Pith. sign in

REVIEW 4 major objections 3 minor 34 references

Dark Matter Axion Detection with Neural Networks at Ultra-Low Signal-to-Noise Ratio

T0 review · 4 major / 3 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read The paper claims that a single-neuron feedforward network, trained on simulated haloscope noise, reaches in about half an hour a detection sensitivity that would otherwise take 100 days of integration — a 5,000-fold reduction in exposure…

desk verdict A plausible simulation study whose headline 5e3 time improvement is an artifact of an invalid accuracy-to-SNR conversion. read the letter →

arxiv 2411.17947 v3 pith:XG5UBTKZ submitted 2024-11-26 hep-ex

classification hep-ex
keywords axiondarkmatterhaloscopeneuralnetworksignal-to-noiseratiothermalnoiseBI-RME3Dfeedforwardlow-noiseamplifier
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 is the first attempt to apply a neural network directly to the axion dark-matter search, with the aim of improving sensitivity rather than just automating data analysis. The authors simulate the dominant noise sources of a haloscope read-out chain — cavity resonant noise and the first low-noise amplifier — using an exact modal method, and train a feedforward network with a single hidden neuron to decide, from each recorded radio-frequency trace, whether an axion signal is present or only noise. They report that at a per-trace signal-to-noise ratio of about 0.036 the network classifies correctly in 99.3% of test cases, which they equate to the significance a conventional analysis would reach only after roughly 5,000 times longer integration. If correct, this means a measurement point that takes 100 days to reach signal-to-noise ratio 3 could instead be completed in about half an hour, or the same exposure could probe smaller axion-photon couplings. The practical value of the claim is that it applies to any haloscope experiment and can run in parallel with standard analysis pipelines.

What carries the argument

The argument is carried by three pieces. First, the BI-RME3D modal method, an exact full-wave technique that turns the axion-photon coupling into circuit quantities: the axion is a current source $I_a$ added to a random Gaussian noise current $I_n$, and the detected voltage is the phasor $V_c=(I_a+I_n)/(Y_w+Y_c)$, so the delivered power $P_w$ carries magnitude and phase information. Second, the Dicke radiometer equation ${\rm SNR}=(P_w/k_BT_{\rm sys})\sqrt{t/\Delta\nu}$, which sets how integration time $t$ converts into signal-to-noise ratio. Third, the feedforward neural network — one hidden neuron, one boolean output — trained on 5000 simulated examples per SNR and tested on 1000, whose accuracy (the fraction of correct 'axion present or not' decisions) is the quantity the whole claim hangs on. The time-improvement factor is then defined by $T=({\rm SNR}_{\rm equiv}/{\rm SNR}_0)^2$, where ${\rm SNR}_{\rm equiv}$ is read off as the number of standard deviations of a standard Gaussian distribution that corresponds to the network's accuracy.

What would settle it

The claim would be settled by measuring the network's false-positive rate on real thermal noise that certainly contains no axion: at a claimed ${\rm SNR}_{\rm equiv}\approx 2.7$, a Gaussian reading predicts a two-sided false-alarm probability of roughly 0.7%, so if the network flags 'axion present' far more often on pure noise data, the accuracy-to-significance mapping is invalid and the time improvement does not exist. A complementary check is to re-run the same training on genuine off-resonance spectra from a working haloscope with a synthetic axion line injected at ${\rm SNR}=0.036$; if accuracy does not reach the simulated 99.3%, the improvement is an artifact of the simulated noise statistics.

Watch

Extended reading notes

Core claim

The central claim is that a very simple neural network can pull an axion signal out of thermal noise far below the signal-to-noise ratio that conventional power averaging requires. The authors model the axion as a current source in a rectangular cavity with the Boundary Integral-Resonant Mode Expansion (BI-RME) modal method, add Gaussian fluctuations normalized to the cavity noise power $k_B T_{\rm cav}\Delta\nu$, include the amplifier in the system temperature $T_{\rm sys}$, and present the network with the resulting complex voltage phasors. For $T_{\rm sys}=1.2$ K and 1000 averages, an input SNR of 0.036 yields 99.3% classification accuracy; the paper converts this accuracy into an equivalent Gaussian significance of about 2.7 standard deviations and, through the relation $T=({\rm SNR}_{\rm equiv}/{\rm SNR}_0)^2$ based on the Dicke radiometer scaling, obtains a time improvement factor of roughly $5\times 10^3$. The concrete consequence stated in the paper is that an experiment needing 100 days of integration to reach ${\rm SNR}=3$ could reach that same sensitivity in about half an hour with the neural network, and that slightly larger input SNRs push accuracy toward 100%.

Load-bearing premise

The entire improvement rests on treating the network's classification accuracy as a Gaussian detection significance — a 99.3% accuracy is taken to mean an equivalent signal-to-noise ratio of about 3 — because the claimed 5,000-fold time saving is just the square of that ratio divided by the input signal-to-noise ratio.

Editorial extensions

If this is right

  • A haloscope measurement point that needs 100 days of integration to reach ${\rm SNR}=3$ could instead reach that sensitivity in about half an hour with the trained network.
  • With the same exposure time as today, a scan would reach better sensitivity and probe lower values of the axion-photon coupling constant $g_{a\gamma\gamma}$.
  • The technique is simple enough to run in parallel with standard spectrum-processing pipelines in any current axion experiment, either as the primary readout or as a cross-check.
  • Accuracy approaches 100% for only slightly larger input SNR, so the benefit grows quickly as the input signal strengthens.
  • The same method transfers to other ultra-feeble-signal searches, notably high-frequency gravitational wave haloscopes, where exposure time is limited by the duration of the signal.

Reading between the lines

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

  • My reading is that the quoted 5,000-fold saving is a single-operating-point number: it is computed at ${\rm SNR}_0=0.036$, and it would shrink substantially if the accuracy-to-Gaussian mapping were replaced by directly measured false-alarm and false-dismissal rates, which the paper does not report.
  • The paper trains and tests on simulated noise only; a natural extension is to validate the single-neuron classifier on recorded off-resonance noise from a running haloscope, where drifts, standing waves, and non-stationary amplifier noise — effects the authors list as future work — would test whether the simulated accuracy survives contact with real data.
  • Because the network outputs a yes/no decision rather than an estimator, the equivalent-SNR construction silently assumes optimal decision thresholds; building the full receiver-operating-characteristic curve across SNR values would give a threshold-independent measure of how much integration time is actually saved.
  • If the method holds up on real noise, it effectively converts part of the exposure-time budget into offline training cost, so the fair comparison is not 100 days versus 30 minutes but 100 days versus 30 minutes plus the one-time cost of building a faithful noise simulator.
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

4 major / 3 minor

Summary. The manuscript proposes using a feedforward neural network to detect axion signals in simulated haloscope data, claiming a 5·10^3 reduction in the integration time needed to reach a given signal-to-noise ratio. The simulation chain combines the BI-RME3D modal method with generated resonant and amplifier noise, and the network performs a boolean classification ('axion present' or 'noise only'). The central quantitative claim is obtained from Eq. (7), which converts classifier accuracy into an equivalent Gaussian SNR and then into a time-improvement factor T. The paper reports accuracies for two system temperatures and several averaging levels, with the largest claimed improvement being a factor of a few thousand at SNR0 = 0.036.

Significance. If the claimed time improvement were physically valid, it would be a practically important result for axion haloscope searches, since it would dramatically shorten the exposure time needed to reach a given sensitivity. The direction of using simulation-trained classifiers is worth exploring, and the use of BI-RME3D to generate realistic signal-plus-noise samples is a reasonable starting point. However, the headline result depends entirely on Eq. (7), whose accuracy-to-SNR conversion is not a valid detection-statistics link. The paper provides no false-alarm rate, no detection-efficiency curve, and no comparison with a matched filter. The internal inconsistency in Table I—where an accuracy below chance still produces a large positive time improvement—shows that the claimed sensitivity gain is an artifact of the conversion rather than a physical effect. The manuscript therefore does not establish its main claim.

major comments (4)
  1. [Section III, Eq. (7) and Table I] The conversion of classifier accuracy to equivalent SNR is unjustified. SNRequiv is defined as the Gaussian quantile z_{(1+acc)/2}, but binary classification accuracy is a threshold-dependent summary that depends on the prior probability of the axion class and on the distribution of the classifier score. It does not, by itself, determine a detection sensitivity or a false-alarm rate. Without a derivation or an explicit link to detection statistics, T is not a physically meaningful time improvement.
  2. [Table I, row '1 average'] Table I contains an internal inconsistency: with SNR0 = 0.0011 and accuracy = 47.7% (worse than chance), Eq. (7) gives SNRequiv = 0.639 and T = 3.4e5. A classifier that performs below chance cannot improve sensitivity, so the accuracy-to-SNR mapping is not merely ad hoc but self-contradictory. This row shows that the time-improvement factor is an artifact of the conversion.
  3. [Section III, Eq. (7)] Because T = (SNRequiv/SNR0)^2, for any fixed accuracy T diverges as SNR0 tends to zero. The claimed factor of 5e3 is therefore not robust: the same formula yields T = 3.4e5 at SNR0 = 0.0011 with nearly chance accuracy, and arbitrarily large improvements could be manufactured by lowering the input SNR. This scaling undermines the significance of the reported numerical values.
  4. [Section III] The paper does not compare the neural network's performance with the optimal matched filter or with the standard radiometer analysis on the same simulated data. It also reports no false-alarm probability, no detection-efficiency curve, and no threshold analysis for the boolean decision. Without these elements, the equivalent SNR claimed for the network cannot be validated as a detection statistic.
minor comments (3)
  1. [Section II] The input representation given to the neural network is not described in enough detail: it is unclear whether the network receives time-domain samples, power spectra, or complex phasors, and at what resolution. Specifying this would improve reproducibility.
  2. [Figure 5 and Table I] The accuracy values are reported as single numbers without error bars or multiple training seeds. Since the accuracy is based on 1000 tests, binomial uncertainties are non-negligible, and the propagated uncertainty on T would be helpful.
  3. [Section III] The text says 'with a SNR = 0.036 the accuracy of the neural network is 99.3%, close to an equivalent SNR of 3', but the corresponding value is 2.697, and the paper should state whether this is considered acceptable and why.

Circularity Check

1 steps flagged · score 8.0 of 10

The headline 5e3 time improvement is defined, not derived: Eq. (7) maps measured network accuracy to SNR_equiv via a Gaussian quantile, making T a transformed measurement.

  1. self definitional [Section III, Eq. (7) and Table I]
    "time improvement T is calculated as the ratio between the time required to reach the equivalent SNR without the neural network, tequiv, and the original SNR of the signal introduced to the neural network, t0: T = tequiv/t0 = (SNRequiv/SNR0)^2, where SNRequiv is calculated as the number of standard deviations in a Gaussian distribution N(0,1) that correspond to the associated neural network accuracy."

    The central claim of a 5e3 integration-time improvement reduces by construction to the measured classifier accuracy. Eq. (7) defines T as (z_(1+acc)/2 / SNR0)^2, so once acc is measured on simulated test data, T is a deterministic transform of that accuracy with no independent detection-statistics content. No false-alarm rate, ROC curve, or matched-filter sensitivity is derived to establish that accuracy corresponds to a physical SNR. The conversion is also internally inconsistent: Table I assigns SNRequiv=0.639 and T=3.4e5 to acc=47.7% (worse than chance), showing the time gain is produced by the hand-chosen Gaussian mapping rather than by any demonstrated sensitivity improvement. The 'prediction' is therefore equivalent to its input.

full rationale

The paper's physical forward model (BI-RME3D axion current, resonant/amplifier noise, Dicke SNR) and the FNN accuracy measurements are independent pieces of work, and I find no load-bearing circularity through self-citation: [13], [29], and the BI-RME references are used as standard methods/prior work and do not by themselves force the conclusion. However, the paper's headline result is not an independently derived sensitivity limit. Eq. (7) defines the time improvement T in terms of an 'equivalent SNR' that is itself defined as the Gaussian quantile of the network's classification accuracy. The claimed factor of 5e3 = (2.697/0.036)^2 is therefore just a rearrangement of the measured 99.3% accuracy. The paper itself notes the practical caveat that training from real measurements would consume the time saved, but that does not address the more basic issue: the SNR-equivalence mapping is adopted by definition rather than derived from detection statistics. Because the main quantitative claim is forced by this definitional conversion, the circularity score is high even though the simulations and classifier are not themselves circular.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new free numeric parameters fitted to real data; the simulation uses standard physical constants (gaγγ, ρDM, Q0, Tsys) and simulation choices (number of averages, number of training cases). The central claim rests on an ad hoc accuracy-to-SNR conversion, listed as the first axiom. The Gaussian noise model and the transferability of the network to real data are additional domain assumptions. No new particles or entities are introduced.

assumptions (5)
  • ad hoc to paper The accuracy of the binary classifier can be converted to an equivalent SNR by treating accuracy as a central Gaussian probability: SNRequiv = z_{(1+acc)/2}.
    Section III defines SNRequiv from accuracy and Eq. 7 uses it to compute T; this conversion has no detection-theoretic derivation and is the main driver of the claimed 5e3 improvement.
  • domain assumption Thermal noise from the cavity and first amplifier is fully described by Gaussian random current phasors normalized to kB Tcav Δν.
    Section II; a standard model, but not validated against real haloscope noise or systematics.
  • domain assumption The BI-RME3D modal method with finite mode truncation yields exact values of the axion-induced port voltage.
    Section II; based on references [12, 28, 29] and prior work by the same group.
  • domain assumption A neural network trained on simulated noise generalizes to real experimental conditions without performance loss.
    The conclusions state that training with precise simulations is required, but no transfer evidence is given.
  • domain assumption The axion signal follows the shifted Maxwell-Boltzmann frequency distribution with Qa = 10^6.
    Section II; standard halo model assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dark Matter Axion Detection with Neural Networks at Ultra-Low Signal-to-Noise Ratio." pith.science (2026). https://pith.science/paper/XG5UBTKZ

@misc{pith2026241117947,
  author       = {Pith},
  title        = {Pith review of: Dark Matter Axion Detection with Neural Networks at Ultra-Low Signal-to-Noise Ratio},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XG5UBTKZ}},
  note         = {Machine review of arXiv:2411.17947}
}
abstract

We present the first analysis of Dark Matter axion detection applying neural networks for the improvement of sensitivity. The main sources of thermal noise from a typical read-out chain are simulated, constituted by resonant and amplifier noises. With this purpose, an advanced modal method employed in electromagnetic modal analysis for the design of complex microwave circuits is applied. A feedforward neural network is used for a boolean decision (there is axion or only noise), and robust results are obtained: the neural network can improve by a factor of $5\cdot 10^{3}$ the integration time needed to reach a given signal to noise ratio. This could either significantly reduce measurement times or achieve better sensitivities with the same exposure durations.

Figures

Figures reproduced from arXiv: 2411.17947 by the authors.

Figure 1
Figure 1. FIG. 1. Readout chain employed for the axion detection [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Resonant noise power extracted from the cavity at a [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 5
Figure 5. FIG. 5. Neural network accuracy for given values of SNR. [PITH_FULL_IMAGE:figures/full_fig_p004_5.png] view at source ↗
Figures from the paper (1 more)
Figure 6
Figure 6. Figure 6: FIG. 6. Accuracy improvement for different number of train [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 14 canonical work pages

  1. [1]

    R. D. Peccei and H. R. Quinn, CP Conservation in the Presence of Pseudoparticles, Phys. Rev. Lett. 38, 1440 (1977). 6

  2. [2]

    J. E. Kim and G. Carosi, Axions and the Strong CP Problem, Rev. Mod. Phys. 82, 557 (2010), [Erratum: Rev.Mod.Phys. 91, 049902 (2019)], arXiv:0807.3125 [hep- ph]

  3. [3]

    Weinberg, A New Light Boson?, Phys

    S. Weinberg, A New Light Boson?, Phys. Rev. Lett. 40, 223 (1978)

  4. [4]

    Wilczek, Problem of Strong P and T Invariance in the Presence of Instantons, Phys

    F. Wilczek, Problem of Strong P and T Invariance in the Presence of Instantons, Phys. Rev. Lett. 40, 279 (1978)

  5. [5]

    Preskill, M

    J. Preskill, M. B. Wise, and F. Wilczek, Cosmology of the Invisible Axion, Phys. Lett. B 120, 127 (1983)

  6. [6]

    L. F. Abbott and P. Sikivie, A Cosmological Bound on the Invisible Axion, Phys. Lett. B 120, 133 (1983)

  7. [7]

    Dine and W

    M. Dine and W. Fischler, The Not So Harmless Axion, Phys. Lett. B 120, 137 (1983)

  8. [8]

    Primakoff, Photoproduction of neutral mesons in nu- clear electric fields and the mean life of the neutral meson, Phys

    H. Primakoff, Photoproduction of neutral mesons in nu- clear electric fields and the mean life of the neutral meson, Phys. Rev. 81, 899 (1951)

Show all 34 references
  1. [9]

    invisible

    P. Sikivie, Detection rates for “invisible”-axion searches, Phys. Rev. D 32, 2988 (1985)

  2. [10]

    Sikivie, Experimental Tests of the Invisible Ax- ion, Phys

    P. Sikivie, Experimental Tests of the Invisible Ax- ion, Phys. Rev. Lett. 51, 1415 (1983), [Erratum: Phys.Rev.Lett. 52, 695 (1984)]

  3. [11]

    Ahyoune et al

    S. Ahyoune et al. , RADES axion search results with a High-Temperature Superconducting cavity in an 11.7 T magnet (2024), arXiv:2403.07790 [hep-ex]

  4. [12]

    Arcioni, M

    P. Arcioni, M. Bozzi, M. Bressan, G. Conciauro, and L. Perregrini, The BI-RME method: An historical overview, in 2014 International Conference on Numer- ical Electromagnetic Modeling and Optimization for RF, Microwave, and Terahertz Applications (NEMO) (2014) pp. 1–4

  5. [13]

    F. Lacy, A. Ruiz-Reyes, and A. Brescia, Machine learning for low signal-to-noise ratio detection, Pattern Recogni- tion Letters 179, 115 (2024)

  6. [14]

    A. K. Yi et al. , Axion Dark Matter Search around 4.55 µeV with Dine-Fischler-Srednicki-Zhitnitskii Sensitivity, Phys. Rev. Lett. 130, 071002 (2023), arXiv:2210.10961 [hep-ex]

  7. [15]

    J. Choi, S. Ahn, B. Ko, S. Lee, and Y. Semertzidis, CAPP-8TB: Axion dark matter search experiment around 6.7 µeV, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrome- ters, Detectors and Associated Equipment 1013, 165667 (2021)

  8. [16]

    J. Kim, O. Kwon, i. m. c. b. u. Kutlu, W. Chung, A. Matlashov, S. Uchaikin, A. F. van Loo, Y. Nakamura, S. Oh, H. Byun, D. Ahn, and Y. K. Semertzidis, Near- Quantum-Noise Axion Dark Matter Search at CAPP around 9.5 µeV, Phys. Rev. Lett. 130, 091602 (2023)

  9. [17]

    O. Kwon, D. Lee, W. Chung, D. Ahn, H. Byun, F. Caspers, H. Choi, J. Choi, Y. Chong, H. Jeong, J. Jeong, J. E. Kim, J. Kim, i. m. c. b. u. Kutlu, J. Lee, M. Lee, S. Lee, A. Matlashov, S. Oh, S. Park, S. Uchaikin, S. Youn, and Y. K. Semertzidis, First Results from an Axion Halos...

  10. [18]

    K. M. Backes, D. A. Palken, S. A. Kenany, et al. , A quantum enhanced search for dark matter axions, Nature 590, 238 (2021)

  11. [19]

    W. U. Wuensch, S. De Panfilis-Wuensch, Y. K. Se- mertzidis, J. T. Rogers, A. C. Melissinos, H. J. Halama, B. E. Moskowitz, A. G. Prodell, W. B. Fowler, and F. A. Nezrick, Results of a laboratory search for cosmic axions and other weakly coupled light particles, Phys. Rev. D 40...

  12. [20]

    Chang, J.-Y

    H. Chang, J.-Y. Chang, Y.-C. Chang, Y.-H. Chang, Y.- H. Chang, C.-H. Chen, C.-F. Chen, K.-Y. Chen, Y.-F. Chen, W.-Y. Chiang, W.-C. Chien, H. T. Doan, W.-C. Hung, W. Kuo, S.-B. Lai, H.-W. Liu, M.-W. OuYang, P.-I. Wu, and S.-S. Yu (TASEH Collaboration), First Re- sults from the ...

  13. [21]

    Hagmann, P

    C. Hagmann, P. Sikivie, N. S. Sullivan, and D. B. Tanner, Results from a search for cosmic axions, Phys. Rev. D42, 1297 (1990)

  14. [22]

    ERDF/EU”. J. Reina-Valero has the support of “Plan de Recuperaci´ on, Transformaci´ on y Resiliencia (PRTR) 2022 (ASF AE/2022/013)

    has been employed. The unloaded quality factor of the cavity is Q0 = 1.1 · 105 for a cryogenic electrical con- ductivity σ = 1 · 109 S/m. A 50 Ω coaxial probe has been connected to the cavity with a relative permittivity of εr = 1. The probe was introduced with a determined le...

  15. [23]

    CST AG, CST Studio Suite (2024), 2024 version

  16. [24]

    J. E. Kim, Weak-Interaction Singlet and Strong CP In- variance, Phys. Rev. Lett. 43, 103 (1979)

  17. [25]

    M. A. Shifman, A. I. Vainshtein, and V. I. Zakharov, Can Confinement Ensure Natural CP Invariance of Strong In- teractions?, Nucl. Phys. B 166, 493 (1980)

  18. [26]

    Y. Kim, D. Kim, J. Jeong, J. Kim, Y. C. Shin, and Y. K. Semertzidis, Effective approximation of electro- magnetism for axion haloscope searches, Physics of the Dark Universe 26, 100362 (2019)

  19. [27]

    M. S. Turner, Periodic signatures for the detection of cosmic axions, Phys. Rev. D 42, 3572 (1990)

  20. [28]

    B. M. Brubaker, L. Zhong, S. K. Lamoreaux, K. W. Lehnert, and K. A. van Bibber, HAYSTAC axion search analysis procedure, Phys. Rev. D 96, 123008 (2017)

  21. [29]

    R. S. Giuseppe Conciauro, Marco Guglielmi, Advanced Modal Analysis (Wiley, 1999)

  22. [30]

    Navarro, B

    P. Navarro, B. Gimeno, A. ´Alvarez Melc´ on, S. A. Cuendis, C. Cogollos, A. D ´ ıaz-Morcillo, J. Gallego, J. G. Barcel´ o, J. Golm, I. Irastorza, A. L. Guerrero, and C. P. Garay, Wide-band full-wave electromagnetic modal anal- ysis of the coupling between dark-matter axions an...

  23. [31]

    R. H. Dicke, The Measurement of Thermal Radiation at Microwave Frequencies, Rev. Sci. Instrum. 17, 268 (1946)

  24. [32]

    Berlin, D

    A. Berlin, D. Blas, R. T. D’Agnolo, S. A. R. Ellis, R. Harnik, Y. Kahn, J. Sch¨ utte-Engel, and M. Wentzel, Electromagnetic cavities as mechanical bars for gravita- tional waves, Phys. Rev. D 108, 084058 (2023)

  25. [33]

    Gatti, L

    C. Gatti, L. Visinelli, and M. Zantedeschi, Cavity detec- tion of gravitational waves: Where do we stand?, Phys. Rev. D 110, 023018 (2024)

  26. [34]

    J. R. Valero, J. R. N. Madrid, D. Blas, A. D. Morcillo, I. G. Irastorza, B. Gimeno, and J. M. Cabrera, High- frequency gravitational waves detection with the BabyI- AXO haloscopes (2024), arXiv:2407.20482 [hep-ex]

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.