Pith. sign in

REVIEW 1 major objections 1 cited by

Lossy Microwave Linear Analog Computer (MiLAC) for Future MIMO: Learning-based Architecture Designs for Spectral and Energy Efficiency Maximization

T0 review · 1 major / 0 minor · reviewed 2026-06-28 · grok-4.3

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

Pith's one-line read A learning framework jointly designs lossy MiLAC architectures and beamforming to maximize MIMO spectral and energy efficiency by balancing interference against hardware losses.

desk verdict LJAPOF frames a joint learning approach to lossy MiLAC architecture and beamforming but its SE/EE gains rest entirely on simulated TAC loss models. read the letter →

arxiv 2606.02369 v1 pith:76EDZT75 submitted 2026-06-01 eess.SP cs.ITmath.IT

Binggui Zhou , Bruno Clerckx This is my paper · ORCID
classification eess.SPcs.ITmath.IT
keywords MiLACMIMOspectralefficiencyenergylearning-basedoptimizationanalogbeamformingarchitecturedesignhardwarelosses
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

The paper shows how microwave linear analog computers can move signal processing into the analog domain for MIMO systems, cutting digital hardware needs but running into losses from tunable admittance components that create interference and drain power. Denser connections help suppress interference yet add more losses and consumption, so the right architecture must trade one against the other. The authors introduce a learning-based joint architecture and performance optimization framework that chooses both the connection pattern and the analog beamforming weights together. Numerical tests indicate the resulting designs beat fixed stem-connected and fully-connected baselines on both spectral efficiency and energy efficiency. If the approach holds, it would let analog processors deliver usable performance gains while respecting real hardware constraints.

What carries the argument

The LJAPOF, a learning-based joint architecture and performance optimization framework that selects MiLAC port interconnections together with analog beamforming weights for either spectral-efficiency or energy-efficiency goals.

What would settle it

Build a hardware prototype using an architecture chosen by the LJAPOF and measure its achieved spectral efficiency and energy efficiency against identical measurements on stem-connected and fully-connected MiLAC prototypes under the same MIMO channel conditions.

Watch

Extended reading notes

Core claim

The central claim is that the proposed LJAPOF unifies MiLAC architecture design and analog beamforming configuration under SE- and EE-oriented objectives, allowing lossy MiLACs to navigate the tradeoff between interference suppression and cumulative hardware losses or power consumption and thereby consistently outperform stem-connected and fully-connected MiLACs.

Load-bearing premise

The simulated loss models used during learning accurately capture the behavior of actual fabricated tunable admittance components so that the discovered architectures deliver the predicted gains in real hardware.

Editorial extensions

If this is right

  • Optimal MiLAC architectures use fewer tunable admittance components than fully-connected versions yet suppress inter-stream interference more effectively than stem-connected versions.
  • The same learning framework produces competitive designs whether the objective is spectral efficiency or energy efficiency.
  • The performance advantage appears across the numerical evaluations presented for lossy MiLAC-aided MIMO systems.

Reading between the lines

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

  • The joint-design idea could be tested on other analog signal-processing hardware where losses and power scale with component count.
  • Incorporating measured non-idealities from a fabricated device back into the learning loop might further improve the selected architectures.
  • The resulting connection patterns could serve as starting points for hybrid analog-digital MIMO transmitters that aim to reduce RF-chain counts.
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

1 major / 0 minor

Summary. The manuscript proposes a learning-based joint architecture and performance optimization framework (LJAPOF) for lossy microwave linear analog computers (MiLACs) in MIMO systems. It jointly optimizes MiLAC architectures and analog beamforming under SE- and EE-oriented objectives to navigate the tradeoff between inter-stream interference suppression and cumulative TAC hardware losses/power consumption. Numerical results are presented to claim that the resulting architectures consistently outperform stem-connected and fully-connected baselines.

Significance. If the numerical gains prove robust, the work would address a practically relevant constraint on analog-domain MIMO processing and demonstrate a learning-based route to architecture selection. The explicit framing of the interference-vs-loss tradeoff is a clear contribution.

major comments (1)
  1. [Numerical Results] Numerical Results section: the central claim of consistent outperformance rests entirely on optimization over a specific simulated TAC loss model. No sensitivity study or hardware-in-the-loop validation is described to test whether the discovered architectures remain superior when the loss model is perturbed by frequency dependence, mutual coupling, or fabrication variance; this directly undermines the claim that the gains arise from intelligent navigation of the tradeoff rather than simulator overfitting.

Simulated Author's Rebuttal

1 responses · 1 unresolved

We thank the referee for the constructive feedback. Below we address the major comment point-by-point and indicate planned revisions.

read point-by-point responses

  1. Referee: [Numerical Results] Numerical Results section: the central claim of consistent outperformance rests entirely on optimization over a specific simulated TAC loss model. No sensitivity study or hardware-in-the-loop validation is described to test whether the discovered architectures remain superior when the loss model is perturbed by frequency dependence, mutual coupling, or fabrication variance; this directly undermines the claim that the gains arise from intelligent navigation of the tradeoff rather than simulator overfitting.

    Authors: We acknowledge that the presented results rely on a fixed simulated TAC loss model. The model parameters are drawn from established TAC characterizations in the literature, and all results are averaged over a large number of independent channel realizations. Nevertheless, we agree that explicit sensitivity analysis is needed to strengthen the claim of robust architecture selection. In the revision we will add a dedicated subsection that perturbs the loss model (frequency dependence, mutual coupling, and fabrication variance) and verifies that the learned architectures retain their SE/EE advantage over the baselines. Hardware-in-the-loop validation lies outside the scope of this theoretical/simulation study. revision: partial

standing simulated objections not resolved
  • Hardware-in-the-loop validation (the manuscript is purely simulation-based)

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity in derivation chain

full rationale

The paper proposes an explicit learning-based optimization framework (LJAPOF) for joint MiLAC architecture and beamforming design under SE/EE objectives, then reports numerical comparisons against stem- and fully-connected baselines. No first-principles derivation, uniqueness theorem, or ansatz is invoked. The results are obtained by direct optimization over the stated TAC loss model; this is not a hidden reduction of outputs to inputs by construction, nor does any load-bearing step rely on self-citation. The work is self-contained as an engineering design study whose validity rests on the fidelity of the simulator rather than on any circular mathematical step.

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

No free parameters, axioms, or invented entities can be identified from the abstract alone.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Lossy Microwave Linear Analog Computer (MiLAC) for Future MIMO: Learning-based Architecture Designs for Spectral and Energy Efficiency Maximization." pith.science (2026). https://pith.science/paper/76EDZT75

@misc{pith2026260602369,
  author       = {Pith},
  title        = {Pith review of: Lossy Microwave Linear Analog Computer (MiLAC) for Future MIMO: Learning-based Architecture Designs for Spectral and Energy Efficiency Maximization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/76EDZT75}},
  note         = {Machine review of arXiv:2606.02369}
}
read the original abstract

Microwave linear analog computers (MiLACs) offer a transformative paradigm for future multiple-input multiple-output (MIMO) systems by shifting complex signal processing into the analog domain, thereby significantly reducing computational complexity, radio-frequency chains, and analog-digital converters, while speeding up computation. However, the practical deployment of MiLACs is severely constrained by the inherent hardware losses of the tunable admittance components (TACs) interconnecting MiLAC ports, which introduce severe inter-stream interference and fundamentally limit the spectral efficiency (SE) of the system. In addition, while denser architectures offer greater spatial degrees of freedom to mitigate inter-stream interference, the cumulative hardware losses and power consumption of massive TACs severely degrade the system's energy efficiency (EE). Consequently, designing architectures for lossy MiLACs emerges as a critical yet unresolved challenge, as it necessitates striking a delicate tradeoff between interference suppression and cumulative hardware losses/power consumption. To address this challenge, this paper investigates the joint MiLAC architecture design and performance (SE/EE) maximization in lossy MiLAC-aided MIMO systems. We propose a novel learning-based joint architecture and performance optimization framework (LJAPOF) that unifies the design of MiLAC architectures and analog beamforming configurations for lossy MiLACs under both SE- and EE-oriented objectives. Numerical results demonstrate that by intelligently navigating the fundamental tradeoff between interference suppression and hardware/power consumption, the proposed LJAPOF can design optimal MiLAC architectures that consistently outperform stem-connected and fully-connected MiLACs in maximizing the system's SE and EE.

Figures

Figures reproduced from arXiv: 2606.02369 by the authors.

Figure 1
Figure 1. The lossy MiLAC-aided point-to-point MIMO system. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. The proposed learning-based joint architecture and performance optimization framework (LJAPOF). [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. SE of ideal and lossy MiLAC-aided MIMO systems [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: SE of lossy MiLAC-aided MIMO systems versus the [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: The SE and the number of active streams of lossy [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: SE and EE of lossy MiLAC-aided MIMO systems and the circuit complexity of the transmitter-side and receiver-side [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: (a) SE–circuit complexity, (b) EE-circuit complexity, and (c) EE-SE tradeoffs in lossy MiLAC-aided MIMO systems. [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. How Many RF Chains Does a Microwave Linear Analog Computer (MiLAC) Need to Match the Fully-Digital Cram\'er-Rao Bound?

    cs.IT 2026-06 unverdicted novelty 7.0 of 10

    MiLAC combiner matches fully digital CRB for DOA estimation of K targets with 2 RF chains per target when row space spans the 2K-dimensional steering-derivative subspace.

Reference graph

Works this paper leans on

37 extracted references · 9 canonical work pages · cited by 1 Pith paper

  1. [1]

    Analog signal process- ing: A possible alternative or complement to dominantly digital radio schemes,

    C. Caloz, S. Gupta, Q. Zhang, and B. Nikfal, “Analog signal process- ing: A possible alternative or complement to dominantly digital radio schemes,”IEEE Microw. Mag., vol. 14, no. 6, pp. 87–103, Sep. 2013

    2013

  2. [2]

    Analog computing for signal processing and communications – part I: Computing with microwave networks,

    M. Nerini and B. Clerckx, “Analog computing for signal processing and communications – part I: Computing with microwave networks,”IEEE Trans. Signal Process., vol. 73, pp. 5183–5197, 2025

    2025

  3. [3]

    Analog computing for signal processing and communications – part II: Toward gigantic MIMO beamforming,

    ——, “Analog computing for signal processing and communications – part II: Toward gigantic MIMO beamforming,”IEEE Trans. Signal Process., vol. 73, pp. 5198–5212, 2025

    2025

  4. [4]

    Capacity of MIMO systems aided by microwave linear analog computers (MiLACs),

    ——, “Capacity of MIMO systems aided by microwave linear analog computers (MiLACs),”arXiv preprint arXiv:2506.05983, Jun. 2025

    work page arXiv 2025

  5. [5]

    Microwave linear analog computer (MiLAC)-aided multiuser MISO: Fundamental limits and beamforming design,

    Z. Wu, M. Nerini, and B. Clerckx, “Microwave linear analog computer (MiLAC)-aided multiuser MISO: Fundamental limits and beamforming design,”arXiv preprint arXiv:2601.10060, Jan. 2026

    work page arXiv 2026

  6. [6]

    Physics-compliant modeling and optimization of MIMO systems aided by microwave linear analog computers,

    M. Nerini and B. Clerckx, “Physics-compliant modeling and optimiza- tion of mimo systems aided by microwave linear analog computers,” arXiv preprint arXiv:2602.19379, 2026

    work page arXiv 2026

  7. [7]

    Channel estimation in MIMO systems aided by microwave linear analog computers (MiLACs),

    Q. Zhang, M. Nerini, and B. Clerckx, “Channel estimation in mimo systems aided by microwave linear analog computers (milacs),”arXiv preprint arXiv:2601.11438, 2026

    work page arXiv 2026

  8. [8]

    MIMO systems aided by microwave linear analog computers: Capacity-achieving architectures with reduced circuit complexity,

    M. Nerini and B. Clerckx, “MIMO systems aided by microwave linear analog computers: Capacity-achieving architectures with reduced circuit complexity,”IEEE Trans. Wireless Commun., pp. 1–15, 2026

    2026

Show all 37 references
  1. [9]

    Modeling and architecture design of reconfigurable intelligent surfaces using scattering parameter network analysis,

    S. Shen, B. Clerckx, and R. Murch, “Modeling and architecture design of reconfigurable intelligent surfaces using scattering parameter network analysis,”IEEE Trans. Wireless Commun., vol. 21, no. 2, pp. 1229–1243, Feb. 2022

    2022

  2. [10]

    Reconfigurable intelligent surfaces 2.0: Beyond diagonal phase shift matrices,

    H. Li, S. Shen, M. Nerini, and B. Clerckx, “Reconfigurable intelligent surfaces 2.0: Beyond diagonal phase shift matrices,”IEEE Commun. Mag., vol. 62, no. 3, pp. 102–108, Mar. 2024

    2024

  3. [11]

    A tutorial on beyond- diagonal reconfigurable intelligent surfaces: Modeling, architectures, system design and optimization, and applications,

    H. Li, M. Nerini, S. Shen, and B. Clerckx, “A tutorial on beyond- diagonal reconfigurable intelligent surfaces: Modeling, architectures, system design and optimization, and applications,”IEEE Commun. Surveys Tuts., vol. 28, pp. 4086–4126, 2026

    2026

  4. [12]

    Beyond-diagonal RIS in multiuser MIMO: Graph theoretic modeling and optimal architectures with low complex- ity,

    Z. Wu and B. Clerckx, “Beyond-diagonal RIS in multiuser MIMO: Graph theoretic modeling and optimal architectures with low complex- ity,”IEEE Trans. Inf. Theory, vol. 71, no. 11, pp. 8506–8523, Nov. 2025

    2025

  5. [13]

    Beyond-diagonal RIS architecture design and optimization under physics-consistent models,

    Z. Wu, M. Nerini, and B. Clerckx, “Beyond-diagonal RIS architecture design and optimization under physics-consistent models,”IEEE Trans. Wireless Commun., vol. 25, pp. 14 086–14 100, 2026

    2026

  6. [15]

    Lossy beyond diagonal reconfigurable intelligent surfaces: Modeling and optimization,

    Y . Peng, H. Li, Z. Wu, and B. Clerckx, “Lossy beyond diagonal reconfigurable intelligent surfaces: Modeling and optimization,”IEEE Trans. Wireless Commun., vol. 25, pp. 7365–7380, 2026

    2026

  7. [16]

    Quantization-aware EE optimization and SE-EE tradeoff for MiLAC-aided MU-MISO beam- forming,

    Y . Zhang, P. Zheng, and T. Y . Al-Naffouri, “Quantization-aware EE optimization and SE-EE tradeoff for MiLAC-aided MU-MISO beam- forming,”arXiv preprint arXiv:2604.24538, 2026

    2026 arXiv

  8. [17]

    Mixed-integer nonlinear programming 2018,

    N. V . Sahinidis, “Mixed-integer nonlinear programming 2018,”Optim. Eng., vol. 20, no. 2, pp. 301–306, Jun. 2019

    2018

  9. [18]

    A low-overhead incorporation-extrapolation based few-shot CSI feedback framework for massive MIMO systems,

    B. Zhou, X. Yang, J. Wang, S. Ma, F. Gao, and G. Yang, “A low-overhead incorporation-extrapolation based few-shot CSI feedback framework for massive MIMO systems,”IEEE Trans. Wireless Com- mun., vol. 23, no. 10, pp. 14 743–14 758, Oct. 2024. 15

    2024

  10. [19]

    Joint training scattering matrix learning and channel estimation for beyond-diagonal reconfigurable intelligent surfaces,

    Y . Peng, B. Zhou, Y . Zheng, D. Mandic, and B. Clerckx, “Joint training scattering matrix learning and channel estimation for beyond-diagonal reconfigurable intelligent surfaces,”arXiv preprint arXiv:2603.25299, 2026

    2026

  11. [20]

    Deep learning-based channel extrapolation for dual-band massive MIMO systems,

    Q. Xiao, K. Li, B. Zhou, and S. Ma, “Deep learning-based channel extrapolation for dual-band massive MIMO systems,”IEEE Wireless Commun. Lett., vol. 15, pp. 2994–2998, 2026

    2026

  12. [21]

    Beyond-diagonal ris under non-idealities: Learning-based architecture discovery and optimization,

    B. Zhou and B. Clerckx, “Beyond-diagonal ris under non-idealities: Learning-based architecture discovery and optimization,”arXiv preprint arXiv:2510.15701, 2025

    2025

  13. [22]

    D. M. Pozar,Microwave Engineering, 4th ed. Hoboken, NJ, USA: John Wiley & Sons, 2011

    2011

  14. [23]

    Clerckx and C

    B. Clerckx and C. Oestges,MIMO Wireless Networks: Channels, Techniques and Standards for Multi-Antenna, Multi-User and Multi-Cell Systems. Academic Press, 2013

    2013

  15. [24]

    Sleep modes for enhanced battery life of 5G mobile terminals,

    M. Lauridsen, G. Berardinelli, F. M. L. Tavares, F. Frederiksen, and P. Mogensen, “Sleep modes for enhanced battery life of 5G mobile terminals,” inProc. IEEE VTC Spring, May 2016, pp. 1–6

    2016

  16. [25]

    Hardware-aware energy efficiency optimization in wireless communi- cations using a gearbox-PHY,

    F. Gast, M. D ¨orpinghaus, P. Sen, A. Nimr, and G. P. Fettweis, “Hardware-aware energy efficiency optimization in wireless communi- cations using a gearbox-PHY,”IEEE Commun. Lett., vol. 28, no. 7, pp. 1584–1588, Jul. 2024

    2024

  17. [26]

    Millimeter wave receiver efficiency: A comprehensive comparison of beamforming schemes with low resolution adcs,

    W. B. Abbas, F. Gomez-Cuba, and M. Zorzi, “Millimeter wave receiver efficiency: A comprehensive comparison of beamforming schemes with low resolution adcs,”IEEE Trans. Wireless Commun., vol. 16, no. 12, pp. 8131–8146, Dec. 2017

    2017

  18. [27]

    Mixed-ADC/DAC multipair massive MIMO relaying systems: Performance analysis and power optimization,

    J. Zhang, L. Dai, Z. He, B. Ai, and O. A. Dobre, “Mixed-ADC/DAC multipair massive MIMO relaying systems: Performance analysis and power optimization,”IEEE Trans. Commun., vol. 67, no. 1, pp. 141– 153, Jan. 2019

    2019

  19. [28]

    Estimating or propagating gradients through stochastic neurons for conditional computation,

    Y . Bengio, N. L ´eonard, and A. Courville, “Estimating or propagating gradients through stochastic neurons for conditional computation,”arXiv preprint arXiv:1308.3432, Aug. 2013

    2013 arXiv

  20. [29]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” inProc. IEEE/CVF CVPR, Jun. 2016, pp. 770–778

    2016

  21. [30]

    Layer normalization,

    J. L. Ba, J. R. Kiros, and G. E. Hinton, “Layer normalization,”arXiv preprint arXiv:1607.06450, Jul. 2016

    2016 arXiv

  22. [31]

    Deep learning-based power allocation for cell-free massive MIMO networks with adaptive access point power control,

    Y .-J. Choi, J.-H. Yu, H.-Y . Jeong, J.-E. Kim, and H.-K. Song, “Deep learning-based power allocation for cell-free massive MIMO networks with adaptive access point power control,” inProc. ICUFN, Jul. 2025, pp. 744–749

    2025

  23. [32]

    Deep learning-based power allocation in massive MIMO systems with SLNR and SINR criterions,

    R. H. Y . Perdana, T.-V . Nguyen, and B. An, “Deep learning-based power allocation in massive MIMO systems with SLNR and SINR criterions,” inProc. ICUFN, Aug. 2021, pp. 87–92

    2021

  24. [33]

    Deep learning power allocation in massive MIMO,

    L. Sanguinetti, A. Zappone, and M. Debbah, “Deep learning power allocation in massive MIMO,” inProc. ACSSC, Oct. 2018, pp. 1257– 1261

    2018

  25. [34]

    Shortcut learning in deep neural networks,

    R. Geirhos, J.-H. Jacobsen, C. Michaelis, R. Zemel, W. Brendel, M. Bethge, and F. A. Wichmann, “Shortcut learning in deep neural networks,”Nat Mach Intell, vol. 2, no. 11, pp. 665–673, Nov. 2020

    2020

  26. [35]

    A 7.5-GS/s 3.8-ENOB 52-mW flash ADC with clock duty cycle control in 65nm CMOS,

    H. Chung, A. Rylyakov, Z. T. Deniz, J. Bulzacchelli, G.-Y . Wei, and D. Friedman, “A 7.5-GS/s 3.8-ENOB 52-mW flash ADC with clock duty cycle control in 65nm CMOS,” inProc. Symp. VLSI Circuits, Jun. 2009, pp. 268–269

    2009

  27. [36]

    Modeling and minimization of transceiver power consumption in wireless networks,

    A. Mezghani and J. A. Nossek, “Modeling and minimization of transceiver power consumption in wireless networks,” inProc. 2011 Int. ITG Workshop Smart Antennas, Feb. 2011, pp. 1–8

    2011

  28. [37]

    Reconfigurable intelligent surface: Power consumption modeling and practical measurement validation,

    J. Wang, W. Tang, J. C. Liang, L. Zhang, J. Y . Dai, X. Li, S. Jin, Q. Cheng, and T. J. Cui, “Reconfigurable intelligent surface: Power consumption modeling and practical measurement validation,”IEEE Trans. Commun., vol. 72, no. 9, pp. 5720–5734, Sep. 2024

    2024

  29. [38]

    PyTorch: An imperative style, high- performance deep learning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Kopf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, “PyTorch: An imperative style, high- p...

    2019

Pith tools

Reviewed June 28, 2026 · model on record in the stance chip above.