Pith. sign in

REVIEW 3 major objections 5 minor 51 references

Index modulation carries a semantic stream's most vital bits, and a new digital system shows the pairing wins.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

Residual-quantized semantic features sent over fluid-antenna index modulation with semantic-aware stream splitting beat prior digital semantic baselines in image reconstruction.

T0 review reviewed 2026-08-01 challenge →

load-bearing objection Competent cross-layer integration of RQ and FA-IM with a genuinely new semantic-aware splitter; the tested results are convincing, but the load-bearing stream-ordering claim is measured on raw error counts and under-validated. the 3 major comments →

arxiv 2607.19934 v1 pith:O4BCAG5S submitted 2026-07-22 cs.IT math.IT

Spatial Semantic Communication: When Semantic Transmission Meets Index Modulation

Xinghao Guo , Yin Xu , Dazhi He , Hanjiang Hong , Zhiyong Chen , Cixiao Zhang , Yiyan Wu , Wenjun Zhang This is my paper
classification cs.IT math.IT
keywords semantic communicationjoint source-channel codingindex modulationfluid antennaresidual quantizationunequal error protectiondigital transmissionimage reconstruction
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper builds a digital semantic communication system called SSC that is the first to marry semantic transmission with index modulation. It converts a neural JSCC encoder's continuous latent features into discrete codeword indices with residual quantization, then sends those indices over a fluid-antenna index modulation link. The core insight is a semantic-aware stream splitter: earlier RQ steps are more semantically important, and the FA-IM port-index stream is more reliable than the symbol stream, so the system assigns important indices to the robust stream. Experiments at equal bit rates show SSC outperforming VQ-based digital baselines and a traditional separation-based codec in PSNR and MS-SSIM across the tested SNR range.

Core claim

On its own terms, the paper establishes that the hierarchical significance of residual-quantization steps can be aligned with the heterogeneous reliability of index-modulation substreams to yield a concrete unequal-error-protection gain. Specifically, for a latent vector quantized in N_q steps, the index chosen at step 1 dominates reconstruction fidelity; meanwhile, in FA-IM with a single fluid antenna and a fixed port set, the port-index bits suffer fewer errors than constellation-symbol bits at the tested configurations. The SSC system therefore maps early RQ indices to the port-index stream and late refinement indices to the symbol stream. With this design, and a three-stage training stra

What carries the argument

The mechanism is residual quantization (RQ) coupled with a stream splitter. RQ decomposes each continuous latent vector into a sequence of codeword indices, each chosen to minimize the residual error, so the first index carries the coarse structure and later indices add refinement. The stream splitter orders the per-step index columns by semantic significance and allocates the leading columns to the FA-IM port-index stream (which conveys log2(N_s) bits by selecting one of N_s pre-selected ports) and the remaining columns to the constellation-symbol stream (which conveys log2(M) bits). The allocation is decided offline by Monte Carlo error counts for the two streams, exploiting their relative

Load-bearing premise

The entire semantic-aware splitting gain rests on the claim that the port-index stream is always more reliable than the symbol stream for any given FA-IM configuration—a property the paper validates only for four parameter combinations and under a specific fading model.

What would settle it

Run FA-IM simulations at a wider range of configurations (e.g., N_s = 8, M = 4, larger W_t, or spatially correlated fading) and compare the error counts of port indices vs. constellation symbols; if the ordering flips at any practical operating point, the proposed fixed splitting becomes anti-UEP and the reported gains should reverse. Alternatively, implement SSC with the splitter reversed (important bits to the symbol stream) and show that it outperforms the proposed order.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Digital semantic communication systems can adopt index modulation without redesigning their analog JSCC backbone; the RQ and splitting modules sit between the encoder and the modulator.
  • Residual quantization provides an exponential effective codebook size with a small physical codebook and low search complexity, which removes a major bottleneck for VQ-based semantic transmitters.
  • The semantic-aware splitting principle transfers to any IM variant with two streams of unequal reliability, such as spatial modulation or subcarrier-index modulation, and to any progressive quantizer with ordered significance.
  • At equal bits-per-pixel, SSC outperforms VQ-based baselines (sDAC, MOC-RVQ, ESC-MVQ) and a separation-based BPG+LDPC codec, suggesting that learned digital semantic systems can close the gap to analog JSCC while remaining bit-compatible.
  • The system degrades gracefully under imperfect channel state information, with transmitter-side estimation errors having a much smaller effect than receiver-side errors, indicating that the abundant FA ports provide transmit diversity.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The offline preconfiguration of stream splitting could become an online adaptive routine: if the error ordering between port-index and symbol streams ever flips (e.g., at different N_s, M, or under correlated fading), the splitter can be recomputed live, turning UEP into dynamic UEP.
  • The same cross-layer alignment of source significance and physical-layer reliability could be applied to other hierarchical source coders, such as progressive JPEG or layered video codecs, when carried over IM links.
  • Future work might test whether the semantic significance ordering (step 1 more important than step 2) is robust across diverse source types; if some sources concentrate importance in later refinement steps, the splitter would need to learn the ordering per source class.
  • A testable extension would be to replace the fixed Monte Carlo decision with a learned importance-aware splitter trained end-to-end, letting the network discover the optimal mapping rather than relying on a hand-set rule.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes a spatial semantic communication (SSC) system that couples a SwinJSCC analog backbone with residual quantization (RQ) and fluid-antenna index modulation (FA-IM). A semantic-aware stream splitter assigns the early, semantically important RQ codeword indices to the FA-IM stream that is deemed more reliable, following the UEP principle. The authors report extensive simulations on Kodak24 claiming that SSC outperforms sDAC, MOC-RVQ, ESC-MVQ, BPG+LDPC, and the SSC w/o IM and SSC w/o SS ablations in PSNR/MS-SSIM over a range of SNRs and rates, while also reporting a large reduction in quantization search complexity.

Significance. If the claims hold, this is one of the first systematic integrations of semantic communications with index modulation, and the cross-layer UEP design is a plausible step beyond treating the physical layer as a homogeneous bit pipe. The paper's strengths include the clean three-module architecture, the three-stage training recipe, the inclusion of an open-source code repository, and a consistent set of ablations (w/o IM, w/o SS, w/o Stage 1) that all point in the expected direction. The complexity comparison with sDAC is also useful. However, the central UEP mechanism rests on a stream-reliability criterion that is not rigorously justified, and the training/evaluation channel mismatch is not validated. These issues make the universal claims in Section V-B stronger than what the evidence supports.

major comments (3)
  1. [§IV-B, Algorithm 1 and Fig. 5] The stream-selection rule compares raw per-slot error counts e_pi and e_sym. Since the port-index stream carries m1=log2 Ns bits/slot and the symbol stream carries m2=log2 M bits/slot, the quantity that matters for the downstream RQ bitstream is the per-bit error rate (or expected number of corrupted bits), not the raw number of erroneous slots. For example, with Ns=4, M=64, e_pi=4e3 and e_sym=6e3 would be read by Algorithm 1 as 'port indices more reliable', but if an index/symbol error flips one bit on average, the per-bit rates are 2e3/1e6 and 1e3/1e6, reversing the ordering. The authors should justify the metric, report per-bit BER curves, and re-examine the splitting decision for every configuration. If the ordering flips at the bit level, the proposed UEP mapping becomes anti-UEP and the claimed gain over 'SSC w/o SS' would reverse.
  2. [§IV-B, Fig. 5] The claim that the relative ordering e_pi vs e_sym is 'a deterministic structural property' invariant to SNR and fast fading is supported by only four parameter combinations (Ns in {2,4}, M in {16,64}) at Wt=2, Np=16, Nr=8. No analytical argument is given, and the paper does not vary Np, Wt, Lp, or consider correlated fading. Since Algorithm 1 is executed offline as a one-time preconfiguration and its entire benefit depends on the ordering, the paper's universal statement in Section V-B is not established. The authors should provide an analytical characterization of e_pi and e_sym, or at minimum a broader Monte Carlo study over the design space.
  3. [§III-B.1 and §V-A/V-B] The network is trained on a BSC with independent bit flips, while evaluation uses a finite-scattering FA-IM channel with correlated fading and joint ML detection. The paper asserts that randomly sampling the BSC bit-flip probability exposes the decoder to a diverse spectrum of error patterns, but no experiment validates that the BSC-trained model generalizes to the actual error patterns produced by FA-IM. A control experiment — for example, fine-tuning the decoder on real FA-IM error samples or training with a simulated FA-IM channel — would substantially strengthen the central claim. Without it, the reported gains may be partly an artifact of the training-time channel abstraction.
minor comments (5)
  1. [§V-A, Figs. 6–12] No error bars, confidence intervals, or training-seed information are reported. Given that some differences between curves are small and the paper claims superiority across the board, the authors should report mean±std over at least three seeds or state that curves are single runs.
  2. [§III-B.1] The mapping from SNR values to the BSC bit-flip probability set is not given. The paper should state the formula that produced the listed p values and clarify the channel averaging assumption used.
  3. [Algorithm 1] The notation e_pi and e_sym is informal. Please define precisely what is counted (e.g., number of erroneous port-index slots / total slots, and symbol slots / total slots) and specify the channel parameters (Lp, Np, Nr, Wt) used in the Monte Carlo preconfiguration.
  4. [§II-C] The claim that the digitization and splitting designs are 'mathematically decoupled from the physical layer antenna architecture' is plausible but unsupported. A short argument or a demonstration on another IM variant (e.g., spatial modulation) would make the claim credible.
  5. [Fig. 5 caption] The y-axis label in the text appears garbled ('0 1 2 3 4 5 6 7 10^5'). Please ensure the figure is legible and add a legend distinguishing port-index errors from symbol errors.

Circularity Check

0 steps flagged

No significant circularity: the stream-splitting design is informed by the paper's own Monte Carlo/observation experiments and is evaluated independently against held-out data and external benchmarks.

full rationale

The paper's derivation chain is not circular in the relevant sense. The proposed SSC system combines an existing JSCC backbone (SwinJSCC), a residual quantizer, and FA-IM. The semantic-aware stream splitting is configured using two empirical inputs: (i) the observation that earlier RQ steps carry more semantic importance (Fig. 4), and (ii) the paper's own Monte Carlo error counts for port-index and symbol streams (Fig. 5, Section IV-A). Algorithm 1 then assigns the semantically important early indices to the stream with lower measured error count. This is a design rule, not a fitted prediction: the relative error ordering is measured, not inferred from the later PSNR results, and the end-to-end PSNR/MS-SSIM evaluations (Figs. 6-12) are independent simulation outcomes on held-out Kodak24 images against baselines including sDAC, MOC-RVQ, ESC-MVQ, and BPG+LDPC. The self-citations in the paper (e.g., refs. [30], [31], [41], [42]) are background references for spatial modulation and FA-IM variants; no load-bearing uniqueness theorem or central result is imported from the authors' own prior work. The concerns raised by the skeptical view—raw per-slot error counts rather than per-bit BER, and limited validation of the SNR/fading invariance of the ordering—are substantive correctness and generalization risks, not circularity, because the paper's claims do not reduce by definition to those measurements. Accordingly, no specific circular step can be exhibited, so the circularity score is 0.

Axiom & Free-Parameter Ledger

6 free parameters · 5 axioms · 0 invented entities

No new physical entities (particles, forces, dimensions) are introduced; 'SSC' is an engineered system composed of existing JSCC, RQ, and FA-IM modules. The main load-bearing assumptions are the BSC training equivalence and the claimed deterministic reliability ordering of FA-IM streams.

free parameters (6)
  • Number of RQ steps Nq = 4
    Set to 4; directly controls bpp and the semantic hierarchy; no sweep shown for Nq (Fig. 4 only demonstrates with Nq=4).
  • Codebook size Ne = 16 (default; 64 in Fig. 9)
    Hand-chosen; together with Nq determines bpp through N'_e=N_e^{N_q}; not derived from theory.
  • Codeword dimension de = 4 for Ne=16; 8 recommended for Ne=64
    Tuned on the Kodak24 evaluation set (Sec V-E); affects quantization fidelity and training stability.
  • Commitment loss weight beta = 0.25
    Hand-picked training hyperparameter; no sensitivity study.
  • EMA decay gamma for codebook = 0.99
    Hand-picked; no sensitivity study.
  • BSC bit-flip probability set = seven values from 1.31e-1 to 2.27e-19
    Predefined set 'calculated across uniformly spaced SNR levels'; the mapping from SNR to p is not given, and this set shapes what noise levels the decoder learns.
axioms (5)
  • domain assumption Independent bit flips at a randomly sampled probability p are an adequate substitute for the actual correlated FA-IM channel during training.
    Used in Stages 1-2 (Sec III-B1, Eq 12); no validation that the trained decoder matches the correlated bit-error patterns of the physical channel.
  • domain assumption The relative error ordering of port-index vs symbol streams is a deterministic, SNR-invariant structural property of FA-IM.
    Section IV-B and Algorithm 1 rely on this to pre-configure the splitter; supported only by four Monte Carlo settings in Fig. 5.
  • domain assumption Perfect CSI at both transmitter and receiver for port-set selection and ML detection; channel estimation errors are modeled as additive scaled Gaussian noise.
    Sec II-A/II-B2; imperfect-CSI study covers epsilon up to 0.4 but the core results use perfect CSI.
  • domain assumption The mmWave finite-scattering planar-wave channel model (Eq 1) with Rayleigh path coefficients.
    Standard geometric channel; invoked in Sec II-A.
  • domain assumption RQ earlier steps are semantically more important, assumed from residual decomposition and the commitment loss.
    Section IV-A uses Fig. 4 (single image / Nq=4) to establish this; the splitter generalizes this to all images and Nq.

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Spatial Semantic Communication: When Semantic Transmission Meets Index Modulation." pith.science (2026). https://pith.science/paper/O4BCAG5S

@misc{pith2026260719934,
  author       = {Pith},
  title        = {Pith review of: Spatial Semantic Communication: When Semantic Transmission Meets Index Modulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/O4BCAG5S}},
  note         = {Machine review of arXiv:2607.19934}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Current digital semantic communication systems have primarily focused on maintaining compatibility with conventional constellation-based modulation. In contrast, index modulation (IM) represents a more spectrally and energy-efficient alternative by exploiting additional dimensions for information conveyance. Recognizing this potential, this paper bridges the gap between IM and semantic communications by proposing a novel spatial semantic communication (SSC) system leveraging cutting-edge fluid antenna-IM (FA-IM) technology. Compatible with existing joint source-channel coding (JSCC) architectures, the proposed SSC system employs the residual quantization (RQ) approach to discretize analog semantic features for subsequent digital IM transmission. Notably, the proposed SSC system synergizes RQ and IM via a semantic-aware stream splitting scheme, which ensures that critical semantic information undergoes less severe channel fading, thereby further optimizing semantic transmission performance. Simulation results validate that the proposed SSC system effectively integrates the high fidelity of RQ, the reliability of semantic-aware splitting, and the spatial efficiency of FA-IM, thereby providing a robust solution for future digital semantic transmission. The open source code is available at: https://github.com/gxh1106/SSC.

Figures

Figures reproduced from arXiv: 2607.19934 by Cixiao Zhang, Dazhi He, Hanjiang Hong, Wenjun Zhang, Xinghao Guo, Yin Xu, Yiyan Wu, Zhiyong Chen.

Figure 1
Figure 1. Figure 1: Block diagram of the proposed SSC system. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Block diagrams of the proposed residual quantizer and dequantizer in the SSC system. (a) Overview of the residual [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The multi-stage training processes for SSC. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Error counts for port indices and symbols of the FA [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Performance comparison of PSNR and MS-SSIM between the proposed SSC system and benchmarks across various [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Performance comparisons between the proposed SSC [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: PSNR and MS-SSIM performance of the proposed SSC [PITH_FULL_IMAGE:figures/full_fig_p012_8.png] view at source ↗
Figure 10
Figure 10. Figure 10: PSNR and MS-SSIM performance of the SSC system [PITH_FULL_IMAGE:figures/full_fig_p013_10.png] view at source ↗
Figure 12
Figure 12. Figure 12: MS-SSIM performance comparison of the proposed [PITH_FULL_IMAGE:figures/full_fig_p014_12.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

51 extracted references

  1. [1]

    5G-advanced toward 6G: Past, present, and future,

    W. Chen, X. Lin, J. Lee, A. Toskala, S. Sun, C. F. Chiasserini, and L. Liu, “5G-advanced toward 6G: Past, present, and future,”IEEE J. Sel. Areas Commun., vol. 41, no. 6, pp. 1592–1619, 2023

    2023

  2. [2]

    Semantic communications for future internet: Fundamentals, applications, and challenges,

    W. Yang, H. Du, Z. Q. Liew, W. Y . B. Lim, Z. Xiong, D. Niyato, X. Chi, X. Shen, and C. Miao, “Semantic communications for future internet: Fundamentals, applications, and challenges,”IEEE Commun. Surveys Tuts., vol. 25, no. 1, pp. 213–250, 2023

    2023

  3. [3]

    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 Process., vol. 69, pp. 2663–2675, 2021

    2021

  4. [4]

    SwinJSCC: Taming swin transformer for deep joint source-channel coding,

    K. Yang, S. Wang, J. Dai, X. Qin, K. Niu, and P. Zhang, “SwinJSCC: Taming swin transformer for deep joint source-channel coding,”IEEE Trans. Cogn. Commun. Netw., vol. 11, no. 1, pp. 90–104, 2025

    2025

  5. [5]

    CDDM: Channel denoising diffusion models for wireless semantic communications,

    T. Wu, Z. Chen, D. He, L. Qian, Y . Xu, M. Tao, and W. Zhang, “CDDM: Channel denoising diffusion models for wireless semantic communications,”IEEE Trans. Wireless Commun., vol. 23, no. 9, pp. 11 168–11 183, 2024

    2024

  6. [6]

    Semantic communication systems for speech transmission,

    Z. Weng and Z. Qin, “Semantic communication systems for speech transmission,”IEEE J. Sel. Areas Commun., vol. 39, no. 8, pp. 2434– 2444, 2021

    2021

  7. [7]

    Wireless deep video semantic transmission,

    S. Wang, J. Dai, Z. Liang, K. Niu, Z. Si, C. Dong, X. Qin, and P. Zhang, “Wireless deep video semantic transmission,”IEEE J. Sel. Areas Commun., vol. 41, no. 1, pp. 214–229, 2023

    2023

  8. [8]

    Large AI model empowered multimodal semantic communications,

    F. Jiang, L. Dong, Y . Peng, K. Wang, K. Yang, C. Pan, and X. You, “Large AI model empowered multimodal semantic communications,” IEEE Commun. Mag., vol. 63, no. 1, pp. 76–82, 2025

    2025

  9. [9]

    Toward compatible semantic communication: A perspective on digital coding and modula- tion,

    G. Zhang, K. Zhou, Y . Cai, Q. Hu, and G. Yu, “Toward compatible semantic communication: A perspective on digital coding and modula- tion,”IEEE Commun. Mag., pp. 1–7, 2025

    2025

  10. [10]

    DeepJSCC- Q: Constellation constrained deep joint source-channel coding,

    T.-Y . Tung, D. B. Kurka, M. Jankowski, and D. G ¨und¨uz, “DeepJSCC- Q: Constellation constrained deep joint source-channel coding,”IEEE J. Sel. Areas Commun., vol. 3, no. 4, pp. 720–731, 2022

    2022

  11. [11]

    A lite distributed semantic communication system for internet of things,

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

    2021

  12. [12]

    Semantic-oriented modulation for wireless communication,

    Y . Wang, X. Zhang, Y . Sun, Q. Cui, and X. Tao, “Semantic-oriented modulation for wireless communication,”IEEE Internet Things J., vol. 12, no. 14, pp. 28 259–28 273, 2025

    2025

  13. [13]

    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., vol. 72, no. 9, pp. 5626–5640, 2024

    2024

  14. [14]

    From analog to digital: Multi-order digital joint coding-modulation for semantic communica- tion,

    G. Zhang, P. Yang, Y . Cai, Q. Hu, and G. Yu, “From analog to digital: Multi-order digital joint coding-modulation for semantic communica- tion,”IEEE Trans. Commun., vol. 73, no. 6, pp. 4257–4271, 2025

    2025

  15. [15]

    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. Cogn. Commun. Netw., vol. 11, no. 1, pp. 75–89, 2025

    2025

  16. [16]

    Alternate learning- based SNR-adaptive sparse semantic visual transmission,

    S. Tong, X. Yu, R. Li, K. Lu, Z. Zhao, and H. Zhang, “Alternate learning- based SNR-adaptive sparse semantic visual transmission,”IEEE Trans. Wireless Commun., vol. 24, no. 2, pp. 1737–1752, 2025

    2025

  17. [17]

    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,”IEEE Trans. Wireless Commun., vol. 25, pp. 4452–4466, 2026

    2026

  18. [18]

    Fully learnable multi-rate quantization for digital semantic communication systems,

    M. Kim and D. J. Ji, “Fully learnable multi-rate quantization for digital semantic communication systems,”IEEE Wireless Commun. Lett., vol. 14, no. 9, pp. 2848–2851, 2025

    2025

  19. [19]

    Neural discrete representation learning,

    A. Van Den Oord, O. Vinyalset al., “Neural discrete representation learning,”Proc. Adv. Neural Inf. Process. Syst. (NIPS), vol. 30, 2017. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 20XX 16

    2017

  20. [20]

    Robust information bottleneck for task-oriented communication with digital modulation,

    S. Xie, S. Ma, M. Ding, Y . Shi, M. Tang, and Y . Wu, “Robust information bottleneck for task-oriented communication with digital modulation,” IEEE J. Sel. Areas Commun., vol. 41, no. 8, pp. 2577–2591, 2023

    2023

  21. [21]

    Robust semantic communications with masked VQ-V AE enabled codebook,

    Q. Hu, G. Zhang, Z. Qin, Y . Cai, G. Yu, and G. Y . Li, “Robust semantic communications with masked VQ-V AE enabled codebook,”IEEE Trans. Wireless Commun., vol. 22, no. 12, pp. 8707–8722, 2023

    2023

  22. [22]

    Vector quantized se- mantic communication system,

    Q. Fu, H. Xie, Z. Qin, G. Slabaugh, and X. Tao, “Vector quantized se- mantic communication system,”IEEE Wireless Commun. Lett., vol. 12, no. 6, pp. 982–986, 2023

    2023

  23. [23]

    sDAC—semantic digital analog converter for semantic communications,

    Z. Bao, H. Liange, X. Liu, C. Li, C. Dong, X. Xu, C. Guo, H. Chen, and P. Zhang, “sDAC—semantic digital analog converter for semantic communications,”IEEE Trans. Commun., pp. 1–1, 2025

    2025

  24. [24]

    ESC-MVQ: End-to- end semantic communication with multi-codebook vector quantization,

    J. Shin, Y . Oh, J. Park, J. Park, and Y .-S. Jeon, “ESC-MVQ: End-to- end semantic communication with multi-codebook vector quantization,” IEEE Trans. Wireless Commun., vol. 25, pp. 3785–3800, 2026

    2026

  25. [25]

    Autoregressive image generation using residual quantization,

    D. Lee, C. Kim, S. Kim, M. Cho, and W.-S. Han, “Autoregressive image generation using residual quantization,” inProc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2022, pp. 11 523–11 532

    2022

  26. [26]

    Taming learnable codebook design and modulation for digital semantic image communication,

    Y . Zhou, Y . Sun, H. Du, G. Chen, X. Xu, H. Chen, P. Zhang, and S. Cui, “Taming learnable codebook design and modulation for digital semantic image communication,”IEEE Internet Things J., vol. 13, no. 1, pp. 473–486, 2026

    2026

  27. [27]

    Low-bitrate high-quality digital semantic communication based on RVQGAN,

    X. Chen, J. Wang, J. Huang, M. Zeng, Z. Zheng, and Z. Fei, “Low-bitrate high-quality digital semantic communication based on RVQGAN,”IEEE Internet Things J., vol. 12, no. 10, pp. 13 525–13 537, 2025

    2025

  28. [28]

    Index modulation techniques for next-generation wireless networks,

    E. Basar, M. Wen, R. Mesleh, M. Di Renzo, Y . Xiao, and H. Haas, “Index modulation techniques for next-generation wireless networks,” IEEE Access, vol. 5, pp. 16 693–16 746, 2017

    2017

  29. [29]

    Spatial modulation,

    R. Y . Mesleh, H. Haas, S. Sinanovic, C. W. Ahn, and S. Yun, “Spatial modulation,”IEEE Trans. V eh. Technol., vol. 57, no. 4, pp. 2228–2241, 2008

    2008

  30. [30]

    Design of capacity-approaching constellation and pre-scaling for spatial modulation,

    X. Guo, Y . Xu, H. Hong, S. Peng, D. He, W. Zhang, and Y .-Y . We, “Design of capacity-approaching constellation and pre-scaling for spatial modulation,” inProc. IEEE V eh. Technol. Conf. (Spring), 2024, pp. 1–5

    2024

  31. [31]

    RIS-aided receive generalized spatial modulation design with reflecting modulation,

    X. Guo, Y . Xu, H. Hong, D. Mi, R. Liu, D. He, W. Zhang, and Y .-Y . Wu, “RIS-aided receive generalized spatial modulation design with reflecting modulation,” inProc. IEEE Global Commun. Conf. (GLOBECOM), 2024, pp. 2563–2568

    2024

  32. [32]

    Orthogonal frequency division multiplexing with index modulation,

    E. Bas ¸ar, ¨U. Ayg ¨ol¨u, E. Panayırcı, and H. V . Poor, “Orthogonal frequency division multiplexing with index modulation,”IEEE Trans. Signal Process., vol. 61, no. 22, pp. 5536–5549, 2013

    2013

  33. [33]

    Fluid antenna systems,

    K.-K. Wong, A. Shojaeifard, K.-F. Tong, and Y . Zhang, “Fluid antenna systems,”IEEE Trans. Wireless Commun., vol. 20, no. 3, pp. 1950–1962, 2021

    1950

  34. [34]

    A tutorial on movable antennas for wireless networks,

    L. Zhu, W. Ma, W. Mei, Y . Zeng, Q. Wu, B. Ning, Z. Xiao, X. Shao, J. Zhang, and R. Zhang, “A tutorial on movable antennas for wireless networks,”IEEE Commun. Surveys Tuts., vol. 28, pp. 3002–3054, 2026

    2026

  35. [35]

    A contemporary survey on fluid antenna systems: Fundamentals and networking perspectives,

    H. Hong, K.-K. Wong, C.-B. Chae, H. Xu, X. Guo, F. R. Ghadi, Y . Chen, Y . Xu, B. Liu, K.-F. Tong, and Y . Zhang, “A contemporary survey on fluid antenna systems: Fundamentals and networking perspectives,” IEEE Trans. Netw. Sci. Eng., pp. 1–24, 2025

    2025

  36. [36]

    Advanced channel coding designs for index-modulated fluid antenna systems,

    E. Faddoul, G. M. Kraidy, C. Psomas, and I. Krikidis, “Advanced channel coding designs for index-modulated fluid antenna systems,” IEEE Trans. Commun., vol. 73, no. 2, pp. 726–739, 2025

    2025

  37. [37]

    Fluid antenna index modulation communications,

    Y . Chen and T. Xu, “Fluid antenna index modulation communications,” IEEE Wireless Commun. Lett., vol. 13, no. 4, pp. 1203–1207, 2024

    2024

  38. [38]

    Position index modulation for fluid antenna system,

    H. Yang, H. Xu, K.-K. Wong, C.-B. Chae, R. Murch, S. Jin, and Y . Zhang, “Position index modulation for fluid antenna system,”IEEE Trans. Wireless Commun., vol. 23, no. 11, pp. 16 773–16 787, 2024

    2024

  39. [39]

    Fluid antenna empowered index modulation for RIS-aided mmwave transmis- sions,

    J. Zhu, Q. Luo, G. Chen, P. Xiao, Y . Xiao, and K.-K. Wong, “Fluid antenna empowered index modulation for RIS-aided mmwave transmis- sions,”IEEE Trans. Wireless Commun., pp. 1–1, 2024

    2024

  40. [40]

    Index modulation for fluid antenna-assisted MIMO communications: System design and performance analysis,

    J. Zhu, G. Chen, P. Gao, P. Xiao, Z. Lin, and A. Quddus, “Index modulation for fluid antenna-assisted MIMO communications: System design and performance analysis,”IEEE Trans. Wireless Commun., pp. 1–1, 2024

    2024

  41. [41]

    Fluid antenna grouping index modulation design for MIMO systems,

    X. Guo, Y . Xu, D. He, C. Zhang, W. Zhang, and Y . Wu, “Fluid antenna grouping index modulation design for MIMO systems,” inProc. IEEE Wireless Commun. Netw. Conf. (WCNC), 2025, pp. 1–6

    2025

  42. [42]

    Fluid antenna index modulation for MIMO systems: Robust transmission and low-complexity detection,

    X. Guo, Y . Xu, D. He, C. Zhang, H. Hong, K.-K. Wong, W. Zhang, and Y . Wu, “Fluid antenna index modulation for MIMO systems: Robust transmission and low-complexity detection,”IEEE Trans. Commun., pp. 1–1, 2025

    2025

  43. [43]

    Modeling and performance analysis for movable antenna enabled wireless communications,

    L. Zhu, W. Ma, and R. Zhang, “Modeling and performance analysis for movable antenna enabled wireless communications,”IEEE Trans. Wireless Commun., vol. 23, no. 6, pp. 6234–6250, 2024

    2024

  44. [44]

    An information-theoretic characterization of MIMO-FAS: Optimization, diversity-multiplexing tradeoff and q-outage capacity,

    W. K. New, K.-K. Wong, H. Xu, K.-F. Tong, and C.-B. Chae, “An information-theoretic characterization of MIMO-FAS: Optimization, diversity-multiplexing tradeoff and q-outage capacity,”IEEE Trans. Wireless Commun., vol. 23, no. 6, pp. 5541–5556, 2024

    2024

  45. [45]

    A novel pixel-based reconfigurable antenna applied in fluid antenna systems with high switching speed,

    J. Zhang, J. Rao, Z. Li, Z. Ming, C.-Y . Chiu, K.-K. Wong, K.-F. Tong, and R. Murch, “A novel pixel-based reconfigurable antenna applied in fluid antenna systems with high switching speed,”IEEE Open J. Antennas Propag., vol. 6, no. 1, pp. 212–228, 2025

    2025

  46. [46]

    Offset spatial modulation and offset space shift keying: Efficient designs for single-rf mimo systems,

    S. Fang, K. Zheng, Y . Xiao, Y . Yang, X. Zeng, and M. Xiao, “Offset spatial modulation and offset space shift keying: Efficient designs for single-rf mimo systems,”IEEE Trans. Commun., vol. 67, no. 8, pp. 5434–5444, 2019

    2019

  47. [47]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, S. Lin, and B. Guo, “Swin transformer: Hierarchical vision transformer using shifted windows,” inProc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2021, pp. 10 012–10 022

    2021

  48. [48]

    Ntire 2017 challenge on single image super-resolution: Dataset and study,

    E. Agustsson and R. Timofte, “Ntire 2017 challenge on single image super-resolution: Dataset and study,” inProc. IEEE Conf. Comput. Vis. Pattern Recognit. Workshops, 2017, pp. 126–135

    2017

  49. [49]

    Kodak lossless true color image suite

    “Kodak lossless true color image suite.” [Online]. Available: http: //r0k.us/graphics/kodak/

  50. [50]

    BPG image format,

    F. Bellard, “BPG image format,” 2018. [Online]. Available: https: //bellard.org/bpg/

    2018

  51. [51]

    An introduction to ldpc codes,

    W. E. Ryanet al., “An introduction to ldpc codes,”CRC Handbook for Coding and Signal Processing for Recording Systems, vol. 5, no. 2, pp. 1–23, 2004

    2004

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.