Pith. sign in

REVIEW 3 major objections 5 minor 66 references

This paper claims that a Viterbi-based search pipeline with an f^{-8/3} time-frequency remapping can detect the long inspiral signals of planetary-mass black-hole binaries in real Hanford O3 data, recovering 95% of injected signals out to d

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 · deepseek-v4-flash

2026-08-01 17:11 UTC pith:CWPWMEPQ

load-bearing objection Useful pipeline paper whose headline distance-reach numbers are in-sample estimates; the sensitivity claim needs a held-out injection set before it can be taken at face value. the 3 major comments →

arxiv 2607.18352 v1 pith:CWPWMEPQ submitted 2026-07-20 astro-ph.IM gr-qc

Search for Planetary-mass Black Holes with an Improved Viterbi Algorithm

Raul Rodriguez , George Alestas , Sachiko Kuroyanagi , Juan Garcia-Bellido This is my paper
classification astro-ph.IM gr-qc
keywords gravitational wavesprimordial black holesplanetary-mass black holesViterbi algorithmhidden Markov modellong-duration inspiralcontinuous-wave searchtime-frequency analysis
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper sets out to detect gravitational waves from binaries of primordial black holes in the planetary-mass range (10^-4 to 10^-1 solar masses), whose inspiral signals last months or longer and are too long for matched-filter searches. The authors build a complete search pipeline around the Viterbi algorithm, which tracks the most probable frequency path without a template bank. To make the search work, they introduce a time-frequency coordinate transformation (f^{-8/3}) that turns the chirp into a straight line, a candidate-isolation procedure, and two complementary detection statistics: n_sigma for power significance and NMSE for consistency with inspiral morphology. Testing on roughly 1000 hours of O3 Hanford data with about 600 injected signals, the pipeline recovers most signals at a fixed false-alarm ratio of 3%, reaching 95% recovery out to 50-135 kpc across most of the mass range. If valid, the method opens the long-duration inspiral window for planetary-mass PBH searches, potentially enabling a discovery that would strongly support the primordial-black-hole hypothesis.

Core claim

The central claim is that a fully operational, template-free search pipeline based on the Viterbi algorithm can detect the long inspiral signals of planetary-mass black-hole binaries in real gravitational-wave data. The key move is to remap the time-frequency plane from f to f^{-8/3}, under which the inspiral track becomes a straight line whose slope is determined solely by the chirp mass. This lets the transition matrix encode a physical preference for slowly-evolving tracks and lets a normalized mean-square-error fit test whether the recovered track matches an inspiral morphology. Validated on ~1000 hours of O3b Hanford data with ~600 injections, the pipeline achieves d_L,95% above 50 kpc

What carries the argument

The central machinery is the combination of three elements: (1) the coordinate transformation f -> f^{-8/3}, which converts the quasi-Newtonian inspiral frequency evolution f_gw(t) ∝ (t_coal - t)^{-3/8} into a straight line with slope K M^{5/3}, so that the chirp mass becomes a single slope parameter; (2) the Viterbi dynamic-programming algorithm over a hidden Markov model, with a transition matrix restricted to center-and-downward frequency jumps, which reconstructs the most probable track without a template bank; and (3) the two detection statistics n_sigma (power significance relative to the noise-only Viterbi maximum) and NMSE (normalized mean-square error of the candidate track against

Load-bearing premise

The detection threshold in the (n_sigma, NMSE) plane was tuned on the same set of ~600 injected signals whose recovery rate is then used to measure the pipeline's distance reach, so the reported 95% recovery distances assume that this tuning is not overtuned to those specific injections or to the favorable face-on, equal-mass, circular configurations used.

What would settle it

Run the same pipeline on a held-out injection population that is not used to optimize the decision boundary, using varied inclinations (with i > 0), eccentric orbits, and realistic sky positions, and compare the resulting d_L,95% curve; if the recovery fraction at the quoted distances drops substantially below 95%, the claimed reach is not representative of a blind search. Alternatively, apply the search to a long stretch of un-injected data with the boundary fixed and check that the observed false-alarm rate matches the nominal 3%.

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

If this is right

  • Enables a search of the intermediate mass window 10^-4 to 10^-1 solar masses, which is inaccessible to both matched-filter CBC searches and standard continuous-wave searches due to spin-up limits.
  • Provides template-free detection and a quick chirp-mass estimate; if a candidate is found, its chirp mass can be characterized almost immediately.
  • The method is signal-agnostic beyond the leading-order inspiral, so it may remain sensitive to higher-order corrections, eccentricity, or environmental effects.
  • With a fixed 3% false-alarm ratio, the reported recovery distances (d_L,95% up to 135 kpc) quantify the Galactic search volume where a single detection could confirm planetary-mass PBHs.

Where Pith is reading between the lines

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

  • If the pipeline is ported to multi-detector networks, the uncorrelated noise between instruments could suppress glitch triggers, potentially extending the reach beyond the single-detector distance limits reported here.
  • The favorable injection choices (face-on, equal-mass, circular orbits) mean the true blind-search sensitivity is likely lower for eccentric or inclined systems; testing against an eccentric injection population, which the authors note as future work, would quantify the degradation.
  • Because the NMSE-based chirp-mass estimate remains accurate slightly beyond the 95% recovery distance, the statistic might be usable as a follow-up sorter for sub-threshold candidates once a more robust background is characterized.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper presents a search pipeline for gravitational waves from planetary-mass primordial black hole inspirals in the mass range 10^-4 to 10^-1 Msun, based on the Viterbi algorithm applied to a remapped time-frequency representation (t, f^{-8/3}) in which chirps become straight lines. The pipeline includes a candidate-isolation procedure and ranks candidates using two statistics: n_sigma (excess Viterbi-track power over noise) and NMSE (morphological consistency with an inspiral). Validation is performed by injecting ~600 TaylorT3 signals into ~1000 hours of O3b LIGO Hanford data. The authors claim that, at a fixed false-alarm ratio of 3%, the pipeline recovers most injections, reaches d_L,95% up to 135 kpc near M ~ 2e-2 Msun, and yields accurate chirp-mass estimates via NMSE minimization. The code is publicly available.

Significance. If the quantitative sensitivity claims hold, the paper would fill a relevant gap between standard CBC searches and continuous-wave searches, and the f^{-8/3} remapping plus the two-statistic ranking are useful methodological ideas. The use of real O3 data, the publicly available implementation, and the explicit reporting of a distance-reach curve are strengths. However, the central distance-reach claim is compromised by an in-sample threshold optimization: the polynomial decision boundary is tuned on the same injection population used to measure d_L,95%. This makes the headline sensitivity numbers optimistic to an unknown degree. The paper is a promising methodological contribution but requires out-of-sample validation before its quantitative sensitivity claims can be accepted.

major comments (3)
  1. [Sec. III.B, Figs. 5 and 6] The polynomial decision boundary in the (n_sigma, NMSE) plane is 'optimized to maximize the number of true positives at fixed FAR' using the same ~600-injection population from which the d_L,95% curve is computed. No held-out injection set, cross-validation, or independent noise-realization split is described. This makes the distance-reach curve an in-sample fitted quantity, directly affecting the abstract claim of recovering most signals with 95% efficiency out to ~135 kpc. The authors should provide an out-of-sample validation, e.g., optimize the boundary on one half of the injections and evaluate on the other half, or use k-fold cross-validation, and report the resulting change in d_L,95%.
  2. [Table I] All injections use inclination iota=0 (face-on), mass ratio q=1, and circular non-spinning orbits. Face-on orientation maximizes the detector response, so the reported d_L,95% values are an upper envelope rather than a representative sensitivity for an isotropic population of PBH binaries. The authors should state this limitation explicitly or, preferably, include at least one additional injection set with iota sampled from an isotropic distribution and/or q<1 to quantify the degradation in distance reach.
  3. [Sec. III.B, threshold calibration paragraph] The 3% false-alarm ratio is calibrated using noise triggers from the same ~1000 h O3b dataset used for the signal-injected runs. The authors themselves state that 'Future work will require a more extensive background characterization,' but the paper does not report the number of independent background triggers or the stability of the fitted polynomial boundary across noise realizations. This matters because the FAR value is a central output of the pipeline; without a measure of its uncertainty, the quoted 3% is not yet a robust statistical claim. Please report the number of noise triggers and demonstrate boundary stability, for example by dividing the background into independent chunks and recomputing the boundary.
minor comments (5)
  1. [Appendix A, near Eq. (A5)] The text says 'This quantity is used in Fig. 8 to characterize the injected population,' but Fig. 8 shows the noise PSD, not the SNR distribution. The reference is likely to Fig. 5 or a missing panel; please correct.
  2. [Sec. II B] The transition matrix is first described as allowing up/center/down (UCD) transitions, but Sec. III A says the prior favors center/down (CD) transitions. Please clarify which transitions are actually allowed in the implemented prior and how the first state is initialized.
  3. [Sec. II D, Eq. (22)] The NMSE normalization by sum of squared track amplitudes can become unstable if the candidate segment crosses zero in the remapped coordinate. Consider adding a small regularization or specifying the behavior in that case.
  4. [References] Reference [57] cites 'Slade, Semantic Scholar' without a title or venue. Please provide a full bibliographic reference for the Hidden Markov Model / Viterbi background.
  5. [Fig. 6 caption] The caption says 'we can recover at least the 95% of the signals within the blue contour,' but the figure shows a distance-reach curve rather than a two-dimensional contour. Please rephrase to match the actual plotted quantity.

Circularity Check

1 steps flagged

Distance-reach claim is in-sample: the polynomial decision boundary is optimized on the same ~600 injections used to measure d_L,95%, with no held-out set.

specific steps
  1. fitted input called prediction [Sec. III.B, Figs. 5 and 6 (pages 8-9)]
    "we construct a polynomial decision boundary in the detection-statistic plane (solid black line in Fig. 5), optimized to maximize the number of true positives at fixed F AR. ... In Fig. 6, we show the distance reach as a function of chirp mass. In particular, we report the dL,95% curve, defined as the luminosity distance up to which signals are recovered in at least 95% of the injections."

    The same ~600-injection population (Table I, Sec. III.B) is used both to optimize the polynomial decision boundary (maximizing true positives at fixed 3% FAR) and then to measure the d_L,95% distance-reach curve. No held-out injection set or cross-validation is described. Consequently, the reported 95% recovery distances are in-sample evaluations of a boundary tuned on those very injections, not independent predictions of blind-search sensitivity. The optimization target (maximize true positives) directly inflates the subsequent recovery fraction. The paper itself notes that 'Future work will require a more extensive background characterization', indicating that the calibration is provisional.

full rationale

The main circularity is localized to the sensitivity claim. The decision boundary in the (n_sigma, NMSE) plane is explicitly 'optimized to maximize the number of true positives at fixed FAR' using the same injected-signal population from which d_L,95% is then computed; with no held-out injections, the distance-reach curve is an in-sample fitted quantity rather than an out-of-sample prediction. Other components are not circular: the Viterbi/n_sigma background is characterized on noise-only data, the NMSE chirp-mass estimate is a direct least-squares fit to the track and is evaluated on injections, and the use of Ref. [52] for the optimal SFT duration and n_sigma is a normal citation to prior work rather than a self-referential uniqueness claim. The paper is self-contained enough that the flaw is not definitional equivalence, but the central performance figure does reduce, by the paper's own procedure, to a parameter tuned on the same data used to measure it.

Axiom & Free-Parameter Ledger

4 free parameters · 5 axioms · 0 invented entities

The central sensitivity claim rests on the standard inspiral frequency evolution (Eq. 1), the assumption that the signal stays in one frequency bin per SFT (Eq. 12), the noise background being stationary enough to calibrate n_sigma, and the injected population (face-on, equal-mass, circular, non-spinning TaylorT3 waveforms) being representative. The main hand-fitted parameters are the decision-boundary coefficients and, to a lesser extent, the isolation hyperparameters and the CD transition prior.

free parameters (4)
  • decision boundary polynomial coefficients = not reported; chosen to maximize true positives at FAR=3%
    Optimized on the injected population in Sec. III.B; directly determines reported recovery rates and d_L,95%.
  • candidate-isolation hyperparameters = 8 windows; retain top 2; iterative expansion
    Hand-chosen design choices in Sec. III.A that affect which segment of the Viterbi track is scored; not fitted but influence sensitivity.
  • Viterbi transition prior = center/down (CD), one-bin jumps
    Chosen in Sec. II.B/III.A to favor inspirals; disallows upward frequency motion and larger steps; affects recovery of fast-evolving high-mass signals.
  • SNR-loss criterion for TSFT selection = 1%
    Used in Sec. III.A to select the 13 coherence times; a different loss would change the TSFT set and the number of SFTs.
axioms (5)
  • domain assumption Leading-order quadrupole circular-orbit inspiral evolution (Eq. 1) is valid over the whole frequency band; higher-order PN terms are not needed.
    Sec. II.A states ISCO lies outside the band for these masses; assumes zero eccentricity and no spin effects.
  • domain assumption Signal frequency stays within a single frequency bin during each SFT for f <= f* (Eq. 12).
    Sec. II.C; if the frequency drifts within a segment, the Viterbi input SNR degrades.
  • domain assumption Noise-only Viterbi background is stationary and can be estimated from raw O3b data to define mu and sigma in Eq. (17).
    Sec. II.D/III.B; real data contains glitches and lines, but the noise-trigger distribution is treated as a stable background.
  • domain assumption Injected TaylorT3 3.5PN waveforms (circular, non-spinning, q=1, iota=0) are representative of real planetary-mass PBH inspirals.
    Table I; eccentricity and inclination are not injected, and the authors note eccentricity is a likely property of PBH binaries (Sec. IV).
  • domain assumption Doppler modulations are negligible over the short TSFT values used, so the track is linear in f^{-8/3} (Eq. 21).
    Sec. II.D; if Doppler shifts were significant, the NMSE straight-line fit would be biased.

pith-pipeline@v1.3.0-alltime-deepseek · 17015 in / 15790 out tokens · 150462 ms · 2026-08-01T17:11:56.301960+00:00 · methodology

0 comments
read the original abstract

Primordial black holes in the planetary-mass range have attracted renewed interest; however, the search for gravitational waves from such binaries remains challenging due to their long-lived nature. In this work, we present, define, and validate a fully operational search pipeline developed to detect planetary-mass binaries during their inspiral phase. We use the Viterbi algorithm, a dynamic programming technique that recovers the most likely track based on a Hidden Markov Model. To enhance its performance, we introduce a novel time-frequency representation of the data and a candidate isolation procedure that separates signals from background noise. The evaluation of candidates is carried out using the two detection statistics, $n_{\sigma}$ and NMSE, which quantify the power significance and the consistency with the expected binary evolution. We then validate the search method using O3 LIGO Hanford data with a population of injected signals. The pipeline is able to recover most of the signals with a fixed false-alarm ratio of $3\%$, covering Galactic scales across most of the parameter space and reaching luminosity distances $ \gtrsim 100$ kpc in the most sensitive region. For each candidate, we also obtain an estimate of the system's chirp mass, whose accuracy remains high throughout the detectable range, enabling a rapid characterization of the system upon detection.

Figures

Figures reproduced from arXiv: 2607.18352 by George Alestas, Juan Garcia-Bellido, Raul Rodriguez, Sachiko Kuroyanagi.

Figure 1
Figure 1. Figure 1: FIG. 1. Time to coalescence as a function of the chirp mass [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Time-frequency representation of O3b LIGO Hanford [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Evolution of the detection statistic [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Phases of the search pipeline applied to an injected long-inspiral signal. From top left to bottom right, the figure [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Triggers distribution for the search in O3b Hanford [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Distance reach curve corresponding to [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7. Relative error of the chirp mass estimation obtained [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8. Average noise power spectral density obtained from [PITH_FULL_IMAGE:figures/full_fig_p012_8.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

66 extracted references · 1 canonical work pages

  1. [1]

    B. P. Abbottet al.(LIGO Scientific, Virgo), Phys. Rev. Lett.116, 061102 (2016)

    2016

  2. [2]

    B. P. Abbottet al.(LIGO Scientific, Virgo), Phys. Rev. X9, 031040 (2019)

    2019

  3. [3]

    Abbottet al.(LIGO Scientific, Virgo), Phys

    R. Abbottet al.(LIGO Scientific, Virgo), Phys. Rev. X 11, 021053 (2021)

    2021

  4. [4]

    The LIGO Scientific Collaboration and the Virgo Collab- oration, Phys. Rev. D109, 022001 (2024)

    2024

  5. [5]

    LIGO Scientific Collaboration, Virgo Collaboration, and KAGRA Collaboration, Phys. Rev. X13, 041039 (2023)

    2023

  6. [6]

    The LIGO Scientific Collaboration, the Virgo Collabora- tion, and the KAGRA Collaboration, The Astrophysical Journal Letters1004, L22 (2026)

    2026

  7. [7]

    A. G. Abacet al.(LIGO Scientific, VIRGO, KAGRA), arXiv:2605.27225 [gr-qc] (2026), arXiv e-print

    Pith/arXiv arXiv 2026

  8. [8]

    Hawking, Mon

    S. Hawking, Mon. Not. Roy. Astron. Soc.152, 75 (1971)

    1971

  9. [9]

    B. J. Carr and S. W. Hawking, Mon. Not. Roy. Astron. Soc.168, 399 (1974)

    1974

  10. [10]

    Garc ´ ıa-Bellido, A

    J. Garc ´ ıa-Bellido, A. D. Linde, and D. Wands, Phys. Rev. D54, 6040 (1996)

    1996

  11. [11]

    B. Carr, F. Kuhnel, and M. Sandstad, Phys. Rev. D94, 083504 (2016)

    2016

  12. [12]

    A. M. Green, arXiv:2402.15211 [astro-ph.CO] (2024), arXiv e-print

    Pith/arXiv arXiv 2024

  13. [13]

    B. Carr, K. Kohri, Y. Sendouda, and J. Yokoyama, Rept. Prog. Phys.84, 116902 (2021)

    2021

  14. [14]

    B. Carr, S. Clesse, J. Garc ´ ıa-Bellido, M. Hawkins, and F. Kuhnel, Phys. Rept.1054, 1 (2024)

    2024

  15. [15]

    Clesse and J

    S. Clesse and J. Garc ´ ıa-Bellido, Phys. Dark Univ.15, 142 (2017)

    2017

  16. [16]

    Garc ´ ıa-Bellido, B

    J. Garc ´ ıa-Bellido, B. Carr, and S. Clesse, Universe8, 12 (2021)

    2021

  17. [17]

    Miller, T

    S. Miller, T. A. Callister, and W. M. Farr, The Astro- physical Journal895, 128 (2020)

    2020

  18. [18]

    Abbottet al.(LIGO Scientific, VIRGO), arXiv:2108.01045 [gr-qc] (2021), arXiv e-print

    R. Abbottet al.(LIGO Scientific, VIRGO), arXiv:2108.01045 [gr-qc] (2021), arXiv e-print

    Pith/arXiv arXiv 2021

  19. [19]

    Abbottet al.(KAGRA, VIRGO, LIGO Scientific), Phys

    R. Abbottet al.(KAGRA, VIRGO, LIGO Scientific), Phys. Rev. X13, 011048 (2023)

    2023

  20. [20]

    Morr´ aset al., Phys

    G. Morr´ aset al., Phys. Dark Univ.42, 101285 (2023)

    2023

  21. [21]

    Prunier, G

    M. Prunier, G. Morr´ as, J. F. N. Siles, S. Clesse, J. Garc ´ ıa- Bellido, and E. Ruiz Morales, Phys. Dark Univ.46, 101582 (2024)

    2024

  22. [22]

    Magee, A.-S

    R. Magee, A.-S. Deutsch, P. McClincy, C. Hanna, C. Horst, D. Meacher, C. Messick, S. Shandera, and M. Wade, Phys. Rev. D98, 103024 (2018)

    2018

  23. [23]

    The LIGO Scientific Collaboration and the Virgo Collab- oration and the KAGRA Collaboration, arXiv:2605.05444 [astro-ph.HE] (2026), arXiv e-print

    Pith/arXiv arXiv 2026

  24. [24]

    A. L. Miller, N. Aggarwal, S. Clesse, and F. De Lillo, Phys. Rev. D105, 062008 (2022)

    2022

  25. [25]

    Abbottet al.(KAGRA, LIGO Scientific, VIRGO), Phys

    R. Abbottet al.(KAGRA, LIGO Scientific, VIRGO), Phys. Rev. D106, 102008 (2022)

    2022

  26. [26]

    A. G. Abacet al.(LIGO Scientific, VIRGO, KAGRA), arXiv:2603.14168 [gr-qc] (2026), arXiv e-print

    arXiv 2026

  27. [27]

    Niikura, M

    H. Niikura, M. Takada, S. Yokoyama, T. Sumi, and S. Masaki, Phys. Rev. D99, 083503 (2019)

    2019

  28. [28]

    Sugiyama, M

    S. Sugiyama, M. Takada, N. Yasuda, and N. Tominaga, arXiv:2602.05840 [astro-ph.CO] (2026), arXiv e-print

    arXiv 2026

  29. [29]

    Collaboration, The Astrophysical Journal Letters951, L8 (2023)

    N. Collaboration, The Astrophysical Journal Letters951, L8 (2023)

    2023

  30. [30]

    Dom` enech and S

    G. Dom` enech and S. Pi, Science China Physics, Mechanics & Astronomy65, 10.1007/s11433-021-1839-6 (2022)

    work page doi:10.1007/s11433-021-1839-6 2022

  31. [31]

    Dom` enech, S

    G. Dom` enech, S. Pi, and A. Wang, arXiv:2602.24061 [astro-ph.CO] (2026), arXiv e-print

    arXiv 2026

  32. [32]

    Tisserandet al., Astronomy & Astrophysics469, 387–404 (2007)

    P. Tisserandet al., Astronomy & Astrophysics469, 387–404 (2007)

    2007

  33. [33]

    Wyrzykowski, J

    L. Wyrzykowski, J. Skowron, S. Koz lowski, A. Udalski, M. K. Szyma´ nski, M. Kubiak, G. Pietrzy´ nski, I. Soszy´ nski, O. Szewczyk, K. Ulaczyk, R. Poleski, and P. Tisserand, Monthly Notices of the Royal Astronomical Society416, 2949–2961 (2011)

    2011

  34. [34]

    B. Carr, A. J. Iovino, G. Perna, V. Vaskonen, and H. Veerm¨ ae, La Rivista del Nuovo Cimento49, 225–274 (2026)

    2026

  35. [35]

    Garc ´ ıa-Bellido and S

    J. Garc ´ ıa-Bellido and S. Clesse, Phys. Dark Univ.19, 144 (2018)

    2018

  36. [36]

    Gorton and A

    M. Gorton and A. M. Green, JCAP08(08), 035, arXiv:2203.04209 [astro-ph.CO]

    Pith/arXiv arXiv

  37. [37]

    M. R. S. Hawkins and J. Garc ´ ıa-Bellido, Mon. Not. Roy. Astron. Soc.544, 1950 (2025), arXiv:2509.05400 [astro- ph.GA]

    arXiv 1950

  38. [38]

    A. L. Miller, S. Clesse, F. De Lillo, G. Bruno, A. Depasse, and A. Tanasijczuk, Phys. Dark Univ.32, 100836 (2021)

    2021

  39. [39]

    A. L. Miller, N. Aggarwal, S. Clesse, F. De Lillo, S. Sachdev, P. Astone, C. Palomba, O. J. Piccinni, and L. Pierini, Phys. Rev. D110, 082004 (2024)

    2024

  40. [40]

    A. L. Miller, N. Aggarwal, S. Clesse, F. De Lillo, S. Sachdev, P. Astone, C. Palomba, O. J. Piccinni, and L. Pierini, Phys. Rev. Lett.133, 111401 (2024)

    2024

  41. [41]

    Sajith Menonet al., Phys

    S. Sajith Menonet al., Phys. Rev. D113, 123042 (2026)

    2026

  42. [42]

    A. L. Miller and L. Pierini, arXiv:2512.10539 [gr-qc] (2025), arXiv e-print

    Pith/arXiv arXiv 2025

  43. [43]

    Abac, A. G. and others, Phys. Rev. D (2026)

    2026

  44. [44]

    Andr´ es-Carcasona, O

    M. Andr´ es-Carcasona, O. J. Piccinni, M. Mart ´ ınez, and L.-M. Mir, PoSEPS-HEP2023, 067 (2024)

    2024

  45. [45]

    Z.-X. Wang, G. Cheng, J. Chen, H.-K. Guo, and A. L. 14 Miller, arXiv:2512.21738 [gr-qc] (2025), arXiv e-print

    arXiv 2025

  46. [46]

    Wang, X.-Y

    Z.-X. Wang, X.-Y. Chen, J. Chen, G. Cheng, H.-K. Guo, and A. L. Miller, arXiv:2512.21739 [gr-qc] (2025), arXiv e-print

    arXiv 2025

  47. [47]

    Viterbi, IEEE Transactions on Information Theory13, 260 (1967)

    A. Viterbi, IEEE Transactions on Information Theory13, 260 (1967)

    1967

  48. [48]

    Suvorova, L

    S. Suvorova, L. Sun, A. Melatos, W. Moran, and R. J. Evans, Phys. Rev. D93, 123009 (2016)

    2016

  49. [49]

    Bayley, G

    J. Bayley, G. Woan, and C. Messenger, Phys. Rev. D100, 023006 (2019)

    2019

  50. [50]

    Bayley, C

    J. Bayley, C. Messenger, and G. Woan, Phys. Rev. D102, 083024 (2020)

    2020

  51. [51]

    Bayley, C

    J. Bayley, C. Messenger, and G. Woan, Phys. Rev. D106, 083022 (2022)

    2022

  52. [52]

    Alestas, G

    G. Alestas, G. Morr´ as, T. S. Yamamoto, J. Garc ´ ıa-Bellido, S. Kuroyanagi, and S. Nesseris, Phys. Rev. D109, 123516 (2024)

    2024

  53. [53]

    The LIGO Scientific Collaboration, the Virgo Collabora- tion, and the KAGRA Collaboration, The Astrophysical Journal Supplement Series267, 29 (2023)

    2023

  54. [54]

    Viterbi tCW Search Pipeline, https://github.com/ raulrgdg/viterbi_search-method_PBHs(2026)

    2026

  55. [55]

    Maggiore,Gravitational Waves: Theory and Experiments, Oxford Master Series in Physics (Oxford University Press, 2007) p

    M. Maggiore,Gravitational Waves: Theory and Experiments, Oxford Master Series in Physics (Oxford University Press, 2007) p. 572

    2007

  56. [56]

    Bayley, soapcw: a viterbi-based search for contin- uous gravitational waves, https://pypi.org/project/ soapcw/(2023)

    J. Bayley, soapcw: a viterbi-based search for contin- uous gravitational waves, https://pypi.org/project/ soapcw/(2023)

    2023

  57. [57]

    Slade, Semantic Scholar (2013)

    G. Slade, Semantic Scholar (2013)

    2013

  58. [58]

    B. P. Abbott, R. Abbott, T. Abbott, S. Abraham, F. Ac- ernese, K. Ackley, C. Adams, V. Adya, C. Affeldt, M. Agathos,et al., Living reviews in relativity23, 1 (2020)

    2020

  59. [59]

    B. P. Abbottet al.(KAGRA, LIGO Scientific, Virgo), Living Rev. Rel.19, 1 (2016)

    2016

  60. [60]

    Aasiet al.(LIGO Scientific), Class

    J. Aasiet al.(LIGO Scientific), Class. Quant. Grav.32, 074001 (2015)

    2015

  61. [61]

    Buonanno, B

    A. Buonanno, B. Iyer, E. Ochsner, Y. Pan, and B. S. Sathyaprakash, Phys. Rev. D80, 084043 (2009)

    2009

  62. [62]

    LIGO Scientific Collaboration, Virgo Collaboration, and KAGRA Collaboration, L VK Algorithm Library - LAL- Suite, Free software (GPL) (2018)

    2018

  63. [63]

    Wette, SoftwareX12, 100634 (2020)

    K. Wette, SoftwareX12, 100634 (2020)

    2020

  64. [64]

    E. B. Wilson, Journal of the American Statistical Associ- ation22, 209 (1927)

    1927

  65. [65]

    Abbottet al.(KAGRA, VIRGO, LIGO Scientific), Astrophys

    R. Abbottet al.(KAGRA, VIRGO, LIGO Scientific), Astrophys. J. Suppl.267, 29 (2023)

    2023

  66. [66]

    B. F. Schutz and M. Tinto, Monthly Notices of the Royal Astronomical Society224, 131 (1987)

    1987