Pith. sign in

REVIEW 2 major objections 5 minor 32 references

Pilot-free neural detector beats CSI-based equalization for THz semantic links

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 ↗

T0 review · glm-5.2

2026-07-09 10:19 UTC pith:IMCBAIU5

load-bearing objection Pilot-free neural detector for semantic communication over THz channels — a real gap, but the CSI-free claim is narrower than it sounds. the 2 major comments →

arxiv 2607.07455 v1 pith:IMCBAIU5 submitted 2026-07-08 eess.SP

Semantic Communications in the THz Band

Fatima Ismail , Hadi Sarieddeen , Jihad Fahs This is my paper
classification eess.SP
keywords semantic communicationterahertz communicationsneural detectorchannel state informationdeep learningequalizationDeepSCTHz MIMO
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.

This paper argues that a deep learning-based semantic communication system (DeepSC), trained only on simple additive white Gaussian noise (AWGN), can operate over realistic terahertz (THz) fading channels without channel-specific retraining. The key insight is a post-equalization argument: under perfect minimum mean square error (MMSE) equalization, a fading channel reduces to an equivalent AWGN channel with rescaled noise, so a decoder trained on AWGN should generalize. To make the system fully free of channel state information (CSI), the authors propose a lightweight Transformer-based neural detector (ND) that replaces conventional equalization. The ND is trained on simulated THz channel data to recover clean transmitted channel-symbol sequences directly from corrupted observations, without pilots. At 0.3 THz, the pilot-free ND matches perfect-CSI MMSE and outperforms pilot-based and noisy-CSI MMSE, while the overall DeepSC system achieves more than 50 percentage-point higher BLEU-1 than a throughput-matched LDPC-coded 64-QAM baseline over 0-12 dB SNR. The ND also remains robust to carrier-frequency offsets of up to 50 MHz without retraining.

Core claim

The central discovery is that AWGN-only training of a semantic communication system suffices for THz fading channels when paired with a receiver that compensates for the channel, and that a compact pilot-free neural detector can serve as that compensator while outperforming conventional MMSE equalization that has access to CSI. The paper also shows that semantic performance is more sensitive to the post-equalization noise distribution than to the fading profile itself, which is why AWGN training transfers across block fading, fast fading, and molecular absorption scenarios.

What carries the argument

The mechanism is a two-stage training procedure: Stage 1 trains the DeepSC encoder-decoder end-to-end over an AWGN channel and freezes it; Stage 2 trains a Transformer-based neural detector (100,802 parameters, two encoder layers, eight attention heads) to map THz-corrupted received signals back to the clean channel-symbol distribution expected by the frozen decoder. The ND uses self-attention to exploit sentence-level dependencies among continuous channel symbols, distinguishing it from constellation-level detectors that assume independent discrete symbols. The THz channel is generated via the TeraMIMO simulator at 0.3 THz with indoor Saleh-Valenzuela parameters and 4096x4096 antenna arrays

Load-bearing premise

The argument for AWGN training rests on the claim that after perfect-CSI MMSE equalization, the residual noise remains white and Gaussian, so the decoder sees an AWGN-like channel. The ND trained on TeraMIMO-simulated indoor THz channels is then assumed to generalize to real THz environments, but this is only tested for one simulator configuration, one indoor scenario, and one antenna array size.

What would settle it

If the post-equalization residual noise in real THz channels is significantly non-Gaussian or colored due to effects not captured by the TeraMIMO simulator, the AWGN-trained DeepSC decoder would degrade, and the ND trained on simulated data would fail to compensate, causing the pilot-free system to underperform CSI-based equalization.

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

If this is right

  • If the post-equalization noise argument holds broadly, semantic communication systems could be deployed across diverse fading environments without costly channel-specific retraining, reducing the engineering burden for new frequency bands.
  • Pilot-free neural detection could eliminate the pilot overhead and CSI estimation pipelines that are especially costly at THz frequencies, simplifying transceiver design for future 6G systems.
  • The robustness to frequency offsets suggests that the ND learns channel-symbol-level structure rather than memorizing exact channel parameters, which could make it adaptable to hardware impairments like oscillator drift.
  • The finding that semantic performance depends more on noise distribution than fading profile implies that future semantic coding efforts should prioritize robustness to non-Gaussian post-equalization residual noise.

Where Pith is reading between the lines

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

  • If the ND implicitly learns the THz channel statistics during Stage 2 training, then its CSI-free claim is contingent on the real channel matching the TeraMIMO simulator's model; deployment in environments not represented in training data could cause silent failures without an explicit error signal.
  • The two-stage approach separates semantic coding from channel compensation, which suggests the ND could be swapped or retrained independently for different THz scenarios while reusing the same AWGN-trained DeepSC backbone, though this modularity is not tested in the paper.
  • The post-equalization noise-whiteness argument may break down under spatially correlated MIMO channels or near-field propagation effects specific to ultra-massive antenna arrays, where residual interference after equalization may not be white or Gaussian.

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

2 major / 5 minor

Summary. This manuscript investigates deep learning-based semantic communication (DeepSC) over terahertz (THz) channels. The authors make three main claims: (1) DeepSC models trained solely under AWGN generalize well to tested THz block- and fast-fading channels when receiver-side compensation is applied, motivated by a post-equalization noise argument; (2) a proposed lightweight, pilot-free neural detector (ND) outperforms MMSE equalization with both perfect and imperfect CSI; and (3) DeepSC is more robust to CSI errors than a throughput-matched traditional coded baseline. The system is evaluated using the TeraMIMO channel simulator at 0.3 THz, and the ND is trained over THz channel realizations from this simulator while the DeepSC encoder-decoder is frozen after AWGN pre-training.

Significance. The paper presents a falsifiable and well-structured experimental methodology. The throughput-matched baseline comparison (4.89 vs 4.8 bits/symbol) is a notable strength, ensuring a fair comparison between semantic and traditional coded systems. The post-equalization noise argument in §3.2.1 provides a parameter-free motivation for AWGN training. The inclusion of 30 independent test passes with reported confidence intervals (max std 1.3e-2 BLEU-1) and receiver ablations (Conv-only, MLP-128) adds rigor to the empirical claims. The proposed pilot-free ND for continuous semantic channel symbols addresses a practical gap in THz communications where CSI acquisition is costly.

major comments (2)
  1. §3.2.1, Eq. (2) context: The post-equalization noise argument assumes that under perfect-CSI MMSE, the fading channel y=hx+n becomes x_hat = alpha*x + n_tilde where n_tilde remains white Gaussian. This motivates AWGN training of the DeepSC encoder-decoder. However, the argument is used to motivate the entire CSI-free system, including the ND. The ND (Stage 2, Algorithm 1) is explicitly trained on THz-corrupted observations from the TeraMIMO simulator with specific indoor Saleh-Valenzuela parameters (T=298.15 K, p=1 atm, 4096x4096 antennas). The 'CSI-free' label means 'no explicit CSI at inference,' not 'channel-agnostic.' The manuscript should clarify this distinction explicitly, as the ND's advantage is structurally dependent on having learned the TeraMIMO channel distribution during Stage 2 training. Without this clarification, the 'CSI-free' claim risks being overstated.
  2. §4.3, Fig. 3(b): The claim that the ND 'outperforms MMSE with perfect CSI' is specific to the TeraMIMO channel distribution used in training and testing. The ND is trained and evaluated on independent realizations from the same simulator with the same parameters. Cross-environment generalization (e.g., outdoor scenarios, different antenna configurations, non-ideal beamforming) is not tested. The ±50 MHz frequency offset test is a very small perturbation within the same model family and does not address distribution shift across different THz environments. The authors should explicitly state that the perfect-CSI MMSE outperformance is specific to the tested TeraMIMO indoor scenario and acknowledge that the ND may fail silently under distribution shift since it has no error signal to detect it.
minor comments (5)
  1. §3.1: The ND architecture description mentions d_model=64, 8 attention heads, and FFN dimension 256, but the number of trainable parameters (100,802) seems surprisingly small for two Transformer encoder layers with these dimensions. A brief breakdown of parameter count would help readers verify this.
  2. Fig. 2(a,b): The legend distinguishes 'AWGN-trained' (dashed) from 'same-channel trained' (solid) models, but the notation 'AWGN (AWGN)', 'Rayleigh (AWGN)', etc., is initially confusing. Consider relabeling for clarity, e.g., 'Trained: AWGN, Tested: Rayleigh'.
  3. §4.2: The latency analysis reports processing times on a V100 GPU, but the batch size and sequence length assumptions for the timing comparison could be stated more explicitly to ensure the comparison between MMSE and ND is fair.
  4. §4.3, Eq. (3): The high-SNR approximation is used to analyze CSI mismatch, but the tested SNR range includes 0-12 dB where the approximation may not hold well. A brief comment on the validity of this approximation at low SNR would strengthen the analysis.
  5. References [14-16] on diffusion-based denoisers are cited but not deeply compared to the proposed ND. A sentence clarifying why diffusion-based approaches are unsuitable for pilot-free text semantic equalization (beyond targeting image transmission) would improve the related work.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive feedback. Both major comments concern the scope and generalizability of the 'CSI-free' label and the ND's outperformance of perfect-CSI MMSE. We agree that the manuscript should explicitly clarify that the ND is trained on the TeraMIMO channel distribution and that its advantages are specific to the tested indoor scenario. We will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: §3.2.1, Eq. (2) context: The post-equalization noise argument assumes that under perfect-CSI MMSE, the fading channel y=hx+n becomes x_hat = alpha*x + n_tilde where n_tilde remains white Gaussian. This motivates AWGN training of the DeepSC encoder-decoder. However, the argument is used to motivate the entire CSI-free system, including the ND. The ND (Stage 2, Algorithm 1) is explicitly trained on THz-corrupted observations from the TeraMIMO simulator with specific indoor Saleh-Valenzuela parameters. The 'CSI-free' label means 'no explicit CSI at inference,' not 'channel-agnostic.' The manuscript should clarify this distinction explicitly, as the ND's advantage is structurally dependent on having learned the TeraMIMO channel distribution during Stage 2 training. Without this clarification, the 'CSI-free' claim risks being overstated.

    Authors: The referee is correct. The post-equalization noise argument in §3.2.1 motivates AWGN pre-training of the DeepSC encoder-decoder: under perfect-CSI MMSE, the effective channel seen by the decoder is approximately AWGN-like, so an AWGN-trained decoder should generalize. This argument does not, however, motivate the ND itself. The ND is trained on THz-corrupted observations drawn from the TeraMIMO simulator (Stage 2, Algorithm 1) and learns to invert the specific channel distribution it was trained on. The 'CSI-free' label accurately describes the inference-time property—no pilots, no explicit channel estimates, no matrix inversions—but it does not mean the ND is channel-agnostic. Its advantage over MMSE is structurally dependent on having learned the TeraMIMO indoor channel statistics during training. We will revise the manuscript to state this distinction explicitly in §3.2.1 and §3.2.2, and will add a clarifying note to the abstract and contributions list so that 'CSI-free' is understood as 'no explicit CSI at inference' rather than 'channel-distribution-independent.' revision: yes

  2. Referee: §4.3, Fig. 3(b): The claim that the ND 'outperforms MMSE with perfect CSI' is specific to the TeraMIMO channel distribution used in training and testing. The ND is trained and evaluated on independent realizations from the same simulator with the same parameters. Cross-environment generalization (e.g., outdoor scenarios, different antenna configurations, non-ideal beamforming) is not tested. The ±50 MHz frequency offset test is a very small perturbation within the same model family and does not address distribution shift across different THz environments. The authors should explicitly state that the perfect-CSI MMSE outperformance is specific to the tested TeraMIMO indoor scenario and acknowledge that the ND may fail silently under distribution shift since it has no error signal to detect it.

    Authors: The referee's assessment is accurate. The ND's outperformance of perfect-CSI MMSE in Fig. 3(b) is observed on independent realizations from the same TeraMIMO indoor scenario with the same Saleh-Valenzuela parameters used in training. We did not test cross-environment generalization (outdoor, different antenna configurations, non-ideal beamforming), and the ±50 MHz frequency offset is indeed a small perturbation within the same model family rather than a cross-distribution test. We agree that the manuscript should explicitly scope the claim to the tested TeraMIMO indoor scenario. We also acknowledge the important point about silent failure under distribution shift: since the ND has no explicit error or confidence signal, it may degrade without detection if deployed in an environment whose statistics differ from training. We will add a statement to §4.3 and the Conclusion explicitly noting that the perfect-CSI MMSE outperformance is specific to the tested TeraMIMO indoor scenario, that cross-environment generalization is untested, and that silent failure under distribution shift is a known limitation of the current design that warrants future work (e.g., distribution-shift detection or online adaptation). revision: yes

Circularity Check

0 steps flagged

No significant circularity; one minor self-citation that is not load-bearing.

full rationale

The paper's central claims are supported by independent experimental evidence rather than circular definitions. The AWGN-training generalization claim (Section 4.1, Fig. 2) is validated by comparing AWGN-trained models against channel-specifically trained models on Rayleigh, Rician, and THz channels — the AWGN-trained models match or exceed the latter, which is an empirical result not forced by construction. The post-equalization noise argument (Section 3.2.1) is a parameter-free derivation motivating AWGN training, not a fitted input renamed as prediction. The ND's superiority over MMSE (Fig. 3) is an empirical comparison against an external baseline (Sionna-based LDPC-coded 64-QAM with MMSE), not a restatement of a fit. The ND is trained on TeraMIMO channel realizations and tested on independent realizations from the same simulator; while this limits cross-environment generalization claims (a correctness risk, not circularity), the training and testing are not the same data, and the comparison against MMSE with perfect CSI is a genuine external benchmark. The only self-citation is to TeraMIMO [4], co-authored by Hadi Sarieddeen, but this is a standard channel simulator used as an experimental tool, not a load-bearing theorem invoked to forbid alternatives or force the result. The paper's derivation chain is self-contained against external benchmarks, and no 'prediction' reduces to its inputs by construction.

Axiom & Free-Parameter Ledger

14 free parameters · 5 axioms · 1 invented entities

The free parameters are mostly standard neural network hyperparameters and DeepSC framework defaults. The key axioms are the post-equalization noise whiteness assumption and the TeraMIMO simulator fidelity. The only invented entity is the ND itself, which has independent evaluation evidence but is circumscribed by the simulation environment.

free parameters (14)
  • lambda (MI loss weight) = 0.0009
    Set following DeepSC [7]; controls the trade-off between semantic distortion and mutual information in the training loss (Eq. 2).
  • d_model (ND) = 64
    Chosen for the ND's internal representation dimension; affects capacity and complexity.
  • Number of Transformer layers (ND) = 2
    Architecture choice for the neural detector; balances complexity and performance.
  • Number of attention heads = 8
    Standard Transformer hyperparameter for the ND.
  • FFN dimension (ND) = 256
    Feed-forward network dimension in the ND Transformer layers.
  • Convolution kernel size (ND) = 3
    Kernel size for the 1D convolutions in the ND; chosen for local symbol-level distortion capture.
  • Training SNR range (Stage 1) = 5-10 dB
    SNR range for AWGN pretraining of DeepSC; affects the model's noise robustness profile.
  • Sentence length L = 30
    Fixed sentence length for training and evaluation.
  • N (complex symbols per word) = 8
    Number of complex channel symbols per word in DeepSC; determines the effective source rate.
  • Training epochs (Stage 1) = ~75
    Number of epochs for AWGN pretraining.
  • Training epochs (Stage 2) = 80
    Number of epochs for ND training over THz channels.
  • Initial learning rate = 1e-4
    Adam optimizer initial learning rate for Stage 2 training.
  • Batch size = 64
    Training and inference batch size in sentences.
  • Pilot overhead (baseline) = 10%
    Fraction of symbols used as pilots for LS-based CSI estimation in the MMSE baseline comparison.
axioms (5)
  • domain assumption Post-equalization noise remains white Gaussian under perfect-CSI MMSE
    Section 3.2.1: the argument that y=hx+n becomes x_hat=alpha*x+n_tilde where n_tilde is white Gaussian. This holds for SISO scalar channels but may not hold for frequency-selective or MIMO channels without further justification.
  • domain assumption TeraMIMO simulator accurately models indoor THz channels
    Section 2: all THz channel realizations are generated using TeraMIMO with default indoor Saleh-Valenzuela parameters. The validity of all results depends on this simulator's fidelity.
  • domain assumption 4096x4096 antenna elements are feasible for THz systems
    Section 4: the simulation uses 4096x4096 antenna elements, cited as enabled by compact THz array technologies [26]. This is an extreme configuration that may not reflect near-term practical deployments.
  • domain assumption BLEU-1 and SBERT adequately measure semantic communication quality
    Section 4: all performance comparisons use BLEU-1 and SBERT scores. These metrics may not fully capture semantic fidelity, especially for partial sentence recovery.
  • domain assumption The scalar beamformed channel h adequately represents the THz MIMO channel
    Section 2: the received signal model y=hx+n uses a scalar h obtained by projecting the MIMO channel onto beamforming vectors. This SISO-like model may miss multi-stream or spatial multiplexing effects.
invented entities (1)
  • Pilot-free neural detector (ND) for continuous semantic channel symbols independent evidence
    purpose: Replaces MMSE equalization by learning to map THz-corrupted observations to clean channel-symbol sequences without explicit CSI
    The ND is evaluated against MMSE with perfect CSI, pilot-based CSI, and noisy CSI, and is tested under frequency offsets. It outperforms baselines, providing falsifiable evidence. However, it is trained and tested on the same simulator (TeraMIMO), so true independence from the channel model is not demonstrated.

pith-pipeline@v1.1.0-glm · 11955 in / 3679 out tokens · 514551 ms · 2026-07-09T10:19:27.504191+00:00 · methodology

0 comments
read the original abstract

Semantic and terahertz (THz)-band communications are algorithmic and spectral enablers of future wireless networks. This work investigates deep learning-based semantic communication (DeepSC) over THz channels. We show that DeepSC models trained solely under additive white Gaussian noise generalize well to the tested THz block- and fast-fading channels when receiver-side compensation is applied. To enable fully data-driven reception, we propose a lightweight neural detector that does not require channel state information (CSI). At 0.3 THz, DeepSC outperforms a throughput-matched traditional coded communication system baseline over 0-12 dB signal-to-noise ratio (SNR), achieving more than 50 percentage-point higher Bilingual Evaluation Understudy unigram (BLEU-1) score. The proposed pilot-free detector outperforms minimum mean square error (MMSE) equalization with both perfect and imperfect CSI and remains robust to frequency offsets up to 50 MHz, highlighting the resilience of semantic communication to THz channel impairments.

Figures

Figures reproduced from arXiv: 2607.07455 by Fatima Ismail, Hadi Sarieddeen, Jihad Fahs.

Figure 1
Figure 1. Figure 1: The DeepSC framework augmented with the proposed ND in a THz MIMO system. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: (a) BLEU and (b) SBERT scores comparing AWGN-trained models (dashed lines) [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: BLEU-1 over THz channels: (a) DeepSC with proposed/ablated NDs and MMSE, [PITH_FULL_IMAGE:figures/full_fig_p006_3.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

32 extracted references · 32 canonical work pages · 3 internal anchors

  1. [1]

    White Paper on Broadband Connectivity in 6G

    Nandana Rajatheva, Italo Atzeni, Emil Bjornson, Andre Bourdoux, Stefano Buzzi, Jean- Baptiste Dore, Serhat Erkucuk, Manuel Fuentes, Ke Guan, Yuzhou Hu, et al. White paper on broadband connectivity in 6g.arXiv preprint arXiv:2004.14247, 2020

    work page internal anchor Pith review Pith/arXiv arXiv 2004

  2. [2]

    Al-Naffouri, and Mohamed-Slim Alouini

    Hadi Sarieddeen, Nasir Saeed, Tareq Y. Al-Naffouri, and Mohamed-Slim Alouini. Next Generation Terahertz Communications: A Rendezvous of Sensing, Imaging, and Localiza- tion.IEEE Communications Magazine, 58(5):69–75, 2020

    work page 2020

  3. [3]

    Terahertz band communication: An old problem revisited and research directions for the next decade

    Ian F Akyildiz, Chong Han, Zhifeng Hu, Shuai Nie, and Josep Miquel Jornet. Terahertz band communication: An old problem revisited and research directions for the next decade. IEEE Transactions on Communications, 70(6):4250–4285, 2022. 7

    work page 2022

  4. [4]

    Al-Naffouri

    Simon Tarboush, Hadi Sarieddeen, Hui Chen, Mohamed Habib Loukil, Hakim Jemaa, Mohamed-Slim Alouini, and Tareq Y. Al-Naffouri. TeraMIMO: A Channel Simulator for Wideband Ultra-Massive MIMO Terahertz Communications.IEEE Transactions on Ve- hicular Technology, 70(12):12325–12341, 2021

    work page 2021

  5. [5]

    Al-Naffouri

    Hadi Sarieddeen, Mohamed-Slim Alouini, and Tareq Y. Al-Naffouri. An Overview of Signal Processing Techniques for Terahertz Communications.Proceedings of the IEEE, 109(10):1628–1665, 2021

    work page 2021

  6. [6]

    Al-Naffouri

    Ahmad Dkhan, Simon Tarboush, Hadi Sarieddeen, Ibrahim Abou-Faycal, and Tareq Y. Al-Naffouri. THz-Band Near-Field RIS Channel Modeling for Linear Channel Estimation. IEEE Communications Letters, pages 1–1, 2025

    work page 2025

  7. [7]

    Deep learning enabled semantic communication systems.IEEE Transactions on Signal Processing, 69:2663–2675, 2021

    Huiqiang Xie, Zhijin Qin, Geoffrey Ye Li, and Biing-Hwang Juang. Deep learning enabled semantic communication systems.IEEE Transactions on Signal Processing, 69:2663–2675, 2021

    work page 2021

  8. [8]

    A unified multi- task semantic communication system with domain adaptation

    Guangyi Zhang, Qiyu Hu, Zhijin Qin, Yunlong Cai, and Guanding Yu. A unified multi- task semantic communication system with domain adaptation. InGLOBECOM 2022-2022 IEEE Global Communications Conference, pages 3971–3976. IEEE, 2022

    work page 2022

  9. [9]

    Semantic communications for speech signals

    Zhenzi Weng, Zhijin Qin, and Geoffrey Ye Li. Semantic communications for speech signals. InICC 2021-IEEE International Conference on Communications, pages 1–6. IEEE, 2021

    work page 2021

  10. [10]

    A robust semantic text communication system.IEEE Transactions on Wireless Communications, 2024

    Xiang Peng, Zhijin Qin, Xiaoming Tao, Jianhua Lu, and Lajos Hanzo. A robust semantic text communication system.IEEE Transactions on Wireless Communications, 2024

    work page 2024

  11. [11]

    Adaptive neural signal detection for massive MIMO.IEEE Transactions on Wireless Communications, 19(8):5635–5648, 2020

    Mehrdad Khani, Mohammad Alizadeh, Jakob Hoydis, and Phil Fleming. Adaptive neural signal detection for massive MIMO.IEEE Transactions on Wireless Communications, 19(8):5635–5648, 2020

    work page 2020

  12. [12]

    A model-driven deep learning network for MIMO detection

    Hengtao He, Chao-Kai Wen, Shi Jin, and Geoffrey Ye Li. A model-driven deep learning network for MIMO detection. In2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP), pages 584–588. IEEE, 2018

    work page 2018

  13. [13]

    A Variational Bayesian Inference-Inspired Unrolled Deep Network for MIMO Detection.IEEE Transac- tions on Signal Processing, 70:423–437, 2022

    Qian Wan, Jun Fang, Yinsen Huang, Huiping Duan, and Hongbin Li. A Variational Bayesian Inference-Inspired Unrolled Deep Network for MIMO Detection.IEEE Transac- tions on Signal Processing, 70:423–437, 2022

    work page 2022

  14. [14]

    Cddm: Channel denoising diffusion models for wireless communications

    Tong Wu, Zhiyong Chen, Dazhi He, Liang Qian, Yin Xu, Meixia Tao, and Wenjun Zhang. Cddm: Channel denoising diffusion models for wireless communications. InGLOBECOM 2023 - 2023 IEEE Global Communications Conference, pages 7429–7434, 2023

    work page 2023

  15. [15]

    Diffsc: Semantic communication framework with enhanced denoising through diffusion probabilis- tic models

    Zeyu Jiang, Xiaohong Liu, Guoxing Yang, Weizhi Li, Aini Li, and Guangyu Wang. Diffsc: Semantic communication framework with enhanced denoising through diffusion probabilis- tic models. InICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 13071–13075, 2024

    work page 2024

  16. [16]

    Dm-mimo: Diffusion models for robust semantic communications over mimo channels

    Yiheng Duan, Tong Wu, Zhiyong Chen, and Meixia Tao. Dm-mimo: Diffusion models for robust semantic communications over mimo channels. In2024 IEEE/CIC International Conference on Communications in China (ICCC), pages 1609–1614. IEEE, 2024

    work page 2024

  17. [17]

    Transrx-6g-v2x: Transformer encoder-based deep neural receiver for next generation of cellular vehicular communications

    Osama Saleem, Soheyb Ribouh, Mohammed Alfaqawi, Abdelaziz Bensrhair, and Pierre Merdrignac. Transrx-6g-v2x: Transformer encoder-based deep neural receiver for next generation of cellular vehicular communications. In2024 IEEE 100th Vehicular Technology Conference (VTC2024-Fall), pages 1–7. IEEE, 2024. 8

    work page 2024

  18. [18]

    Mobilebert: a compact task-agnostic bert for resource-limited devices

    Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou. Mobilebert: a compact task-agnostic bert for resource-limited devices. InProceedings of the 58th annual meeting of the association for computational linguistics, pages 2158–2170, 2020

    work page 2020

  19. [19]

    Adam: A Method for Stochastic Optimization.Interna- tional Conference on Learning Representations, 12 2014

    Diederik Kingma and Jimmy Ba. Adam: A Method for Stochastic Optimization.Interna- tional Conference on Learning Representations, 12 2014

    work page 2014

  20. [20]

    Super Convergence Cosine Annealing with Warm-Up Learning Rate

    Zhao Liu. Super Convergence Cosine Annealing with Warm-Up Learning Rate. InCAIBDA 2022; 2nd International Conference on Artificial Intelligence, Big Data and Algorithms, pages 1–7, 2022

    work page 2022

  21. [21]

    Yuhang Chen, Renwang Li, Chong Han, Shu Sun, and Meixia Tao. Hybrid spherical- and planar-wave channel modeling and estimation for terahertz integrated um-mimo and irs systems.IEEE Transactions on Wireless Communications, 22(12):9746–9761, 2023

    work page 2023

  22. [22]

    Channel measurement and analysis in an indoor corridor scenario at 300 ghz

    Yuanbo Li, Yiqin Wang, Yi Chen, Ziming Yu, and Chong Han. Channel measurement and analysis in an indoor corridor scenario at 300 ghz. InICC 2022 - IEEE International Conference on Communications, pages 2888–2893, 2022

    work page 2022

  23. [23]

    Channel measurement and path-loss characterization for low-terahertz indoor scenarios

    Jia He, Yi Chen, Yiqin Wang, Ziming Yu, and Chong Han. Channel measurement and path-loss characterization for low-terahertz indoor scenarios. In2021 IEEE International Conference on Communications Workshops (ICC Workshops), pages 1–6, 2021

    work page 2021

  24. [24]

    Rafiqul, and A H M Zahirul Alam

    Noor Hidayah Muhamad Adnan, Islam Md. Rafiqul, and A H M Zahirul Alam. Effects of inter element spacing on large antenna array characteristics. In2017 IEEE 4th International Conference on Smart Instrumentation, Measurement and Application (ICSIMA), pages 1– 5, 2017

    work page 2017

  25. [25]

    Ultramassive MIMO systems at terahertz bands: Prospects and challenges.IEEE Vehicular Technology Magazine, 15(4):33–42, 2020

    Alice Faisal, Hadi Sarieddeen, Hayssam Dahrouj, Tareq Y Al-Naffouri, and Mohamed-Slim Alouini. Ultramassive MIMO systems at terahertz bands: Prospects and challenges.IEEE Vehicular Technology Magazine, 15(4):33–42, 2020

    work page 2020

  26. [26]

    Akyildiz and Josep Miquel Jornet

    Ian F. Akyildiz and Josep Miquel Jornet. Realizing ultra-massive mimo (1024×1024) communication in the (0.06–10) terahertz band.Nano Communication Networks, 8:46–54,

    work page

  27. [27]

    Electromagnetic Communication in Nano-scale

    work page

  28. [28]

    Bleu: a method for automatic evaluation of machine translation

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. Bleu: a method for automatic evaluation of machine translation. InProceedings of the 40th annual meeting of the Association for Computational Linguistics, pages 311–318, 2002

    work page 2002

  29. [29]

    Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks

    Nils Reimers and Iryna Gurevych. Sentence-bert: Sentence embeddings using siamese bert-networks.arXiv preprint arXiv:1908.10084, 2019

    work page internal anchor Pith review Pith/arXiv arXiv 1908

  30. [30]

    Akyildiz

    Josep Miquel Jornet and Ian F. Akyildiz. Channel Modeling and Capacity Analysis for Electromagnetic Wireless Nanonetworks in the Terahertz Band.IEEE Transactions on Wireless Communications, 10(10):3211–3221, 2011

    work page 2011

  31. [31]

    Sionna: An Open-Source Library for Next-Generation Physical Layer Research

    Jakob Hoydis, Sebastian Cammerer, Fay¸ cal Ait Aoudia, Avinash Vem, Nikolaus Binder, Guillermo Marcus, and Alexander Keller. Sionna: An open-source library for next- generation physical layer research.arXiv preprint arXiv:2203.11854, 2022

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  32. [32]

    An improved least squares (LS) chan- nel estimation method based on CNN for OFDM systems.Electronic Research Archive, 31:5780–5792, 01 2023

    Hua Yang, Xuan Geng, Heng Xu, and Yichun Shi. An improved least squares (LS) chan- nel estimation method based on CNN for OFDM systems.Electronic Research Archive, 31:5780–5792, 01 2023. 9

    work page 2023