Pith. sign in

REVIEW 2 minor 256 references

Bayesian step-down procedure nearly matches the Bayes oracle for sparse signals under known dependence

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 · grok-4.3

2026-06-26 09:54 UTC pith:2BDIJH6M

load-bearing objection BSD matches the oracle closely in simulations under known covariance and adds an admissibility result, but the known-covariance scope limits how far the near-oracle claim travels.

arxiv 2606.22490 v1 pith:2BDIJH6M submitted 2026-06-21 stat.ME

Bayesian Model Pursuit and Near-Oracle Sparse Signal Discovery Under Dependence

Prasenjit Ghosh , Arijit Chakrabarti This is my paper
classification stat.ME
keywords sparse signal discoveryBayesian step-downmodel pursuitcovariance dependenceBayes oracleBayes risksupport recoveryadmissibility
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 authors introduce the Bayesian Step-Down procedure that uses posterior guidance to pursue sparse signal models while respecting the known covariance dependence. They define a Bayes oracle for one-factor dependence models and show through simulations that BSD performs nearly identically to this oracle in Bayes risk and in recovering the true support of signals. The procedure also closely tracks another method called MRD-GBS despite their different statistical foundations. BSD further admits a residual representation that establishes its admissibility for any covariance structure and simplifies computations. These results map out the Bayes-risk limits achievable for sparse discovery when dependence is present and known.

Core claim

BSD adopts a posterior-guided model-pursuit strategy that sequentially accumulates evidence for competing sparse signal configurations while explicitly incorporating the data's covariance structure. Across a broad range of dimensions, sparsity levels, and dependence structures, BSD exhibits near-oracle behavior and is often virtually indistinguishable from the Bayes Oracle in terms of Bayes risk and support recovery performance. A similarly close agreement is observed between BSD and MRD-GBS. BSD admits a residual representation, thereby yielding admissibility under arbitrary covariance dependence and substantial computational simplifications.

What carries the argument

Bayesian Step-Down (BSD) procedure: a sequential posterior-guided model pursuit that incorporates covariance structure to select sparse signal configurations

Load-bearing premise

The covariance matrix is known and fixed exactly, rather than needing to be estimated from the data.

What would settle it

Running simulations with covariance estimated from finite samples and checking whether BSD's Bayes risk deviates substantially from the oracle would falsify the near-oracle claim under realistic conditions.

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

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

0 major / 2 minor

Summary. The manuscript introduces the Bayesian Step-Down (BSD) procedure for sparse signal discovery under arbitrary known covariance dependence. It defines a Bayes Oracle for a class of sparse one-factor dependence models and reports simulation studies across dimensions, sparsity levels, and dependence structures in which BSD exhibits near-oracle behavior in Bayes risk and support recovery, often indistinguishable from the Bayes Oracle. BSD is compared to the Bayes Oracle, MRD-GBS, MRD, and Benjamini-Hochberg; close agreement is also noted between BSD and MRD-GBS. The paper further shows that BSD admits a residual representation, yielding admissibility under arbitrary covariance dependence.

Significance. If the simulation results hold, the work provides valuable insight into the attainable Bayes-risk frontier for sparse signal discovery when covariance is known exactly. The admissibility result via residual representation and the breadth of the simulation study constitute clear strengths; the near-oracle performance supplies a useful benchmark when exact oracle calculations are unavailable.

minor comments (2)
  1. [Simulation studies] The simulation section should explicitly state the number of replications, the precise data-generation mechanism for non-one-factor covariance structures, and any exclusion rules applied to the reported metrics so that the near-oracle claim can be reproduced without ambiguity.
  2. [Abstract] The citation 'Ghosh and Chakrabarti (2026)' appears in the abstract; confirm the reference list entry and year are correct.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive summary of the manuscript, the recognition of the simulation studies and admissibility result as strengths, and the recommendation for minor revision. No specific major comments were provided in the report.

Circularity Check

0 steps flagged

No significant circularity; self-citation present but not load-bearing

full rationale

The paper defines its own Bayes Oracle for one-factor models and compares BSD performance to it via explicit simulation and risk calculations under known covariance; no equation reduces the reported near-oracle behavior to a fitted input or self-referential definition. The single self-citation to Ghosh and Chakrabarti (2026) for MRD-GBS is used only for comparative benchmarking and does not justify the central BSD-oracle agreement or the residual-representation admissibility claim. The derivation chain remains self-contained against the stated external benchmarks and assumptions.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no explicit free parameters, axioms, or invented entities can be extracted or audited.

pith-pipeline@v0.9.1-grok · 5815 in / 1019 out tokens · 16288 ms · 2026-06-26T09:54:43.634608+00:00 · methodology

0 comments
read the original abstract

Sparse signal discovery is a fundamental problem in large-scale inference, where the goal is to identify a small number of active signals hidden among a large collection of null effects. Despite the prevalence of dependence in modern applications, relatively little is known about how much dependence can be exploited for efficient sparse signal recovery from a Bayes-risk perspective. In this paper, we develop a Bayesian Step-Down (BSD) procedure for sparse signal discovery under arbitrary known covariance dependence. BSD adopts a posterior-guided model-pursuit strategy that sequentially accumulates evidence for competing sparse signal configurations while explicitly incorporating the data's covariance structure. To assess its effectiveness, we introduce a Bayes Oracle for a class of sparse one-factor dependence models and compare BSD with the Oracle, the recently proposed MRD-GBS procedure of Ghosh and Chakrabarti (2026), the original MRD procedure of Cohen et al. (2009), and the Benjamini-Hochberg method. Our simulation studies reveal a striking phenomenon: across a broad range of dimensions, sparsity levels, and dependence structures, BSD exhibits near-oracle behavior and is often virtually indistinguishable from the Bayes Oracle in terms of Bayes risk and support recovery performance. Remarkably, a similarly close agreement is observed between BSD and MRD-GBS despite their fundamentally different Bayesian and frequentist foundations. These findings provide new insight into the attainable Bayes-risk frontier for sparse signal discovery under dependence and suggest that BSD may serve as a useful benchmark when exact Oracle calculations are unavailable. Finally, we show that BSD admits a residual representation, thereby yielding admissibility under arbitrary covariance dependence and substantial computational simplifications.

Figures

Figures reproduced from arXiv: 2606.22490 by Arijit Chakrabarti, Prasenjit Ghosh.

Figure 1
Figure 1. Figure 1: Bayes misclassification risks of the competing procedures under the [PITH_FULL_IMAGE:figures/full_fig_p031_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Bayes misclassification risks of the competing procedures under the [PITH_FULL_IMAGE:figures/full_fig_p033_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Bayes misclassification risks of the competing procedures under the [PITH_FULL_IMAGE:figures/full_fig_p034_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Empirical false discovery rates under the six one-factor dependence struc [PITH_FULL_IMAGE:figures/full_fig_p036_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Empirical false discovery rates under the six one-factor dependence struc [PITH_FULL_IMAGE:figures/full_fig_p037_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Empirical false discovery rates under the six one-factor dependence struc [PITH_FULL_IMAGE:figures/full_fig_p038_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Empirical false non-discovery rates under the six one-factor dependence [PITH_FULL_IMAGE:figures/full_fig_p040_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Empirical false non-discovery rates under the six one-factor dependence [PITH_FULL_IMAGE:figures/full_fig_p041_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Empirical false non-discovery rates under the six one-factor dependence [PITH_FULL_IMAGE:figures/full_fig_p042_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Empirical powers under the six one-factor dependence structures when [PITH_FULL_IMAGE:figures/full_fig_p044_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Empirical powers under the six one-factor dependence structures when [PITH_FULL_IMAGE:figures/full_fig_p045_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Empirical powers under the six one-factor dependence structures when [PITH_FULL_IMAGE:figures/full_fig_p046_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Average number of rejections under the six one-factor dependence struc [PITH_FULL_IMAGE:figures/full_fig_p048_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Average number of rejections under the six one-factor dependence struc [PITH_FULL_IMAGE:figures/full_fig_p049_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Average number of rejections under the six one-factor dependence struc [PITH_FULL_IMAGE:figures/full_fig_p050_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Bayes misclassification risks of BSD, MRD–GBS, MRD–CSX, and BH [PITH_FULL_IMAGE:figures/full_fig_p098_16.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

256 extracted references · 1 canonical work pages · 1 internal anchor

  1. [1]

    Journal of Multivariate Analysis , volume =

    Ledoit, Olivier and Wolf, Michael , title =. Journal of Multivariate Analysis , volume =

  2. [2]

    and Levina, Elizaveta , title =

    Bickel, Peter J. and Levina, Elizaveta , title =. The Annals of Statistics , volume =

  3. [3]

    Tony and Liu, Weidong , title =

    Cai, T. Tony and Liu, Weidong , title =. Journal of the American Statistical Association , volume =

  4. [4]

    Biostatistics , volume =

    Friedman, Jerome and Hastie, Trevor and Tibshirani, Robert , title =. Biostatistics , volume =

  5. [5]

    2025 , eprint =

    Paul, Sayantan and Chakrabarti, Arijit , title =. 2025 , eprint =

    2025

  6. [6]

    Annals of the Institute of Statistical Mathematics , year =

    Paul, Sayantan and Chakrabarti, Arijit , title =. Annals of the Institute of Statistical Mathematics , year =

  7. [7]

    Asymptotic Bayes Optimality Under Sparsity of Some Multiple Testing Procedures , journal =

    Frommlet, Florian and Bogdan, Ma. Asymptotic Bayes Optimality Under Sparsity of Some Multiple Testing Procedures , journal =. 2013 , volume =

    2013

  8. [8]

    2015 , eprint =

    Prasenjit Ghosh and Arijit Chakrabarti , title =. 2015 , eprint =

    2015

  9. [9]

    Covariance-Adaptive Residualization and Stagewise Calibration for Dependent Multiple Testing

    Ghosh, Prasenjit and Chakrabarti, Arijit , title =. arXiv preprint arXiv:2606.07466 , year =. 2606.07466 , archivePrefix =

    work page internal anchor Pith review Pith/arXiv arXiv

  10. [10]

    2025 , eprint =

    Paul, Sayantan and Ghosh, Prasenjit and Chakrabarti, Arijit , title =. 2025 , eprint =

    2025

  11. [11]

    and Castillo, I

    Abraham, K. and Castillo, I. and Roquain, E. , title =. The Annals of Statistics , volume =

  12. [12]

    Lehmann, E. L. and Casella, G. , title =

  13. [13]

    Lehmann, E. L. and Romano, J. P. , title =

  14. [14]

    Armagan and D

    A. Armagan and D. B. Dunson and M. Clyde , editor =. Generalized Beta Mixtures of Gaussians , booktitle =

  15. [15]

    , title =

    Clyde, Merlise and George, Edward I. , title =. Statistical Science , volume =

  16. [16]

    and Sillanp

    O'Hara, Robert B. and Sillanp. A Review of Bayesian Variable Selection Methods: What, How and Which , journal =

  17. [17]

    Shaffer, J. P. , editor =. Simultaneous testing. , booktitle =

  18. [18]

    Cournot, A. A. , title =. Reprinted 1984 as Volume 1 of Cournot's Oeuvres Completes , publisher =

    1984

  19. [19]

    Berry, D. A. , editor =. Multiple Comparisons, Multiple Tests and Data Dredging: A Bayesian Perspective , booktitle =

  20. [20]

    Irwin, J. O. , title =. Biometrika , volume =

  21. [21]

    Biometrika , volume =

    Student , title =. Biometrika , volume =

  22. [22]

    and Hotelling, H

    Working, H. and Hotelling, H. , title =. Journal of the American Statistical Association , volume =

  23. [23]

    Fisher, R. A. , title =

  24. [24]

    , title =

    Newman, D. , title =. Biometrika , volume =

  25. [25]

    Tukey, J. W. , title =. Biometrics , volume =

  26. [26]

    Duncan, D. B. , title =. Unpublished doctoral thesis, Iowa State College, Ames, Iowa , year =

  27. [27]

    Duncan, D. B. , title =. Virginia Journal of Science , volume =

  28. [28]

    Duncan, D. B. , title =. Virginia Polytechnique Institute Technical Report , year =

  29. [29]

    Duncan, D. B. , title =. Biometrics , volume =

  30. [30]

    Duncan, D. B. , title =. The Annals of Mathematical Statistics , volume =

  31. [31]

    Duncan, D. B. , title =. Technometrics , volume =

  32. [32]

    Duncan, D. B. and Bonner, R. G. , title =. Virginia Agri. Exp. Sta. Tech. Rept., Blacksburg, Va , year =

  33. [33]

    A method for judging all contrasts in the analysis of variance , journal =

    Scheff. A method for judging all contrasts in the analysis of variance , journal =

  34. [34]

    and Peritz, E

    Marcus, R. and Peritz, E. and Gabriel, K. R. , title =. Biometrika , volume =

  35. [35]

    Roy, S. N. , title =. The Annals of Mathematical Statistics , volume =

  36. [36]

    Roy, S. N. and Bose, R. C. , title =. The Annals of Mathematical Statistics , volume =

  37. [37]

    Armagan and D

    A. Armagan and D. B. Dunson and J. Lee , title =. Statistica Sinica , volume =

  38. [38]

    Cai, T. T. and Jin, J. and Low, M. J. , title =. The Annals of Statistics , volume =

  39. [39]

    and Jin, J

    Donoho, D. and Jin, J. , title =. The Annals of Statistics , volume =

  40. [40]

    van der Pas, S. L. and Salomond, J. B. and Schmidt-Hieber, J. , title =. Electronic Journal of Statistics , volume =

  41. [41]

    George, E. I. and Foster, D. P. , title =. Biometrika , volume =

  42. [42]

    Donoho, D. L. and Johnstone, I. M. , title =. Journal of the American Statistical Association , volume =

  43. [43]

    and Li, R

    Fan, J. and Li, R. , title =. Journal of the American Statistical Association , volume =

  44. [44]

    and Fan, J

    Antoniadis, A. and Fan, J. , title =. Journal of the American Statistical Association , volume =

  45. [45]

    and Benjamini, Y

    Abramovich, F. and Benjamini, Y. , title =. In Wavelets and Statistics. Lecture Notes in Statistics , volume =

  46. [46]

    and Braun, H

    Benjamini, Y. and Braun, H. , title =. The Annals of Statistics , volume =

  47. [47]

    and Benjamini, Y

    Abramovich, F. and Benjamini, Y. , title =. Computational Statistics and Data Analysis , volume =

  48. [48]

    and Birg

    Barron, A. and Birg. Risk bounds for model selection via penalization. , journal =

  49. [49]

    Gaussian model selection

    Birg. Gaussian model selection. , journal =

  50. [50]

    Bayesian Estimation of Sparse Signals with a Continuous Spike-and-Slab Prior , year =

    Ro. Bayesian Estimation of Sparse Signals with a Continuous Spike-and-Slab Prior , year =

  51. [51]

    and Datta, J

    Bhadra, A. and Datta, J. and Polson, N. G. and Willard, B. , title =. Bayesian Analysis , volume =

  52. [52]

    , title =

    Tibshirani, R. , title =. Journal of the Royal Statistical Society: Series B , volume =

  53. [53]

    Donoho, D. L. and Johnstone, I. M. , title =. Biometrika , volume =

  54. [54]

    Donoho, D. L. and Johnstone, I. M. , title =. Probability Theory Related Fields , volume =

  55. [55]

    Donoho, D. L. and Johnstone, I. M. , title =. Bernoulli , volume =

  56. [56]

    Donoho, D. L. and Johnstone, I. M. , title =. The Annals of Statistics , volume =

  57. [57]

    Pinsker, M. S. , title =. Problems Information Transmission , volume =

  58. [58]

    Efroimovich, S. Y. and Pinsker, M. S. , title =. Automat. Remote Control (translated from Avtomatika i Telemekhanika) , volume =

  59. [59]

    and Greenshtein, V

    Abramovich, F. and Greenshtein, V. and Pensky, M. , title =. The Annals of Statistics , volume =

  60. [60]

    and Angelini, C

    Abramovich, F. and Angelini, C. , title =. Sankhy

  61. [61]

    Brown, L. D. and Greenshtein, E. , title =. The Annals of Statistics , volume =

  62. [62]

    Zhang, C. H. and Huang, J. , title =. The Annals of Statistics , volume =

  63. [63]

    Zhang, C. H. , title =. The Annals of Statistics , volume =

  64. [64]

    , title =

    Zou, H. , title =. Journal of the American Statistical Association , volume =

  65. [65]

    and Hastie, T

    Zou, H. and Hastie, T. , title =. Journal of the Royal Statistical Society: Series B (Statistical Methodology) , volume =

  66. [66]

    and Schmidt-Hieber, J

    Castillo, I. and Schmidt-Hieber, J. and van der Vaart, A. W. , title =. 2014 , eprint =

    2014

  67. [67]

    and Chakrabarti, A

    Ghosh, P. and Chakrabarti, A. , title =. 2015 , eprint =

    2015

  68. [68]

    and Mizera, I

    Koenker, R. and Mizera, I. , title =. Journal of the American Statistical Association , volume =

  69. [69]

    and Benjamini, Y

    Abramovich, F. and Benjamini, Y. and Donoho, D. L. and Johnstone, I. M. , title =. The Annals of Statistics , volume =

  70. [70]

    Golubev, G. K. , title =. Problemy Peredachi Informatsii , volume =

  71. [71]

    Cai, T. T. and Jin, J. , title =. The Annals of Statistics , volume =

  72. [72]

    , title =

    Gelman, A. , title =. Bayesian Analysis , volume =

  73. [73]

    , title =

    Jin, J. , title =. Journal of the Royal Statistical Society: Series B , volume =

  74. [74]

    Ingster, Y. I. , title =. Mathematical Methods of Statistics , volume =

  75. [75]

    and Rice, J

    Meinshausen, N. and Rice, J. , title =. The Annals of Statistics , volume =

  76. [76]

    and Benjamini, Y

    Gavrilov, Y. and Benjamini, Y. and Sarkar, Sanat K. , title =. The Annals of Statistics , volume =

  77. [77]

    Benjamini and Y

    Y. Benjamini and Y. Hochberg , title =. Journal of the Royal Statistical Society: Series B , volume =

  78. [78]

    and Pati, D

    Bhattacharya, A. and Pati, D. and Pillai, N. and Dunson, D. B. , title =. 2014 , eprint =

    2014

  79. [79]

    and Pati, D

    Bhattacharya, A. and Pati, D. and Pillai, N. and Dunson, D. B. , title =. 2012 , eprint =

    2012

  80. [80]

    Donoho, D. L. , title =. Aide-Memoire of a Lecture at AMS Conference on Math Challenges of the 21st Century. , year =

Showing first 80 references.