Pith. sign in

REVIEW 3 major objections 2 minor 2 cited by

WiFo-CF: Wireless Foundation Model for CSI Feedback

T0 review · 3 major / 2 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read WiFo-CF is a wireless foundation model for CSI feedback designed so that one pretrained network handles heterogeneous channel dimensions, feedback rates, and data distributions, and reports better in- and out-of-distribution accuracy than c

desk verdict A plausible foundation-model pitch for CSI feedback, but the abstract alone cannot support the performance claims—the OOD generalization claim is especially underspecified. read the letter →

arxiv 2508.04068 v2 pith:V2UJQ34H submitted 2025-08-06 eess.SP

Xuanyu Liu , Shijian Gao , Boxun Liu , Xiang Cheng , Liuqing Yang This is my paper · ORCID
classification eess.SP
keywords CSIfeedbackwirelessfoundationmodelself-supervisedpretrainingmixtureofexpertschannelstateinformationindoorlocalizationout-of-distributiongeneralizationheterogeneousdataset
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper aims to show that CSI feedback—the compression and reporting of radio channel measurements—can be handled by one pretrained foundation model instead of many task-specific codecs. It introduces WiFo-CF, which combines a multi-user, multi-rate self-supervised pretraining scheme with a Mixture of Shared and Routed Expert (S-R MoE) architecture, and supports it with the first heterogeneous channel feedback dataset. The asserted payoff is that a single model works across varying channel dimensions, feedback rates, and data distributions, and remains accurate on out-of-distribution data from simulated and real-world settings. If true, it would let wireless systems deploy one adaptive compression model rather than retraining per configuration, and would make CSI representations reusable for tasks like indoor localization.

What carries the argument

The load-bearing mechanism is S-R MoE—a Mixture of Shared and Routed Expert layer in which a subset of experts is shared across all configurations and a subset is routed per input, so common channel structure and configuration-specific details are separated inside one network. Around this sits the multi-user, multi-rate self-supervised pretraining objective and the heterogeneous CSI dataset, which supplies the diversity in channel dimensions, feedback rates, and data distributions that the expert routing needs to learn meaningful shared/specialized separation.

What would settle it

Take a group of channel configurations whose antenna counts, feedback rates, and environments are deliberately absent from the pretraining dataset, and compare WiFo-CF's reconstruction error against a per-configuration autoencoder trained on those settings; the out-of-distribution claim is false if the foundation model cannot match the baseline without fine-tuning.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that heterogeneous CSI feedback can be unified in a single foundation model rather than split into per-configuration autoencoders. WiFo-CF is trained with multiple users and multiple feedback rates under a self-supervised objective, then uses an S-R MoE inside the encoder-decoder to keep shared channel structure in shared experts while routing configuration-specific behavior through specialized experts. Backed by the first heterogeneous channel feedback dataset, the authors report that the resulting model outperforms fixed-configuration baselines on both in-distribution and out-of-distribution data in simulation and on real-world measurement

Load-bearing premise

The load-bearing premise is that the heterogeneous pretraining data represents the real-world distribution shifts a deployed model will meet, and that S-R MoE routing separates shared from configuration-specific knowledge without supervision; if either fails, the out-of-distribution claim collapses.

Editorial extensions

If this is right

  • A single WiFo-CF model can replace per-configuration CSI feedback codecs, since varying channel dimensions and feedback rates are handled within one framework.
  • Deployment to a new wireless environment no longer requires training a dedicated autoencoder from scratch; the pretrained model is intended to work directly on out-of-distribution data.
  • The same learned representations support downstream tasks, so CSI feedback pretraining can be reused for indoor localization rather than treated as a compression-only objective.
  • The first heterogeneous channel feedback dataset becomes a reusable pretraining resource for the field.
  • If the pretraining approach scales, larger heterogeneous datasets could produce increasingly general CSI representations.

Reading between the lines

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

  • I would expect the S-R MoE's value to show up most clearly in the tail of rare configurations rather than in average compression gains; an ablation varying the ratio of shared to routed experts would test whether shared experts truly carry configuration-agnostic structure.
  • The self-supervised organization suggests the same multi-user, multi-rate recipe could transfer to other physical-layer codec tasks, such as beamforming feedback or channel prediction, where per-configuration models are currently the norm.
  • If the out-of-distribution claim holds, WiFo-CF points toward a standards-friendly deployment model: one model shipped to devices, with configuration changes handled by input conditioning rather than model updates—a consequence the paper does not spell out.
  • A direct extension would be to quantify how much fine-tuning is needed per new environment; the paper demonstrates transfer but does not report the adaptation cost.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 2 minor

Summary. The manuscript proposes WiFo-CF, a wireless foundation model for CSI feedback that combines a multi-user, multi-rate self-supervised pre-training strategy with a Mixture of Shared and Routed Expert (S-R MoE) architecture. The abstract claims that this single model accommodates heterogeneous configurations—varying channel dimensions, feedback rates, and data distributions—and achieves superior performance on both in-distribution and out-of-distribution data across simulated and real-world scenarios. The paper also introduces what it describes as the first heterogeneous channel feedback dataset and claims that the learned representations transfer to CSI-based indoor localization. The reviewed material contains the abstract only; no equations, experimental setup, baselines, or quantitative results are provided.

Significance. If the claims hold, the contribution would be significant: a single pretrained model that handles diverse CSI feedback configurations and transfers to downstream tasks could reduce the need for per-configuration training and improve deployment flexibility. The introduction of a large-scale heterogeneous dataset would also be a useful community resource. However, the significance is entirely conditional: the abstract offers no verifiable evidence, no comparisons, and no architectural or experimental details, so the practical value cannot be assessed from the reviewed material.

major comments (3)
  1. [Abstract (entire)] The central claims—'superior performance on both in-distribution and out-of-distribution data' and 'effectively facilitate adaptation'—are stated without any quantitative results. No baselines, metrics, error bars, dataset sizes, or evaluation protocol are presented. The reader cannot verify even the direction of the claimed improvements. At minimum, a valid evaluation against existing CSI feedback methods (e.g., CsiNet, CRNet) on common benchmarks is required to support these claims.
  2. [Abstract, OOD claim] The term 'out-of-distribution' is undefined. The paper must specify how OOD test sets are constructed relative to the pretraining data: which axes (channel dimensions, feedback rates, data distributions) are shifted, what overlap analysis was performed, and whether the same users or sites appear in both training and testing. Without this, the reported OOD gains may simply reflect diversity within the pretraining distribution rather than genuine generalization to unseen configurations.
  3. [Abstract, S-R MoE architecture] The Mixture of Shared and Routed Expert design is described as if it separates shared from configuration-specific knowledge, but no mechanism or analysis is provided. The routing behavior is asserted rather than derived. The paper should define how routing is trained, how the shared/routed split is enforced, and include ablations showing that this separation is actually responsible for any performance gain over a simpler unified model (e.g., a single transformer or a dense MoE).
minor comments (2)
  1. [Abstract, dataset description] The 'first heterogeneous channel feedback dataset' is not characterized: number of users, channel dimensions, SNR/feedback-rate ranges, data sources (simulated/real-world), and availability are all unspecified. This metadata is essential for readers to gauge the scale and diversity of the pretraining corpus.
  2. [Abstract, terminology] The term 'foundation model' is used without specifying the scale or pre-training objective. Clarity on the model size, pre-training loss, and any constraints (e.g., input tokenization for variable dimensions) would help position the work relative to existing CSI deep-learning methods.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity identified: the abstract claims pretraining on a heterogeneous dataset and evaluation on OOD data, with no equations, fitted parameters, or self-citations that reduce the result to its inputs.

full rationale

The available text is only the abstract. It describes a pretraining dataset, a self-supervised multi-user/multi-rate pretraining strategy, and a Mixture of Shared and Routed Expert architecture, followed by a claim of superior performance on in-distribution and out-of-distribution data. No derivation chain, equation, or fitted-input/prediction relation is present, so there is no specific reduction to exhibit. The claim that OOD performance is enabled by 'diverse patterns' in the dataset could in principle be questioned if the OOD test sets overlapped with pretraining data, but the abstract gives no definitions or experimental protocol that would make the claim circular by construction. The absence of detail about what 'out-of-distribution' means is an empirical validation concern, not a circularity concern under the hard rules. There are no self-citations and no imported uniqueness theorems. Accordingly, the honest finding is no significant circularity, score 0.

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

All premises are extracted from the abstract because the full text is unavailable. The central gap is the untested assumption that pretrained representations on a heterogeneous dataset transfer to unseen configurations and downstream tasks.

assumptions (3)
  • domain assumption Deep learning can effectively compress CSI feedback
    The abstract relies on prior deep learning-based CSI feedback schemes as the baseline; this is background for the field and is not derived in the paper.
  • domain assumption Self-supervised pretraining on a large heterogeneous CSI dataset produces representations that transfer across channel dimensions, rates, and distributions
    The central generalization claim rests on this assumption; the abstract gives no proof or analysis of why pretraining transfers.
  • ad hoc to paper The S-R MoE routing is sufficient to capture both shared and configuration-specific feedback patterns
    This is a design choice introduced by the paper; the abstract does not provide an ablation or theoretical justification.

how reviews work

0 comments
Cite this review

Pith. "Pith review of WiFo-CF: Wireless Foundation Model for CSI Feedback." pith.science (2026). https://pith.science/paper/V2UJQ34H

@misc{pith2026250804068,
  author       = {Pith},
  title        = {Pith review of: WiFo-CF: Wireless Foundation Model for CSI Feedback},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/V2UJQ34H}},
  note         = {Machine review of arXiv:2508.04068}
}
read the original abstract

Deep learning-based channel state information (CSI) feedback schemes demonstrate strong compression capabilities but are typically constrained to fixed system configurations, limiting their generalization and flexibility. To address this challenge, WiFo-CF, a novel wireless foundation model tailored for CSI feedback, is proposed, uniquely accommodating heterogeneous configurations such as varying channel dimensions, feedback rates, and data distributions within a unified framework through its key innovations: (1) a multi-user, multi-rate self-supervised pre-training strategy; and (2) a Mixture of Shared and Routed Expert (S-R MoE) architecture. Supporting the large-scale pre-training of WiFo-CF is the first heterogeneous channel feedback dataset, whose diverse patterns enable the model to achieve superior performance on both in-distribution and out-of-distribution data across simulated and real-world scenarios. Furthermore, the learned representations effectively facilitate adaptation to downstream tasks such as CSI-based indoor localization, validating WiFo-CF's scalability and deployment potential.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Hierarchical Wireless Foundation Model for Multi-Task Optimization

    eess.SP 2026-07 conditional novelty 6.0 of 10

    A hierarchical wireless foundation model with a shared channel encoder and prompt-conditioned decoder solves beamforming, scheduling, channel estimation, and beam selection with competitive performance and large laten...

  2. Towards channel foundation models (CFMs): Motivations, methodologies and opportunities

    eess.SP 2025-07 conditional novelty 4.0 of 10

    A survey and position paper proposing channel foundation models, with experiments on two pretrained CSI models showing gains over a vanilla ViT baseline.

Reference graph

Works this paper leans on

44 extracted references · 38 canonical work pages · cited by 2 Pith papers

  1. [1]

    E. G. Larsson, O. Edfors, F. Tufvesson, and T. L. Marzetta, `` Massive MIMO for Next Generation Wireless Systems ,'' IEEE Commun. Mag., vol. 52, no. 2, pp. 186--195, Feb. 2014

    work page 2014

  2. [2]

    L. Lu, G. Y. Li, A. L. Swindlehurst, A. Ashikhmin, and R. Zhang, `` An Overview of Massive MIMO: Benefits and Challenges ,'' IEEE J. Sel. Top. Signal Process., vol. 8, no. 5, pp. 742--758, Apr. 2014

    work page 2014

  3. [3]

    Rusek et al., `` Scaling Up MIMO: Opportunities and Challenges with Very Large Arrays ,'' IEEE Signal Process

    F. Rusek et al., `` Scaling Up MIMO: Opportunities and Challenges with Very Large Arrays ,'' IEEE Signal Process. Mag., vol. 30, no. 1, pp. 40--60, Jan. 2012

    work page 2012

  4. [4]

    Cheng et al., `` Intelligent Multi-Modal Sensing-Communication Integration: Synesthesia of Machines ,'' IEEE Commun

    X. Cheng et al., `` Intelligent Multi-Modal Sensing-Communication Integration: Synesthesia of Machines ,'' IEEE Commun. Surv. Tutorials , vol. 26, pp. 258--301, Firstquarter 2024

    work page 2024

  5. [5]

    B. Liu, S. Gao, Z. Yang, X. Cheng, and L. Yang, ``Beam pattern modulation embedded hybrid transceiver optimization for integrated sensing and communication,'' IEEE Trans. Wireless Commun. , vol. 24, no. 6, pp. 4966--4980, Jun. 2025

    work page 2025

  6. [6]

    Zhang, S

    H. Zhang, S. Gao, X. Cheng, and L. Yang, `` Integrated Sensing and Communications Towards Proactive Beamforming in mmWave V2I via Multi-Modal Feature Fusion (MMFF) ,'' IEEE Trans. Wireless Commun. , vol. 23, no. 11, pp. 15\,721--15\,735, Nov. 2024

    work page 2024

  7. [7]

    S. T. Chung and A. J. Goldsmith, `` Degrees of freedom in adaptive modulation: a unified view ,'' IEEE Trans. Commun., vol. 49, no. 9, pp. 1561--1571, Sep. 2001

    work page 2001

  8. [8]

    S. Sadr, A. Anpalagan, and K. Raahemifar, `` Radio resource allocation algorithms for the downlink of multiuser OFDM communication systems ,'' IEEE Commun. Surv. Tutorials, vol. 11, no. 3, pp. 92--106, Aug. 2009

    work page 2009

Show all 44 references
  1. [9]

    Huang, R

    K. Huang, R. W. Heath, and J. G. Andrews, `` Limited feedback beamforming over temporally-correlated channels ,'' IEEE Trans. Signal Process., vol. 57, no. 5, pp. 1959--1975, Aug. 2009

    1959

  2. [10]

    K. Kim, T. Kim, D. J. Love, and I. H. Kim, `` Differential feedback in codebook-based multiuser MIMO systems in slowly varying channels ,'' IEEE Trans. Commun., vol. 60, no. 2, pp. 578--588, Feb. 2012

    2012

  3. [11]

    Daubechies, M

    I. Daubechies, M. Defrise, and C. De Mol, `` An iterative thresholding algorithm for linear inverse problems with a sparsity constraint ,'' Commun. Pure Appl. Math., vol. 57, no. 11, pp. 1413--1457, Aug. 2004

    2004

  4. [12]

    C. A. Metzler, A. Maleki, and R. G. Baraniuk, `` From denoising to compressed sensing ,'' IEEE Trans. Inf. Theory, vol. 62, no. 9, pp. 5117--5144, Apr. 2016

    2016

  5. [13]

    C. Li, W. Yin, and Y. Zhang, `` User’s guide for TVAL3: TV minimization by augmented lagrangian and alternating direction algorithms ,'' CAAM report, vol. 20, no. 46-47, p. 4, 2009

    2009

  6. [14]

    S. Gao, X. Cheng, and L. Yang, `` Estimating Doubly-Selective Channels for Hybrid mmWave Massive MIMO Systems: A Doubly-Sparse Approach ,'' IEEE Trans. Wireless Commun., vol. 19, no. 9, pp. 5703--5715, May 2020

    2020

  7. [15]

    Wen, W.-T

    C.-K. Wen, W.-T. Shih, and S. Jin, `` Deep learning for massive MIMO CSI feedback ,'' IEEE Wireless Commun. Lett., vol. 7, no. 5, pp. 748--751, Mar. 2018

    2018

  8. [16]

    Guo, C.-K

    J. Guo, C.-K. Wen, S. Jin, and G. Y. Li, `` Convolutional neural network-based multiple-rate compressive sensing for massive MIMO CSI feedback: Design, simulation, and analysis ,'' IEEE Trans. Wireless Commun., vol. 19, no. 4, pp. 2827--2840, Jan. 2020

    2020

  9. [17]

    Wang, C.-K

    T. Wang, C.-K. Wen, S. Jin, and G. Y. Li, `` Deep learning-based CSI feedback approach for time-varying massive MIMO channels ,'' IEEE Wireless Commun. Lett., vol. 8, no. 2, pp. 416--419, Oct. 2018

    2018

  10. [18]

    Z. Lu, J. Wang, and J. Song, `` Multi-resolution CSI feedback with deep learning in massive MIMO system ,'' in Int. Conf. Commun. (ICC). 1em plus 0.5em minus 0.4em Dublin, Ireland: IEEE, Jun. 2020, pp. 1--6

    2020

  11. [19]

    Q. Cai, C. Dong, and K. Niu, `` Attention model for massive MIMO CSI compression feedback and recovery ,'' in Wireless Commun. Netw. Conf. (WCNC). 1em plus 0.5em minus 0.4em Marrakech, Morocco: IEEE, Apr. 2019, pp. 1--5

    2019

  12. [20]

    Ji and M

    S. Ji and M. Li, `` CLNet: Complex input lightweight neural network designed for massive MIMO CSI feedback ,'' IEEE Wireless Commun. Lett., vol. 10, no. 10, pp. 2318--2322, Jul. 2021

    2021

  13. [21]

    M. B. Mashhadi, Q. Yang, and D. G \"u nd \"u z, `` Distributed deep convolutional compression for massive MIMO CSI feedback ,'' IEEE Trans. Wireless Commun., vol. 20, no. 4, pp. 2621--2633, Dec. 2020

    2020

  14. [22]

    Zhang and A

    Y. Zhang and A. Alkhateeb, `` Zone-specific CSI feedback for massive MIMO: A situation-aware deep learning approach ,'' IEEE Wireless Commun. Lett., vol. 13, no. 12, pp. 3320--3324, Sep. 2024

    2024

  15. [23]

    B. Liu, X. Liu, S. Gao, X. Cheng, and L. Yang, `` LLM4CP: Adapting Large Language Models for Channel Prediction ,'' J. Commun. Inf. Networks, vol. 9, no. 2, pp. 113--125, Jun. 2024

    2024

  16. [24]

    X. Liu, S. Gao, B. Liu, X. Cheng, and L. Yang, `` LLM4WM: Adapting LLM for Wireless Multi-Tasking ,'' IEEE Trans. Mach. Learn. Commun. Networking, vol. 3, pp. 835--847, Jul. 2025

    2025

  17. [25]

    Cheng, B

    X. Cheng, B. Liu, X. Liu, E. Liu, and Z. Huang, `` Foundation Model Empowered Synesthesia of Machines (SoM): AI-native Intelligent Multi-Modal Sensing-Communication Integration ,'' IEEE Trans. Network Sci. Eng., 2025, early Access

    2025

  18. [26]

    Z. Han, L. Bai, Z. Huang, and X. Cheng, `` LLM4SG: Large Language Models for Scatterer Generation via Synesthesia of Machines ,'' arXiv preprint arXiv:2505.17879, 2025

    2025 arXiv

  19. [27]

    Alikhani, G

    S. Alikhani, G. Charan, and A. Alkhateeb, `` Large Wireless Model (LWM): A Foundation Model for Wireless Channels ,'' arXiv preprint arXiv:2411.08872, 2024

    2024 arXiv

  20. [28]

    B. Liu, S. Gao, X. Liu, X. Cheng, and L. Yang, `` WiFo: Wireless Foundation Model for Channel Prediction ,'' Sci. China Inf. Sci., vol. 68, no. 8, p. 162302, May 2025

    2025

  21. [29]

    Emery, A

    J. Emery, A. H. Karkan, J.-F. Frigon, and F. Leduc-Primeau, ``A foundation model for massive mimo precoding with an adaptive per-user rate-power tradeoff,'' arXiv preprint arXiv:2507.18587, 2025

    2025 arXiv

  22. [30]

    J. Guo, Y. Cui, C.-K. Wen, and S. Jin, `` Prompt-enabled large AI models for CSI feedback ,'' arXiv preprint arXiv:2501.10629, 2025

    2025 arXiv

  23. [31]

    Joham, W

    M. Joham, W. Utschick, and J. A. Nossek, `` Linear transmit processing in MIMO communications systems ,'' IEEE Trans. Signal Process., vol. 53, no. 8, pp. 2700--2712, Aug. 2005

    2005

  24. [32]

    Q. Shi, M. Razaviyayn, Z.-Q. Luo, and C. He, `` An iteratively weighted MMSE approach to distributed sum-utility maximization for a MIMO interfering broadcast channel ,'' IEEE Trans. Signal Process., vol. 59, no. 9, pp. 4331--4340, Apr. 2011

    2011

  25. [33]

    Dosovitskiy et al., `` An image is worth 16x16 words: Transformers for image recognition at scale ,'' arXiv preprint arXiv:2010.11929, 2020

    A. Dosovitskiy et al., `` An image is worth 16x16 words: Transformers for image recognition at scale ,'' arXiv preprint arXiv:2010.11929, 2020

    2010 arXiv

  26. [34]

    Vaswani et al., `` Attention Is All You Need ,'' in Adv

    A. Vaswani et al., `` Attention Is All You Need ,'' in Adv. Neural Inf. Process. Syst. (NeurIPS), vol. 30, Long Beach, CA, USA, Dec. 2017, pp. 6000--6010

    2017

  27. [35]

    Zhang and R

    B. Zhang and R. Sennrich, `` Root mean square layer normalization ,'' Adv. Neural Inf. Process. Syst. (NeurIPS), vol. 32, Dec. 2019

    2019

  28. [36]

    Liu et al., `` Deepseek-v3 technical report ,'' arXiv preprint arXiv:2412.19437, 2024

    A. Liu et al., `` Deepseek-v3 technical report ,'' arXiv preprint arXiv:2412.19437, 2024

    2024 arXiv

  29. [37]

    Jaeckel, L

    S. Jaeckel, L. Raschkowski, K. B \"o rner, and L. Thiele, `` QuaDRiGa: A 3-D Multi-Cell Channel Model With Time Evolution for Enabling Virtual Field Trials ,'' IEEE Trans. Antennas Propag., vol. 62, no. 6, pp. 3242--3256, Mar. 2014

    2014

  30. [38]

    Alkhateeb, `` DeepMIMO: A generic deep learning dataset for millimeter wave and massive MIMO applications ,'' arXiv preprint arXiv:1902.06435, 2019

    A. Alkhateeb, `` DeepMIMO: A generic deep learning dataset for millimeter wave and massive MIMO applications ,'' arXiv preprint arXiv:1902.06435, 2019

    1902 arXiv

  31. [39]

    Cheng et al., `` SynthSoM: A Synthetic Intelligent Multi-Modal Sensing-Communication Dataset for Synesthesia of Machines (SoM) ,'' Sci

    X. Cheng et al., `` SynthSoM: A Synthetic Intelligent Multi-Modal Sensing-Communication Dataset for Synesthesia of Machines (SoM) ,'' Sci. Data, vol. 12, no. 819, May 2025

    2025

  32. [40]

    Hoydis et al., `` Sionna: An open-source library for next-generation physical layer research ,'' arXiv preprint arXiv:2203.11854, 2022

    J. Hoydis et al., `` Sionna: An open-source library for next-generation physical layer research ,'' arXiv preprint arXiv:2203.11854, 2022

    2022 arXiv

  33. [41]

    Shepard, J

    C. Shepard, J. Ding, R. E. Guerra, and L. Zhong, `` Understanding real many-antenna MU-MIMO channels ,'' in Asilomar Conf. Signals Syst. Computers. 1em plus 0.5em minus 0.4em Pacific Grove, CA, USA: IEEE, Nov. 2016, pp. 461--467

    2016

  34. [42]

    orner, and S. ten Brink, `` A Distributed Massive MIMO Channel Sounder for

    F. Euchner, M. Gauger, S. D\"orner, and S. ten Brink, `` A Distributed Massive MIMO Channel Sounder for "Big CSI Data"-driven Machine Learning ,'' in Int. ITG Workshop Smart Antennas (WSA), Eurecom, France, Nov. 2021

    2021

  35. [43]

    Fedus, B

    W. Fedus, B. Zoph, and N. Shazeer, `` Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity ,'' J. Mach. Learn. Res., vol. 23, no. 120, pp. 1--39, Jan. 2022

    2022

  36. [44]

    F. Qi, J. Guo, Y. Cui, X. Li, C. Wen, and S. Jin, `` Deep learning-based CSI feedback in Wi-Fi systems ,'' in Int. Symp. Compu. Eng. Intell. Commun. (ISCEIC). 1em plus 0.5em minus 0.4em Nanjing, China: IEEE, Nov. 2024, pp. 43--48

    2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.