulation order is generally preferable' is confounded with bit rate: at a fixed number of transmitted symbols, increasing the modulation order M transmits more bits, so the performance gain may simply reflect a larger transmission budget. The comparison should be made at equal bit rate (for example by varying the number of symbols per image accordingly) or the claim should be restricted to the specific rate regime tested. As written, the 'generally preferable' conclusion is not supported by the reported experiment.
[Section IV.B.2 and IV.C] The proposed design principle assumes that a representation space learned on continuous analog symbols remains a good initialization for the digital fine-tuning stage, i.e., that quantization does not irreversibly destroy the semantic structure. This premise is load-bearing but not directly validated. The only evidence is the relative performance in Fig. 4, which is subject to the confounds described above. Please provide a direct analysis of the transfer assumption—for example, representation-space similarity before and after fine-tuning, or the performance of randomly initialized digital training with the same total compute—or clearly identify this as an open hypothesis.
[Section V.A.1] The claimed advantage of digital SC for multiround transmission is supported only by 'Our initial results, illustrated in Fig. 5' with no numerical details, experimental protocol, or comparison baselines. As stated, this claim is not verifiable. Please either include the quantitative setup and results or explicitly mark this as a preliminary illustration rather than a demonstrated advantage.

minor comments (4)

[Fig. 4 caption] The caption text contains corrupted font substitutions (the sequence of /uni... glyphs) and the figure appears to display unreadable labels. Please regenerate the figure and caption with a standard font so that the scheme names and axes are legible.
[Section III.B] The statement that symbol quantization 'can be viewed as a special case when the modulation order equals the number of quantization levels in scalar quantization' is imprecise, because symbol quantization operates on the full constellation vector while scalar quantization operates element-wise. Clarify the intended equivalence.
[References] Several key claims about benefits of digital SC cite the authors' own prior work ([13] for multiround reliability); please either provide more independent evidence or summarize the relevant results in-text so the reader can assess the strength of the support.
[General] No code, data, or model checkpoints are provided. Given the small scale of the case study, releasing code and training details would substantially increase the reproducibility of the central claims.

Circularity Check

0 steps flagged · score 2.0 of 10

No definitional circularity: the multistage analog-pretrain/digital-fine-tune proposal is a recommended recipe evaluated by a case study, and the self-citations are peripheral; the central comparison is not forced by construction.

full rationale

The paper's load-bearing proposal—analog pre-training followed by digital fine-tuning (Section IV.B.2)—is presented as a design principle and is evaluated empirically in Section IV.C, not derived from a fitted quantity or from the cited prior work. The comparisons STE vs. STE (Fine-tune) vs. STE (Irregular) involve possible experimental confounds (training budget, rate matching, seeds), but those are methodological validity concerns, not circularity: the reported advantage is not defined as equal to the training strategy by construction. The irregular constellation is produced by K-means on analog outputs, yet the claim that it helps after fine-tuning is an empirical outcome, not an input to its own proof. Self-citations are present—[13] for multi-hop distortion accumulation and [7] for Gumbel-Softmax—but they are peripheral to the central case study; [13] supports only the future-direction discussion of multiround reliability. No uniqueness theorem is imported from the authors' prior work, and no ansatz is smuggled in via citation. The observation that higher modulation order is preferable may be rate-confounded, but that is a benchmark-construction issue rather than a circular step. No equation in the paper reduces a predicted result to a fitted parameter or to a self-citation. Hence no significant circularity; the score reflects only a minor, non-load-bearing self-citation.

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

The proposed recipe depends on experimental design choices (modulation order, cluster count, bit width) that are not reported, and on two domain assumptions: that analog-pretrained features transfer to digital fine-tuning, and that semantic representations are non-uniform in a way that makes standard QAM suboptimal. No new physical or mathematical entities are introduced.

free parameters (3)

Modulation order M = not stated (e.g., 4, 16, 64 per Fig. 4)
The modulation order is varied in the case study and the conclusion that higher M improves performance depends on this choice; exact values and rate matching are not reported.
Number of K-means clusters K for irregular constellation = not stated
The 'STE (Irregular)' scheme uses K-means clustering to generate constellation points; the choice of K (number of constellation symbols) is a free design parameter not reported.
Quantization bit-width / number of levels = not stated
For scalar or vector quantization, the number of discrete levels determines rate and is chosen by the designer; not reported in the case study.

assumptions (3)

ad hoc to paper The JSC encoder trained in the analog domain produces a representation space that is a good initialization for digital fine-tuning.
Section IV.B.2 states this as a design principle; the paper provides only one empirical demonstration, so this is a load-bearing domain assumption rather than a proven fact.
domain assumption Channel noise on digital symbols can be modeled as independent AWGN on the received constellation, and errors are limited to threshold crossings.
Used throughout the case study and in Section V.A for the reliability advantage; this is a standard communications assumption, but it shapes all results.
ad hoc to paper Semantic representations are non-uniformly distributed (as shown in Fig. 3) so that standard uniform constellations are suboptimal.
Section IV.B.3 asserts this based on 'our analysis of the output from an analog SC system'; the evidence is illustrative, not quantitative, and is load-bearing for the irregular-constellation recommendation.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Towards Compatible Semantic Communication: A Perspective on Digital Coding and Modulation." pith.science (2026). https://pith.science/paper/IY6GYQKX

@misc{pith2026241218876,
  author       = {Pith},
  title        = {Pith review of: Towards Compatible Semantic Communication: A Perspective on Digital Coding and Modulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IY6GYQKX}},
  note         = {Machine review of arXiv:2412.18876}
}

read the original abstract

Semantic communication (SC) is emerging as a pivotal innovation within the 6G framework, aimed at enabling more intelligent transmission. This development has led to numerous studies focused on designing advanced systems through powerful deep learning techniques. Nevertheless, many of these approaches envision an analog transmission manner by formulating the transmitted signals as continuous-valued semantic representation vectors, limiting their compatibility with existing digital systems. To enhance compatibility, it is essential to explore digitized SC systems. This article systematically identifies two promising paradigms for designing digital SC: probabilistic and deterministic approaches, according to the modulation strategies. For both, we first provide a comprehensive analysis of the methodologies. Then, we put forward the principles of designing digital SC systems with a specific focus on informativeness and robustness of semantic representations to enhance performance, along with constellation design. Additionally, we present a case study to demonstrate the effectiveness of these methods. Moreover, this article also explores the intrinsic advantages and opportunities provided by digital SC systems, and then outlines several potential research directions for future investigation.

Figures

Figures reproduced from arXiv: 2412.18876 by the authors.

**Figure 1.** The illustration of the paradigm shift of SC systems. (a) Separation-based systems denote the conventional scheme that exploits a combination of [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 2.** The illustrations of the probabilistic method and deterministic method. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** The illustration of the proposed multistage design strategy. For the constellation illustration, we view half of the semantic representation output by a [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: The performance of different schemes with different modulation [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

**Figure 5.** Figure 5: Three advantages of digital SC system for further investigation. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 14 canonical work pages

[1]

Digital Semantic Communications: An Alternating Multi-Phase Training Strategy with Mask Attack

M. Gong, S. Wang, S. Bi, Y . Wu, and L. Qian, “Digital semantic communications: An alternating multi-phase training strategy with mask attack,” arXiv preprint arXiv:2408.04972 , 2024

work page Pith review arXiv 2024
[2]

Toward intelligent communications: Large model empowered semantic communica- tions,

H. Xie, Z. Qin, X. Tao, and Z. Han, “Toward intelligent communications: Large model empowered semantic communica- tions,” IEEE Commun. Mag. , pp. 1–7, Jul. to appear, 2024, doi: 10.1109/MCOM.001.2300807

work page doi:10.1109/mcom.001.2300807 2024
[3]

Learning discrete representations via information maximizing self-augmented training,

W. Hu, T. Miyato, S. Tokui, E. Matsumoto, and M. Sugiyama, “Learning discrete representations via information maximizing self-augmented training,” in Proc. Int. Conf. Mach. Learn. , vol. 70, Aug. 2017, pp. 1558–1567

work page 2017
[4]

Neural joint source-channel coding,

K. Choi, K. Tatwawadi, A. Grover, T. Weissman, and S. Ermon, “Neural joint source-channel coding,” in Proc. Int. Conf. Mach. Learn. (ICML) , vol. 97, Long Beach, California, USA, June 2019, pp. 1182–1192

work page 2019
[5]

Joint coding-modulation for digital semantic communications via variational autoencoder,

Y . Bo, Y . Duan, S. Shao, and M. Tao, “Joint coding-modulation for digital semantic communications via variational autoencoder,” IEEE Trans. Commun., to appear, 2024, doi: 10.1109/TCOMM.2024.3386577

work page arXiv 2024
[6]

Variational inference: A review for statisticians,

D. M. Blei, A. Kucukelbir, and J. D. McAuliffe, “Variational inference: A review for statisticians,” J. Amer. Stat. Assoc. , vol. 112, no. 518, pp. 859–877, Jul. 2017

work page 2017
[7]

A unified multi- task semantic communication system for multimodal data,

G. Zhang, Q. Hu, Z. Qin, Y . Cai, G. Yu, and X. Tao, “A unified multi- task semantic communication system for multimodal data,” IEEE Tran. Commun., vol. 72, no. 7, pp. 4101–4116, July 2024

work page 2024
[8]

OFDM-based digital semantic communication with importance awareness,

C. Liu, C. Guo, Y . Yang, W. Ni, and T. Q. S. Quek, “OFDM-based digital semantic communication with importance awareness,” IEEE Trans. Commun., vol. 72, no. 10, pp. 6301–6315, Oct. 2024

work page 2024

Show all 15 references

[9]

Deep source-channel coding for sentence semantic transmission with HARQ,

P. Jiang, C.-K. Wen, S. Jin, and G. Y . Li, “Deep source-channel coding for sentence semantic transmission with HARQ,” IEEE Trans. Commun., vol. 70, no. 8, pp. 5225–5240, Aug. 2022

2022
[10]

End-to-end optimized image compression,

J. Ball ´e, V . Laparra, and E. P. Simoncelli, “End-to-end optimized image compression,” in Proc. Int. Conf. Learn. Represent. (ICLR) , Toulon, France, Apr. 2017

2017
[11]

Joint source-channel coding for channel-adaptive digital semantic communications,

J. Park, Y . Oh, S. Kim, and Y .-S. Jeon, “Joint source-channel coding for channel-adaptive digital semantic communications,” IEEE Trans. Cognitive Commun. Netw. , to appear, 2024, doi:10.1109/TCCN.2024.3422496

2024
[12]

Constellation design for deep joint source-channel coding,

M. Wang, J. Li, M. Ma, and X. Fan, “Constellation design for deep joint source-channel coding,” IEEE Signal Process. Lett. , vol. 29, pp. 1442–1446, Mar. 2022

2022
[13]

Alleviating distortion accumu- lation in multi-hop semantic communication,

G. Zhang, Q. Hu, Y . Cai, and G. Yu, “Alleviating distortion accumu- lation in multi-hop semantic communication,” IEEE Commun. Lett. , vol. 28, no. 2, pp. 308–312, Feb. 2024

2024
[14]

Is semantic communication secure? a tale of multi-domain adversarial attacks,

Y . E. Sagduyu, T. Erpek, S. Ulukus, and A. Yener, “Is semantic communication secure? a tale of multi-domain adversarial attacks,” IEEE Commun. Mag. , vol. 61, no. 11, pp. 50–55, Nov. 2023

2023
[15]

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, Aug. 2024. Guangyi Zhang [S] (zhangguangyi@zju.edu.cn) is currently pursuin...

2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

Digital Semantic Communications: An Alternating Multi-Phase Training Strategy with Mask Attack

M. Gong, S. Wang, S. Bi, Y . Wu, and L. Qian, “Digital semantic communications: An alternating multi-phase training strategy with mask attack,” arXiv preprint arXiv:2408.04972 , 2024

work page Pith review arXiv 2024

[2] [2]

Toward intelligent communications: Large model empowered semantic communica- tions,

H. Xie, Z. Qin, X. Tao, and Z. Han, “Toward intelligent communications: Large model empowered semantic communica- tions,” IEEE Commun. Mag. , pp. 1–7, Jul. to appear, 2024, doi: 10.1109/MCOM.001.2300807

work page doi:10.1109/mcom.001.2300807 2024

[3] [3]

Learning discrete representations via information maximizing self-augmented training,

W. Hu, T. Miyato, S. Tokui, E. Matsumoto, and M. Sugiyama, “Learning discrete representations via information maximizing self-augmented training,” in Proc. Int. Conf. Mach. Learn. , vol. 70, Aug. 2017, pp. 1558–1567

work page 2017

[4] [4]

Neural joint source-channel coding,

K. Choi, K. Tatwawadi, A. Grover, T. Weissman, and S. Ermon, “Neural joint source-channel coding,” in Proc. Int. Conf. Mach. Learn. (ICML) , vol. 97, Long Beach, California, USA, June 2019, pp. 1182–1192

work page 2019

[5] [5]

Joint coding-modulation for digital semantic communications via variational autoencoder,

Y . Bo, Y . Duan, S. Shao, and M. Tao, “Joint coding-modulation for digital semantic communications via variational autoencoder,” IEEE Trans. Commun., to appear, 2024, doi: 10.1109/TCOMM.2024.3386577

work page arXiv 2024

[6] [6]

Variational inference: A review for statisticians,

D. M. Blei, A. Kucukelbir, and J. D. McAuliffe, “Variational inference: A review for statisticians,” J. Amer. Stat. Assoc. , vol. 112, no. 518, pp. 859–877, Jul. 2017

work page 2017

[7] [7]

A unified multi- task semantic communication system for multimodal data,

G. Zhang, Q. Hu, Z. Qin, Y . Cai, G. Yu, and X. Tao, “A unified multi- task semantic communication system for multimodal data,” IEEE Tran. Commun., vol. 72, no. 7, pp. 4101–4116, July 2024

work page 2024

[8] [8]

OFDM-based digital semantic communication with importance awareness,

C. Liu, C. Guo, Y . Yang, W. Ni, and T. Q. S. Quek, “OFDM-based digital semantic communication with importance awareness,” IEEE Trans. Commun., vol. 72, no. 10, pp. 6301–6315, Oct. 2024

work page 2024