Pith. sign in

REVIEW 3 major objections 7 minor 39 references

It's all In the (Exponential) Family: An Equivalence between Maximum Likelihood Estimation and Control Variates for Sketching Algorithms

T0 review · 3 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper proves that, inside an exponential family, an optimally corrected control variate estimator and the maximum likelihood estimator have identical asymptotic variance reduction, and turns that identity into a fixed-point algorithm…

desk verdict Theorem 4 is a genuinely useful variance equivalence in exponential families, but the paper oversells the CV-EM algorithm as converging to the MLE and the trace example produces a biased estimator, not an MLE. read the letter →

arxiv 2601.22378 v3 pith:45ETT7XJ submitted 2026-01-29 stat.ML cs.LGstat.AP

Keegan Kang , Kerong Wang , Ding Zhang , Rameshwar Pratap , Bhisham Dev Verma , Benedict H.W. Wong This is my paper · ORCID
classification stat.MLcs.LGstat.AP MSC 62F1062E2065C05
keywords exponentialfamiliesmaximumlikelihoodestimationcontrolvariatessketchingalgorithmsfixedpointiterationEMalgorithmvariancereductionfeaturehashing
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 sets out to prove that, in exponential families, maximum likelihood estimation and control variate estimation are asymptotically the same operation: if the parameters to be estimated equal the expected sufficient statistics, the asymptotic variance reduction of the MLE is exactly the variance reduction of an optimal CVE. That equality makes it possible to compute the MLE by iterating control-variate corrections instead of solving score equations, an EM-style fixed point procedure the paper calls CV-EM. The authors argue this removes two practical obstacles to using CVEs in sketching algorithms: theoretical variance calculations invalidated when unknown quantities are plugged into the control weights, and reproducibility failures caused by different root-finding or weight-estimation choices. On the bivariate Normal examples of feature hashing and random projections, CV-EM is shown to be numerically stable and fast, matching the best root-finding methods while avoiding Newton-Raphson's small-sample outliers.

What carries the argument

The load-bearing object is the pair of Jacobian matrices of the exponential family: V_{ij} = ∂μ_i/∂η_j, which is also the scaled covariance of the sufficient statistics, and its inverse ($V^{{-1}}$)_{ij} = ∂η_i/∂μ_j, which exists because η and μ are in one-to-one relation via dμ = V dη. The MLE variance uses n times the upper-left block of $V^{{-1}}$; the optimal CVE variance uses the Schur complement A - $BD^{{-1}}$B^T of the same block decomposition of V. The condition μ_i = ν_i makes the Fisher information reduce to that upper-left block, and the Schur complement identity then identifies the two variance expressions. This duality is the mechanism that turns control-variate corrections into an EM-style fixed point algorithm for the MLE.

What would settle it

Choose an exponential family with known parameters ν_K and target ν_E where at least one μ_i ≠ ν_i, as in estimating a nonlinear function of the sufficient-statistic means; compute the MLE asymptotic variance from the inverse Fisher information and the optimal CVE variance from the Schur complement, and check whether they differ. A simpler analogue: run Algorithm 1 on the chi-squared trace-estimation setting of Section 6 and verify that the fixed point, if it converges, has variance different from the claimed MLE-variance identity.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is Theorem 4: for observations from an exponential family, with parameters split into an estimated block ν_E and a known block ν_K, if every mean parameter μ_i = ∂ψ/∂η_i = E(y_i) equals the corresponding ν_i, then the asymptotic variance of the MLE of any linear combination of ν_E equals the variance of the CVE that uses the sufficient statistics y_j from the known block as control variates. The proof computes the Fisher information of the MLE via the Jacobian ∂η/∂μ and the variance of the optimal CVE via the Schur complement of the covariance block matrix, then shows the two expressions are the same block-matrix identity when μ_i = ν_i. The theorem is used to derive Algorithm 1, a fixed-point iteration for the MLE whose updates are optimal control-variate corrections, and to argue that iterating CV weights converges to the MLE. It also gives a heuristic recovery of the diagonal estimator for Hutchinson's trace problem from a closed-form fixed point.

Load-bearing premise

The argument depends on each parameter being estimated being numerically equal to the expected value of the corresponding sufficient statistic on that block; when this equality fails, the MLE and CVE variance expressions no longer coincide and the main theorem does not apply.

Editorial extensions

If this is right

  • CV-EM converges to the MLE by repeatedly applying optimal control-variate corrections, so one can obtain MLE-level variance without solving score equations.
  • Under the theorem's condition, the plug-in theoretical variance of a CVE is no longer invalid: after convergence its variance is the MLE variance at that sketch size.
  • Standardizing on CV-EM removes implementation-dependent discrepancies in MLE/CVE comparisons, since different root-finding methods give different small-sample behavior.
  • For Hutchinson's trace estimator, the fixed-point heuristic produces the diagonal estimator with variance expression 2(||M||_F^2 - Σm_ii^2)/k, giving a variance analysis that the original estimator lacked.

Reading between the lines

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

  • If the equivalence is taken as a design principle rather than just a theorem, it suggests deriving MLEs in new sketching settings by first computing expectation and covariance identities for sufficient statistics and then iterating the control-variate corrections, avoiding score-function algebra.
  • The condition μ_i = ν_i is narrow; the paper's own chi-squared trace example falls outside it, so the general claim should be read as an equivalence for the mean-parameter block rather than for all exponential-family parameterizations.
  • A testable extension would be to run CV-EM on multinomial frequency estimators under the known-marginal condition, where the equivalence should also hold, and to compare update counts with the same root-finding baselines.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. The paper studies the relationship between maximum likelihood estimation (MLE) and control variate estimation (CVE) in exponential families. The main theoretical result, Theorem 4, states that when the mean parameters μi coincide with the target parameters νi, the asymptotic variance of the MLE of a linear combination of the estimated parameters equals the variance of the optimal CVE that uses the remaining sufficient statistics as control variates. The authors then propose Algorithm 1 (CV-EM), a fixed-point iteration derived from the score equations, and claim it is an Expectation-Maximization algorithm converging to the MLE. They use this to discuss reproducibility of MLE/CVE-based sketching algorithms and, in Section 6, to derive an estimator for the matrix trace. Experiments on feature hashing and random projections, both based on a zero-mean bivariate Normal model, compare the fixed-point iteration with Newton-Raphson, secant, and two CVE variants.

Significance. If Theorem 4 is taken as proven, the paper establishes a clean variance-equivalence identity between two estimation paradigms; the proof is self-contained and the algebraic steps are consistent with the stated condition μi=νi. The experimental section is detailed and suggests that the fixed-point iteration can be numerically competitive. However, the paper's central advertised consequences, namely that Algorithm 1 converges to the MLE and that the method can find MLEs from known CV weights, are not established by the arguments provided. The trace example in Section 6 explicitly violates the theorem's conditions and the derived estimator is acknowledged to be biased, so it does not support the 'finding the MLE' claim. The contribution is therefore significant in its core variance result but requires major revision to separate proven facts from heuristic claims.

major comments (3)
  1. [Section 3, Appendix B.1-B.2] The paper claims in the abstract and in Section 1.2 that Algorithm 1 (CV-EM) is an EM algorithm converging to the MLE. Appendix B.1 only derives a fixed-point form of the score equation (Eq. 26); it does not prove that the Gauss-Seidel iteration converges, nor that a fixed point is the MLE rather than another stationary point. Appendix B.2 explicitly leaves the convergence rate as an open problem and acknowledges possible wrong-root convergence, with Table 1 reporting 0.26% of cubics having three real roots at k=10. Without a contraction, monotonicity, or global-convergence argument, the assertion that the algorithm 'converges to the MLE' is unsupported and load-bearing for the paper's practical contributions.
  2. [Section 6, Appendix D] The section is advertised as 'Finding the MLE via CVE weights' and the abstract claims the algorithm leads to finding the MLE when CV weights are known. However, the paper states that the conditions of Theorem 4 are not satisfied for Hutchinson's estimator, and the closed-form fixed point is Bekas's diagonal estimator (Bekas et al., 2007), which the text acknowledges is biased. No likelihood for tr(M) is analyzed, and the sentence 'the MLE would be a biased estimator' is not a derivation of an MLE. Thus the example does not substantiate Contribution Three.
  3. [Section 4, Observation 3] The paper claims that after convergence of Algorithm 1, the estimator has 'the variance of the MLE at that specific number of observations (or sketch size) k.' Theorem 4 is an asymptotic statement, proved via asymptotic normality of the MLE and the Fisher information. No finite-sample variance result for the MLE is proved, so the finite-k variance claim in Section 4 is not justified by the theorem. The experiments at small k compare MSE, but the theoretical link to the MLE variance at finite k is missing.
minor comments (7)
  1. [Section 3, Algorithm 1] The term 'EM algorithm' is used for what is actually a score-equation fixed-point iteration. The derivation in Appendix B.1 does not involve latent variables or a standard E-step; the authors should either justify the terminology or rename it, for example as 'CV fixed-point iteration'.
  2. [Appendix B, proof of Theorem 4] The step 'Using the differential relationship in Theorem 3 and the fact that μi=νi' to obtain V_n ∂η(νE)/∂νE = (1/n)[I_t;0] is terse. Please spell out the chain-rule calculation, since this is the point where the theorem's key condition enters.
  3. [Section 6] The sentence 'the MLE would be a biased estimator' is confusing because the estimator derived is Bekas's estimator, not a proven MLE. Rephrase to state clearly that the fixed-point limit is a known biased estimator and is not claimed to be an MLE.
  4. [Abstract, Section 5] The abstract says the algorithm is 'faster and numerically stable' without mentioning that the experiments cover only the zero-mean bivariate Normal distribution. Add the caveat to avoid overclaiming generality.
  5. [Appendix A, Eq. (8)] The text says 'where n/(σ11σ22-σ12^2) is factored out' but the factoring is not displayed in the equation. Showing the intermediate step would improve readability.
  6. [Theorem 2 and Theorem 4] The notation V_n and V is used inconsistently: Theorem 2 uses V_n for Cov(y) and also defines V, while Theorem 4 partitions V as the matrix of ∂μ/∂η. Define these once and use them consistently throughout.
  7. [Title] The title contains 'F or' (with a space), which appears to be a typo for 'For'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the MLE/CVE variance equivalence is a derived identity under the stated mu=nu condition, not an input to itself.

full rationale

Theorem 4's derivation is self-contained: the MLE asymptotic variance is computed from the score equation and the CVE variance from the covariance partition, and the equality V_MLE = V_CVE follows from the Schur complement identity plus the Jacobian relation dmu = V deta, cited to Efron. The condition mu_i = nu_i is an explicit assumption that makes the block structure of the Jacobian identity; it is not a restatement of the conclusion nor fitted from the data. Algorithm 1 is presented as a fixed-point rearrangement of the score equation in Appendix B.1, and the paper explicitly leaves convergence rate as an open question and concedes possible wrong-root convergence at small k; these are unproven algorithmic claims, not circular reductions. Section 6 explicitly says the conditions of Theorem 4 are not satisfied, derives Bekas' estimator, and acknowledges it is biased, so no fitted parameter is being relabeled as an MLE prediction. Self-citations (Kang 2021; Verma et al. 2022; Kang and Wong 2018) appear only as examples and motivation, not as load-bearing justification for the theorem or algorithm. No circular step is present.

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

The theoretical core rests only on standard exponential family and matrix identities plus the mu = nu condition; no free parameters are fitted to data. The convergence of Algorithm 1 is an additional empirical claim, not an axiom used in the proof.

assumptions (5)
  • standard math Exponential family cumulant identities: E(y) = grad psi(eta) and Cov(y) = (1/n) grad^2 psi(eta), plus the bijective map dmu = V deta.
    Invoked as Theorem 2 and Theorem 3 in Section 2, cited to Efron (2022); the proof of Theorem 4 relies on these identities and on the inverse function theorem.
  • standard math Asymptotic normality and efficiency of the MLE: the MLE converges to N(nu, i_nu^{-1}).
    Used in Section 2.1 and the proof of Theorem 4 to define the MLE asymptotic variance as the inverse Fisher information; cited to Lehmann (1999).
  • standard math Schur complement inversion formula for block matrices.
    Theorem 5 in Appendix B, cited to Golub and Van Loan (2013); used to equate the MLE variance with the CVE variance.
  • domain assumption Target parameters coincide with mean parameters: mu_i = nu_i for all i (or a linear bijection handled by Corollary 1).
    Assumed in Theorem 4; it is what makes the MLE score equation and the CVE fixed point rearrange to the same equations. The authors list this as a limitation in Section 7.
  • domain assumption The most variance-reducing control variates come from sufficient statistics y, and there exist p-t control variates with known means for the p-t known parameters.
    Assumptions One and Two in Section 2.2; these motivate the construction of the CVE but are not proven. They are not needed for the algebraic equality itself once the y-based CVs are chosen.

how reviews work

0 comments
Cite this review

Pith. "Pith review of It's all In the (Exponential) Family: An Equivalence between Maximum Likelihood Estimation and Control Variates for Sketching Algorithms." pith.science (2026). https://pith.science/paper/45ETT7XJ

@misc{pith2026260122378,
  author       = {Pith},
  title        = {Pith review of: It's all In the (Exponential) Family: An Equivalence between Maximum Likelihood Estimation and Control Variates for Sketching Algorithms},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/45ETT7XJ}},
  note         = {Machine review of arXiv:2601.22378}
}
read the original abstract

Maximum likelihood estimators (MLE) and control variate estimators (CVE) have been used in conjunction with known information across sketching algorithms and applications in machine learning. We prove that under certain conditions in an exponential family, an optimal CVE will achieve the same asymptotic variance as the MLE, giving a fixed point algorithm for the MLE. Experiments show the fixed point algorithm is faster and numerically stable compared to other root finding algorithms for the MLE for the bivariate Normal distribution, and we expect this to hold across distributions satisfying these conditions. We show how this algorithm leads to reproducibility for algorithms using MLE / CVE, and demonstrate how the algorithm leads to finding the MLE when the CV weights are known.

Figures

Figures reproduced from arXiv: 2601.22378 by the authors.

Figure 1
Figure 1. Idea behind estimation via sketching algorithms. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The exponential family exp  n⃗η(νE, νK) T ⃗y exp {nψ(⃗η(νE, νK))} g(X ) for VMLE = VCV E. Algorithm 1: EM Algorithm via Control Vari￾ates (CV-EM) Preconditions : Want to estimate some linear combination of νE = {ν1, . . . , νt} and know νK = {νt+1, . . . , νp} Require :t control variate expressions of the form f P i(ν1, . . . , νt) = yi + p j=t+1 cˆij (ν1, . . . , νt)(yj−µj ) with ˆcij known Result : MLE of νE Get … view at source ↗
Figure 3
Figure 3. MSE plot when ∥⃗x1∥ 2 = ∥⃗x2∥ 2 with θ = π 12 . FHMLE-NR MLE-SecantCV-EM CV-Init CV-Emp 0.2 0.4 0.6 0.8 1 Inner Product Estimates Estimates at k = 20 [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (10 more)
Figure 6
Figure 6. Figure 6: Plot of mean squared error (MSE) of respective estimates of inner product between [PITH_FULL_IMAGE:figures/full_fig_p023_6.png]
Figure 7
Figure 7. Figure 7: MSE plot when ∥⃗x1∥ 2 = ∥⃗x2∥ 2 with θ = π 12 for feature hashing. A lower value is better. (ii) Proportion of outliers in 10, 000 estimates: A smaller proportion of outliers suggests better performance [PITH_FULL_IMAGE:figures/full_fig_p023_7.png]
Figure 8
Figure 8. Figure 8: Boxplots of estimated inner products via the baseline feature hashing estimate ( [PITH_FULL_IMAGE:figures/full_fig_p024_8.png]
Figure 9
Figure 9. Figure 9: Boxplots of update steps for MLE-NR, MLE-Secant, and CV-EM at k = {20, 40, 60, 80, 100} for feature hashing. A lower and narrower boxplot is better. From [PITH_FULL_IMAGE:figures/full_fig_p025_9.png]
Figure 10
Figure 10. Figure 10: Plot of mean squared error (MSE) of respective estimates of inner product between [PITH_FULL_IMAGE:figures/full_fig_p027_10.png]
Figure 11
Figure 11. Figure 11: MSE plot when ∥⃗x1∥ 2 = ∥⃗x2∥ 2 with θ = π 12 for random projections. A lower value is better. RPMLE-NR MLE-Secant CV-EM CV-InitCV-Emp -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Inner Product Estimates Estimates at k = 10 RPMLE-NR MLE-Secant CV-EM CV-InitCV-Emp 0.2 0.3 0.4 0.5 0.6…
Figure 12
Figure 12. Figure 12: Boxplots of estimated inner products via the baseline random projection estimate ( [PITH_FULL_IMAGE:figures/full_fig_p028_12.png]
Figure 13
Figure 13. Figure 13: Boxplots of update steps for MLE-NR, MLE-Secant, and CV-EM at k = {20, 40, 60, 80, 100} for random projections. A lower and narrower boxplot is better. From [PITH_FULL_IMAGE:figures/full_fig_p029_13.png]
Figure 14
Figure 14. Figure 14: Boxplots of empirical α and C at k = 100 for MLE-NR, MLE-Secant, and CV-EM across all vector pairs at 25,000 observations for feature hashing. The boxplots in Figures 14 and 15 show the empirical α and C for each run. The interquartile range of α for Newton Raphson is…
Figure 15
Figure 15. Figure 15: Boxplots of empirical α and C at k = 100 for MLE-NR, MLE-Secant, and CV-EM across all vector pairs at 25,000 observations for random projections. D Further Discussion on Hutchinson’s Estimator This section of the appendix elaborates on our heuristic of how CV-EM can b…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 36 canonical work pages

  1. [1]

    P., Pennington, J., Johnson, M

    Adams, R. P., Pennington, J., Johnson, M. J., Smith, J., Ovadia, Y., Patton, B., and Saunderson, J. (2018). Estimating the Spectral Density of Large Implicit Matrices . arXiv preprint arXiv:1802.03451

    arXiv 2018

  2. [2]

    Amari, S.-i. (1987). Differential geometrical theory of statistics , volume 10. Hayward California

    work page 1987

  3. [3]

    and Nagaoka, H

    Amari, S.-i. and Nagaoka, H. (2000). Methods of information geometry , volume 191. American Mathematical Soc

    work page 2000

  4. [4]

    Atkinson, K. E. (2008). An Introduction to Numerical Analysis . John Wiley & Sons

    work page 2008

  5. [5]

    Avron, H. (2010). Counting triangles in large graphs using randomized matrix trace estimation. In Workshop on Large-scale Data Mining: Theory and Applications , volume 10, page 9

    work page 2010

  6. [6]

    Axler, S. (2015). Linear Algebra Done Right . Springer

    work page 2015

  7. [7]

    J., and Yu, B

    Balakrishnan, S., Wainwright, M. J., and Yu, B. (2017). Statistical guarantees for the em algorithm: From population to sample-based analysis

    work page 2017

  8. [8]

    Baston, R. A. and Nakatsukasa, Y. (2022). Stochastic diagonal estimation: Probabilistic bounds and an improved algorithm. arXiv preprint arXiv:2201.10684

    arXiv 2022

Show all 39 references
  1. [9]

    Bekas, C., Kokiopoulou, E., and Saad, Y. (2007). An estimator for the diagonal of a matrix. Applied numerical mathematics , 57(11-12):1214--1229

    2007

  2. [10]

    Charikar, M. S. (2002). Similarity Estimation Techniques From Rounding Algorithms . In Proceedings of the Thirty-Fourth Annual ACM Symposium on Theory of Computing , pages 380--388

    2002

  3. [11]

    and Muthukrishnan, S

    Cormode, G. and Muthukrishnan, S. (2005). An Improved Data Stream Summary: the Count-Min Sketch and its Applications . Journal of Algorithms , 55(1):58--75

    2005

  4. [12]

    and Kressner, D

    Cortinovis, A. and Kressner, D. (2021). On randomized trace estimates for indefinite matrices with an application to determinants. Foundations of Computational Mathematics , pages 1--29

    2021

  5. [13]

    Csisz \'a r, I. (1984). Information geometry and alternating minimization procedures. Statistics and Decisions, Dedewicz , 1:205--237

    1984

  6. [14]

    Efron, B. (1978). The Geometry of Exponential Families . The Annals of Statistics , pages 362--376

    1978

  7. [15]

    Efron, B. (2022). Exponential Families in Theory and Practice . Cambridge University Press

    2022

  8. [16]

    Glynn, P. W. and Szechtman, R. (2002). Some New Perspectives on the Method of Control Variates . In Monte Carlo and Quasi-Monte Carlo Methods 2000: Proceedings of a Conference held at Hong Kong Baptist University, Hong Kong SAR, China, November 27--December 1, 2000 , pages 27-...

    2002

  9. [17]

    Golub, G. H. and Van Loan, C. F. (2013). Matrix Computations . JHU Press

    2013

  10. [18]

    Hutchinson, M. F. (1989). A Stochastic Estimator of the Trace of the Influence Matrix for Laplacian Smoothing Splines . Communications in Statistics-Simulation and Computation , 18(3):1059--1076

    1989

  11. [19]

    and Motwani, R

    Indyk, P. and Motwani, R. (1998). Approximate Nearest Neighbors: Towards Removing the Curse of Dimensionality . In Proceedings of the Thirtieth Annual ACM Symposium on Theory of Computing , pages 604--613

    1998

  12. [20]

    Kang, K. (2021). Correlations Between Random Projections and the Bivariate Normal . Data Mining and Knowledge Discovery , 35(4):1622--1653

    2021

  13. [21]

    and Wong, W

    Kang, K. and Wong, W. P. (2018). Improving Sign Random Projections with Additional Information . In International Conference on Machine Learning , pages 2479--2487. PMLR

    2018

  14. [22]

    Keener, R. W. (2010). Theoretical Statistics: Topics for a Core Course . Springer Science & Business Media

    2010

  15. [23]

    Lavenberg, S. S. and Welch, P. D. (1981). A Perspective on the Use of Control Variables to Increase the Efficiency of Monte Carlo Simulations . Management Science , 27(3):322--335

    1981

  16. [24]

    Lehmann, E. L. (1999). Elements of large-sample theory . Springer

    1999

  17. [25]

    and Rademacher, L

    Leroux, B. and Rademacher, L. (2024). Euclidean distance compression via deep random features. Advances in Neural Information Processing Systems , 37:53261--53284

    2024

  18. [26]

    Li, P., Church, K., and Hastie, T. (2008). One Sketch for All: Theory and Application of Conditional Random Sampling . Advances in Neural Information Processing Systems , 21

    2008

  19. [27]

    J., and Church, K

    Li, P., Hastie, T. J., and Church, K. W. (2006). Improving Random Projections Using Marginal Information . In Learning Theory: 19th Annual Conference on Learning Theory, COLT 2006, Pittsburgh, PA, USA, June 22-25, 2006. Proceedings 19 , pages 635--649. Springer

    2006

  20. [28]

    and Li, P

    Li, X. and Li, P. (2022). C- M in H ash: Improving minwise hashing with circulant permutation. In Chaudhuri, K., Jegelka, S., Song, L., Szepesvari, C., Niu, G., and Sabato, S., editors, Proceedings of the 39th International Conference on Machine Learning , volume 162 of Procee...

    2022

  21. [29]

    McLachlan, G. J. and Krishnan, T. (2007). The EM algorithm and extensions . John Wiley & Sons

    2007

  22. [30]

    A., Musco, C., Musco, C., and Woodruff, D

    Meyer, R. A., Musco, C., Musco, C., and Woodruff, D. P. (2021). Hutch++: Optimal stochastic trace estimation. In Symposium on Simplicity in Algorithms (SOSA) , pages 142--155. SIAM

    2021

  23. [31]

    and Herschel, M

    Nauman, F. and Herschel, M. (2022). An Introduction to Duplicate Detection . Springer Nature

    2022

  24. [32]

    J., Girolami, M., and Chopin, N

    Oates, C. J., Girolami, M., and Chopin, N. (2017). Control Functionals for Monte Carlo Integration . Journal of the Royal Statistical Society Series B: Statistical Methodology , 79(3):695--718

    2017

  25. [33]

    and Kulkarni, R

    Pratap, R. and Kulkarni, R. (2021). Variance Reduction in Frequency Estimators via Control Variates Method . In Uncertainty in Artificial Intelligence , pages 183--193. PMLR

    2021

  26. [34]

    D., and Kulkarni, R

    Pratap, R., Verma, B. D., and Kulkarni, R. (2021). Improving tug-of-war sketch using control-variates method. In SIAM Conference on Applied and Computational Discrete Algorithms (ACDA21) , pages 66--76. SIAM

    2021

  27. [35]

    Roche, A. (2011). Em algorithm and variants: An informal tutorial. arXiv preprint arXiv:1105.1476

    2011 arXiv

  28. [36]

    Shao, J. (2003). Mathematical Statistics . Springer Science & Business Media

    2003

  29. [37]

    F., Oates, C

    South, L. F., Oates, C. J., Mira, A., and Drovandi, C. (2023). Regularized zero-variance control variates. Bayesian Analysis , 18(3):865--888

    2023

  30. [38]

    D., Pratap, R., and Thakur, M

    Verma, B. D., Pratap, R., and Thakur, M. (2022). Variance Reduction in Feature Hashing using MLE and Control Variate Method . Machine Learning , 111(7):2631--2662

    2022

  31. [39]

    Weinberger, K., Dasgupta, A., Langford, J., Smola, A., and Attenberg, J. (2009). Feature Hashing for Large Scale Multitask Learning . In Proceedings of the 26th Annual International Conference on Machine Learning , pages 1113--1120

    2009

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.