G-HIVE: Parameter Estimation and Approximate Inference for Multivariate Response Generalized Linear Models with Hidden Variables

Inbeom Lee; Yang Ning

REVIEW 3 major objections 6 minor 18 references

G-HIVE: Parameter Estimation and Approximate Inference for Multivariate Response Generalized Linear Models with Hidden Variables

T0 review · 3 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read In multivariate GLMs with hidden confounders, an inverse-variance reweighting restores the linear-model structure, so that a spectral projection removes the first-order bias and yields Gaussian approximate inference for the projected parame

desk verdict A substantial GLM extension of hidden-variable regression with a real gap between the claimed Berry–Esseen justification and the conditions under which it holds. read the letter →

arxiv 2509.00196 v1 pith:XMWZ3VKF submitted 2025-08-29 stat.ME math.STstat.TH

Inbeom Lee , Yang Ning This is my paper · ORCID

classification stat.MEmath.STstat.TH MSC 62J1262F1262H25

keywords generalizedlinearmodelshiddenvariablesunmeasuredconfoundersmultivariateresponsemodifiedquasi-likelihoodfactororthogonalprojectiondebiasingBerry-Esseenbound

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

Unobserved variables that influence both the covariates and the outcome bias ordinary regression, and the standard fix for linear models fails once the outcome model is nonlinear, because the bias no longer lies in the span of the hidden-factor loadings. This paper shows the failure has a remedy: fit each response with a quasi-likelihood whose score is divided by the model variance function b'', so that the leading bias of the fitted coefficient matrix F* again lies in the hidden-factor space and can be removed by the orthogonal projection P_B-perp. The remaining approximation bias per response is O(1/p), smaller the more covariates one has, and the projection matrix is recovered by PCA on the covariance of the reweighted residuals. Because the asymptotic bias still dominates the stochastic error at the sqrt(n) scale, exact inference on the original parameter Theta is replaced by second-order approximate inference on P_B-perp F*, with a Berry-Esseen bound controlling how close the Gaussian intervals are to correct. If right, this gives a practical pipeline for debiased estimation and uncertainty quantification in multivariate binary or count regression under hidden confounding.

What carries the argument

The load-bearing device is the inverse-variance reweighted estimating equation E[(Y_m - b'(F*_m X))/b''(F*_m X) times X^T] = 0, equivalently a modified quasi-likelihood whose integrand divides by the variance function b''. This weight reshapes the leading misspecification bias from a term involving the diagonal matrix D, which does not lie in the column space of B and cannot be projected away, into B_m Sigma_Z A^T Sigma_X^-1, a row of the hidden loading matrix B, which the projection P_B-perp = I - B(B^T B)^-1 B^T annihilates. The companion mechanism is PCA: the covariance of the reweighted residuals takes the form diagonal plus B(.)B^T, so under the factor-model pervasiveness assumption its

What would settle it

Simulate a GLM with hidden confounders where Assumption 1 is violated by a controlled amount, forcing a nonzero component of Theta along a known direction in the column space of B, and check whether Theta-hat converges to Theta, to P_B-perp(Theta + BL), or to something else. A direct numerical check of the claimed rates is also decisive: compute F* and P_B-perp F* from a very large Monte Carlo sample and verify the O(1/sqrt(p)) and O(1/p) decay of the two per-response biases as p increases. A further check: in a setting with a measured confounder treated as hidden, compare G-HIVE interval cove

Watch

Extended reading notes

Core claim

The paper's central claim is that the orthogonality property that makes hidden-variable correction work in linear regression, namely that the first-order misspecification bias lies in the column space of the hidden loading matrix B, can be restored in multivariate GLMs by replacing the ordinary score with an inverse-variance-weighted score. Define F* by E[(Y_m - b'(F*_m X))/b''(F*_m X) times X^T] = 0; then F*_m - Theta_m = B_m E(ZZ^T) A^T {E(XX^T)}^-1 + O(1/p), whose leading term is proportional to a row of B and is therefore killed by the projection P_B-perp. The paper proves the per-response norms ||F* - Theta||_F/sqrt(M) = O(1/sqrt(p)) and ||P_B-perp F* - Theta||_F/sqrt(M) = O(1/p), estim

Load-bearing premise

The load-bearing premise is that the true regression coefficients are orthogonal to the column space of the hidden loadings (P_B Theta = 0); since B is never observed, this condition cannot be checked in the data, and if it fails the projection removes part of the genuine covariate-response association and the estimator targets a different quantity.

Editorial extensions

If this is right

Because the residual bias after projection is O(1/p) per response, collecting more covariates actively improves the deconfounding: the approximation gap between the estimable target and Theta shrinks as p grows.
Confidence intervals built from Theta-hat's limiting Gaussian law cover the projected parameter P_B-perp F*, a second-order approximation of Theta, at the nominal level, with the Berry-Esseen bound quantifying the error; coverage for Theta itself is not guaranteed, and simulations show the naive MLE's intervals badly undercover.
When the number of responses M grows faster than the number of covariates p (roughly M asymp p^alpha with alpha > 2 in the paper's rate simplification), G-HIVE's convergence rate beats the naive MLE that ignores hidden variables, so more response variables are a blessing of dimensionality for estimating the projection.
The eigenvalue-ratio rule selects the number of hidden factors K without tuning parameters, making the pipeline fully data-driven, and simulation results show the data-driven choice performs close to the oracle versions that know K or the true projection.
The framework covers any GLM with bounded second derivative of the log-partition function (linear, logistic, Poisson), extending hidden-variable correction beyond the linear-regression settings previously solved.

Reading between the lines

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

Because Assumption 1 (P_B Theta = 0) is untestable from observed data, since B is never seen, a natural companion analysis is a sensitivity study: perturb the estimated projection toward directions of non-orthogonality and report how much Theta-hat and the coverage statements shift.
The reweight-then-project mechanism is a candidate template for other misspecified-model settings where bias has a structured column-space component; the paper's own open directions (general dependent noise, p > n via regularization) mark the natural next tests.
The Berry-Esseen regime conditions (r_p < 1/3 and r_M > (1 - r_p) or (1/2 + r_p)) imply a practical rule of thumb: the number of responses M must comfortably exceed the number of covariates p before the Gaussian approximation is trustworthy.
The deconfounding check demonstrated on the NHANES data suggests a general validation pattern: treat a measured variable as hidden, and compare G-HIVE's recovered coefficients with those from a model that observes it.

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.

Desk Editor's Note

read the letter

This paper extends Bing et al. from linear to multivariate GLMs with hidden variables. The reweighted estimating equation plus projection is genuinely new, and the bias expansions in Theorems 1–4 are nontrivial and seem correct. The estimation side is solid: the approximation bias rates (1/sqrt(p) and 1/p) and the stochastic error rates are coherent, and the PCA step is well motivated. I would struggle to call the inference guarantee established, though. Theorem 5 gives a Berry–Esseen bound containing a term delta_2 that depends on ||F^*v||_2. Then, in the simplification after the theorem, the authors say they 'further assume' Rn=O(1), K=O(1), and ||F^*v||_2=O(1). That extra condition is not part of the theorem's assumptions and is not implied by their scaling. Their own simulation DGP normalizes each row of Theta to unit norm, so for typical unit v, ||F^*v||_2 is about sqrt(M/p), which grows under the stated rates M~n^{rM}, p~n^{rp} with rM>1/2+rp. Plugging that in, delta_2 does not go to zero; it diverges. So the claim that the Berry–Esseen bound 'justifies the validity of the proposed approximate inference approach' is not supported in the regime they analyze. The authors are transparent that they are adding a condition, but the abstract and the discussion around Theorem 5 give an unconditional impression. This is a repair, not a rewrite: the bound in (30) is correct as a statement about the Gaussian approximation error, but it needs a formal assumption like ||F^*v||_2=O(1) and a discussion of when it holds, or a sharper treatment that keeps the dependence on ||F^*v||_2 and shows it vanishes under the same scaling. Two smaller issues: no code or data, which limits easy verification, and Assumption 1 (PB Theta=0) is untestable but standard for this literature. On the whole this is a serious paper with a strong estimation story and an inference section that overstates its guarantee. It deserves a careful referee who will push on the Berry–Esseen condition. I'd bring it to a reading group on deconfounding or high-dimensional latent variable models, and I would likely cite it once the inference gap is fixed.

Referee Report

3 major / 6 minor

Summary. The paper studies multivariate-response generalized linear models with hidden variables under a factor-model assumption relating X to latent Z. It defines a pseudo-true parameter F* through inverse-variance reweighted estimating equations and proposes an estimator \hat Θ = \hat P_B^\perp \hat F, where \hat F is a modified quasi-likelihood estimator and \hat P_B^\perp is obtained from PCA on reweighted residuals. The authors prove that the first-order approximation bias is O(1/√p) and the projected second-order bias is O(1/p), give convergence rates for \hat F, \hat P_B^\perp, and \hat Θ, and provide a Berry–Esseen bound for linear functionals of \hat Θ centered at P_B^\perp F*. Simulations and two real-data analyses illustrate the method against the naive MLE.

Significance. If the technical gaps noted below are addressed, this is a meaningful contribution. Extending surrogate-variable analysis and spectral deconfounding from linear regression to non-linear multivariate GLMs is non-trivial, and the proposed reweighted estimating equations are a sensible device to restore the orthogonality that fails for the ordinary misspecified GLM score. The explicit bias rates O(1/√p) and O(1/p) are interesting, and the paper is unusually candid about the fact that the inference target is the surrogate P_B^\perp F*, not Θ itself. The proofs are detailed and the method is transparently algorithmic. The main reservations are that the stated vanishing Berry–Esseen bound requires an extra normalization on F* that is not part of the assumptions and is violated in the paper's own simulation regime, and that the sample-splitting argument in the proof of Theorem 5 is not fully matched to the final averaged estimator.

major comments (3)

[§4, Theorem 5 / Eq. (32)] The simplified Berry–Esseen bound (32) is obtained by 'further assum[ing] Rn=O(1), K=O(1), and ||F^*v||_2=O(1)' (Section 4). The last condition is not part of Assumptions 1–4 and is not stated in Theorem 5. Under the simulation DGP of Section 5.1 (rows of Θ normalized to unit L2 norm) and the scaling M≍n^{rM}, p≍n^{rp} with rp<1/3, rM>1/2+rp, ||F^*v||_2 is of order √(M/p) for a generic unit v (e.g., v=e_j). Substituting this into δ2 gives a leading term Rn√n/p, which diverges because p=o(√n) in that regime. Hence the claim that the Berry–Esseen bound goes to zero and 'justifies the validity of the proposed approximate inference approach' (Abstract, §3.2, Discussion) is not established for growing M. The paper must either add the norm condition to the theorem statements and propagate it through the claims, or temper the asymptotic validity claim.
[Appendix B.5, proof of Theorem 5 / Algorithm 1] The proof repeatedly says 'due to sample splitting, for simplicity we can equivalently assume that \hat F_m is independent of Y_i and X_i'. But Algorithm 1 averages the two folds to produce \hat F and \hat Σ, and \hat G_m in (20) is computed on all n observations with the averaged \hat F_m. The averaged estimator is not independent of either fold, and the held-out independence used for the concentration steps is not satisfied by the final inferential object. The manuscript should spell out a cross-fitting scheme that supports the expansion in (71) and the variance estimator (31), or provide a proof under the actual dependence; otherwise the Gaussian approximation guarantee in Theorem 5 is not fully established.
[§3.2, §5.5, and Abstract] The confidence interval is for u^T P_B^\perp F^*v, not for u^T Θv. The paper is transparent in Section 3.2 and Table 1, but the abstract and Section 7 phrase the result as general 'uncertainty quantification' and 'validity of the proposed approximate inference approach' without the surrogate caveat. Since the gap between P_B^\perp F^* and Θ is only O(1/p) in Frobenius norm, for finite p it can exceed the parametric width; the favorable coverage for Θ in Table 1 is a finite-p simulation result rather than a theoretical guarantee. Please state plainly in the abstract and main text that the CI covers the surrogate P_B^\perp F^*, and that no frequentist coverage claim is made for Θ.

minor comments (6)

[§3.1, after (13)] The displayed definitions say 'F^* := [F_1^{*T},...,F_M^{*T}]^T ∈ R^{M×K}' and similarly for \hat F; the correct dimension is R^{M×p}, since each F_m is 1×p. Please fix this typo.
[Figures 1 and 2] The y-axis label 'Scaled Squared Frobenius Error' is ambiguous. The text defines ||\hat Θ − Θ||_F^2 / √(pM), but it would be helpful to state why this normalization is used and to show Monte Carlo error bars, especially since r ranges from 20 to 500.
[§5.3] The 'true' F^* is obtained by maximizing the modified quasi-likelihood with n=2×10^5. This is an approximation to the population F^*, not the exact estimand; please state this clearly and report Monte Carlo variability.
[§5.5 / Table 1] Coverage probabilities are averaged over r=100 repetitions. At the 95% level the Monte Carlo standard error is about 0.04, which is not negligible. Report standard errors or use more repetitions, and separately report coverage for the surrogate P_B^\perp F^* and for Θ.
[Reproducibility] No code or data availability statement is provided. Given the many implementation choices (initial values for the non-concave quasi-likelihood, data-driven K estimation, eigenvalue ratio cutoff), releasing code would substantially strengthen the paper.
[§4, Eq. (31)] The rate for \hat s_n^2/n contains a term of the form 'p √(K log(p∨M)/n) log(M∨n)' inside brackets; the parentheses are difficult to parse. Please clean up the display and double-check that all logarithm powers are consistent with the proof.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the reweighted estimating equations, PCA projection, and bias expansions are explicit derivations; self-citations are not load-bearing.

full rationale

The derivation chain is self-contained. The key estimand F* is defined by the explicit estimating equation (5), E[(Y_m - b'(F*_m X))/b''(F*_m X) X^T] = 0, which does not presuppose Theta or B. Theorem 1 then derives the bias expansion F*_m - Theta_m = B_m Sigma_Z A^T Sigma_X^{-1} + O(1/p) via Taylor expansion around the true linear predictor; the leading term lying in the column space of B is a mathematical consequence of the chosen inverse-variance reweighting, not an input defining F*. The projection estimator Theta_hat = P_B_perp_hat F_hat combines the modified quasi-likelihood maximizer (10) with PCA on the residual covariance (12), and Theorems 2-5 obtain rates and a Berry-Esseen bound by concentration arguments and Davis-Kahan, not by fitting the target. The paper is also explicit that the confidence interval targets u^T P_B_perp F* v, not Theta itself (Section 3.2, Theorem 5), so there is no disguised prediction. The self-citations to Bing et al. (2022), one of which is co-authored by Ning, appear in Assumption 1 and Remark 1, but the identifiability assumption PB Theta = 0 is also attributed to Lee et al. (2017) and Wang et al. (2017), and the eigenvalue-ratio justification is supplemented by the paper's own Section C; neither step reduces a central claim to a self-citation. The Berry-Esseen simplification in Section 4 adds the condition ||F* v||_2 = O(1) beyond Theorem 5's stated assumptions; this is a missing-assumption/scaling concern about when the bound vanishes, not a circular reduction to a fitted input or self-citation. Overall, no significant circularity is present.

Assumptions & free parameters 1 free parameters · 6 assumptions · 2 invented entities

The central claim rests on the factor model, the identifiability orthogonality condition, and pervasiveness assumptions. No constants are fitted in the theory; the only user-chosen quantity is K, the number of hidden factors. The introduced constructs F* and P_B^perp F* are mathematical tools rather than empirical entities.

free parameters (1)

K (number of hidden factors) = estimated by eigenvalue ratio in (14) or user-specified
Algorithm 1 requires K; the data-driven version estimates it, and the theoretical guarantees assume K is known or consistently estimated.

assumptions (6)

domain assumption Assumption 1: PB*Theta = 0
Identifiability of Theta; without it, the projection can remove true signal. Invoked in Section 2.2 and throughout.
domain assumption Factor model X = AZ + W with W independent of Z and Sigma_W = tau I_p (tau=1)
Used to derive the form of the bias and the identifiability of the projection; stated in Eq (2) and Section 2.1.
domain assumption Assumption 2: sub-Gaussian W and Z, bounded X, sub-exponential GLM errors
Required for concentration inequalities in the proofs; stated as Assumption 2.
domain assumption Assumption 3: b'' bounded between C1 and C2, derivatives b', b''', b'''' bounded
Standard in GLM analysis; used to control Taylor expansions and Hessian perturbations.
domain assumption Assumption 4: pervasiveness of A and B: eigenvalues of A^T A ~ p, B^T B ~ M, eigenvalues of Sigma_Z bounded
Required for the PCA step to recover the column space of B; stated in Assumption 4.
domain assumption Canonical link GLM (1) and conditional independence of Y_m given X, Z
The entire model setup; stated in Eq (1) and Section 2.1.

invented entities (2)

F* (pseudo-true parameter via reweighted estimating equations)
purpose: Intermediate estimand whose first-order bias lies in col(B), enabling debiasing by projection.
Introduced in Eq (5); a mathematical construct, not an empirically anchored entity.
P_B^perp F* (second-order target)
purpose: Inference target that is O(1/p) from Theta; used for confidence intervals.
Shifted target for approximate inference; defined through F* and the projection P_B^perp.

how reviews work

0 comments

Cite this review

Pith. "Pith review of G-HIVE: Parameter Estimation and Approximate Inference for Multivariate Response Generalized Linear Models with Hidden Variables." pith.science (2026). https://pith.science/paper/XMWZ3VKF

@misc{pith2026250900196,
  author       = {Pith},
  title        = {Pith review of: G-HIVE: Parameter Estimation and Approximate Inference for Multivariate Response Generalized Linear Models with Hidden Variables},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XMWZ3VKF}},
  note         = {Machine review of arXiv:2509.00196}
}

read the original abstract

In practice, there often exist unobserved variables, also termed hidden variables, associated with both the response and covariates. Existing works in the literature mostly focus on linear regression with hidden variables. However, when the regression model is non-linear, the presence of hidden variables leads to new challenges in parameter identification, estimation, and statistical inference. This paper studies multivariate response generalized linear models (GLMs) with hidden variables. We propose a unified framework for parameter estimation and statistical inference called G-HIVE, short for 'G'eneralized - 'HI'dden 'V'ariable adjusted 'E'stimation. Specifically, based on factor model assumptions, we propose a modified quasi-likelihood approach to estimate an intermediate parameter, defined through a set of reweighted estimating equations. The key of our approach is to construct the proper weight, so that the first-order asymptotic bias of the estimator can be removed by orthogonal projection. Moreover, we propose an approximate inference framework for uncertainty quantification. Theoretically, we establish the first-order and second-order asymptotic bias and the convergence rate of our estimator. In addition, we characterize the accuracy of the Gaussian approximation of our estimator via the Berry-Esseen bound, which justifies the validity of the proposed approximate inference approach. Extensive simulations and real data analysis results show that G-HIVE is feasibly implementable and can outperform the baseline method that ignores hidden variables.

Figures

Figures reproduced from arXiv: 2509.00196 by the authors.

**Figure 2.** The (left) graph shows the estimation error [PITH_FULL_IMAGE:figures/full_fig_p016_2.png] view at source ↗

**Figure 3.** Models (1) and (2) and variables in the real dataset NHANES (2017-2018) in the context of confounding (Pearl (2009)). 50 [PITH_FULL_IMAGE:figures/full_fig_p050_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 17 canonical work pages

[1]

Ahn, S. C. and Horenstein, A. R. (2013). Eigenvalue ratio test for the number of factors. Econometrica 81 1203–1227. Bai, J. (2003). Inferential theory for factor models of large dimensions. Econometrica 71 135–171. Bartholomew, D. J., Knott, M. and Moustaki, I. (2011). Latent variable models and factor analysis: A unified approach. John Wiley & Sons. Bin...

work page 2013
[2]

Diabetes Care

diagnosis and classification of diabetes: Standards of care in diabetes—2025. Diabetes Care

work page 2025
[3]

Income” on the response variables without taking into account the effect of “Age

While it is infeasible in general to discern the ground truth coefficient values for real data analyses, it is apparent that many coefficients are well aligned with basic knowledge of the relationships between the covariates and the response variables. For instance, fasting glucose levels are known to be higher in individuals with diabetes (ElSayed et al....

work page 2025
[4]

Thus, we have 1 n B(Z − ˜Z)(Z − ˜Z)T BT − B E h 1 n (Z − ˜Z)(Z − ˜Z)T i BT F = Op M p r K n

Under Assumptions 1 - 4 and the regime of K < n, we have the following bound: E 1 n B(Z − ˜Z)(Z − ˜Z)T BT − B E h 1 n (Z − ˜Z)(Z − ˜Z)T i BT F ≤ C · M p · r K n for some fixed constant C >0. Thus, we have 1 n B(Z − ˜Z)(Z − ˜Z)T BT − B E h 1 n (Z − ˜Z)(Z − ˜Z)T i BT F = Op M p r K n . Proof. Note that for a K × 1 centered sub-Gaussian random vector V , if ...

work page 2017
[5]

true” effect of “Income

We assume the latter model shows the “true” effect of “Income” on “Hypertension”, “Diabetes”, and “Arthritis” after accounting for the effect of “Age”. Comparing the coefficient values corresponding to “Income” on “Arthritis” between Table 4a and Table 4b, it is apparent that G-hive’s 0 .14 is closer to the “true” value of 0 .07 compared to the more posit...

work page 2017
[6]

≥ λp and ˆλ1 ≥

Let Σ, ˆΣ ∈ Rp×p be symmetric with eigenvalues λ1 ≥ ... ≥ λp and ˆλ1 ≥ ... ≥ ˆλp. Fix 1 ≤ r ≤ s ≤ p and assume that min(λr−1 − λr, λs − λs+1) > 0 where we define λ0 = ∞ and λp+1 = −∞. Let d = s − r + 1 and let V = [vr, vr+1, ..., vs] ∈ Rp×d and ˆV = [ˆvr, ˆvr+1, ...,ˆvs] ∈ Rp×d have orthonormal columns satisfying 46 Σvj = λjvj and ˆΣˆvj = ˆλj ˆvj for j = ...

work page 2015
[7]

ERIC. Ning, Y. and Liu, H. (2017). A general theory of hypothesis tests and confidence regions for sparse high dimensional models. The Annals of Statistics 45 158–195. Ouyang, J. , Tan, K. M. and Xu, G. (2023). High-dimensional inference for generalized linear models with hidden confounding. The Journal of Machine Learning Research 24 14030–14090. Parikh,...

work page 2017
[9]

Then, for every t ≥ 0, we have P ( nX i=1 aiXi ≥ t ) ≤ 2 exp − cB · min t2 V 2 · ||a||2 2 , t V · ||a||max where V = max 1≤i≤n ||Xi||ψ1

(Bernstein Inequality) Let X1, ..., Xn be independent, mean-zero sub-exponential random vari- ables, and let a = (a1, ..., an) ∈ Rn. Then, for every t ≥ 0, we have P ( nX i=1 aiXi ≥ t ) ≤ 2 exp − cB · min t2 V 2 · ||a||2 2 , t V · ||a||max where V = max 1≤i≤n ||Xi||ψ1 . The proof for this lemma can be found in Vershynin (2018). Note that if you setT = min...

work page 2018

Show all 18 references

[10]

The same logic can be applied to A to show that its smallest eigenvalue is on the order of p with high probability

By standard high probability bounds for Gaussian covariance estimation (Vershynin (2018)) and Weyl’s inequality (Horn and Johnson (1994)), it can be shown that the smallest eigenvalue of BT B is on the order of M . The same logic can be applied to A to show that its smallest e...

2018
[11]

Do you now smoke cigarettes every day, some days, or not at all?

(Concentration of the Euclidean Norm of α−Sub-exponential Random Variables) Let X1, ..., Xn be independent, mean-zero α-sub-exponential random variables satisfying ||Xi||ψα ≤ V , for some α ∈ (0, 1] and E(X 2 i ) ≤ w2 i . For an n × n matrix Q, let A = QT Q = (aij). Then for a...

2021
[12]

We know from Assumption 4 that λj Σ ≍ M/p for 1 ≤ j ≤ K, and λj Σ = 0 for K + 1 ≤ j ≤ M

Recall that Σ = B Σ−1 Z + AT A −1 BT . We know from Assumption 4 that λj Σ ≍ M/p for 1 ≤ j ≤ K, and λj Σ = 0 for K + 1 ≤ j ≤ M . Similar to the derivation in the proof for Theorem 3, we know by Weyl’s inequality that for all 1 ≤ j ≤ M , λj( ˆΣ) − λj(Σ) ≤ ˆΣ − Σ op ≲ M p p√ M +...

2022
[15]

Let a ∈ Rn

(Concentration for α − Sub-exponential Random Variables) Let X1, ..., Xn be independent, mean-zero α-sub-exponential random variables satisfying ||Xi||ψα ≤ V , for some α ∈ (0, 1]. Let a ∈ Rn. For any t >0, we have P ( nX i=1 aiXi ≥ t ) ≤ 2 exp − cα · min t2 V 2 · ||a||2 2 , t...

2021
[47]

W ang, J., Zhao, Q., Hastie, T

Cambridge university press. W ang, J., Zhao, Q., Hastie, T. and Owen, A. B. (2017). Confounder adjustment in multiple hypothesis testing. Annals of statistics 45

2017
[48]

High dimensional covariance matrix estimation using a factor model

F an, J., F an, Y.and L v, J.(2008). High dimensional covariance matrix estimation using a factor model. Journal of Econometrics 147 186–197. F an, J., Liao, Y. and Mincheva, M. (2013). Large covariance estimation by thresholding principal orthogonal complements. Journal of th...

2008
[232]

, Guo, B

Chang, J. , Guo, B. and Yao, Q. (2015). High dimensional stochastic regression with latent factors, endogeneity and nonlinearity. Journal of Econometrics 189 297–312. Chicco, D. and Rovelli, C. (2019). Computational prediction of diagnosis and feature selection on mesothelioma...

2015
[1320]

Horn, R. A. and Johnson, C. R. (1994). Topics in matrix analysis . Cambridge university press. Hu, F. B. , Manson, J. E. , Stampfer, M. J. , Colditz, G. , Liu, S. , Solomon, C. G. and Willett, W. C. (2001). Diet, lifestyle, and the risk of type 2 diabetes mellitus in women. Ne...

1994
[1863]

W ang, Y.and Shah, R. (2025). Latent confounding in high-dimensional nonlinear models. arXiv preprint arXiv:2508.06274 . Wedderburn, R. W. (1974). Quasi-likelihood functions, generalized linear models, and the gauss—newton method. Biometrika 61 439–447. White, H. (1982). Maxim...

2025 arXiv
[2022]

ElSayed, N. A. , McCoy, R. G. , Aleppo, G. , Balapattabi, K., Beverly, E. A. , Briggs Early, K., Bruemmer, D., Ebekozien, O., Echouffo-Tcheugui, J. B., Ekhlaspour, L. et al. (2025)

2025

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

Ahn, S. C. and Horenstein, A. R. (2013). Eigenvalue ratio test for the number of factors. Econometrica 81 1203–1227. Bai, J. (2003). Inferential theory for factor models of large dimensions. Econometrica 71 135–171. Bartholomew, D. J., Knott, M. and Moustaki, I. (2011). Latent variable models and factor analysis: A unified approach. John Wiley & Sons. Bin...

work page 2013

[2] [2]

Diabetes Care

diagnosis and classification of diabetes: Standards of care in diabetes—2025. Diabetes Care

work page 2025

[3] [3]

Income” on the response variables without taking into account the effect of “Age

While it is infeasible in general to discern the ground truth coefficient values for real data analyses, it is apparent that many coefficients are well aligned with basic knowledge of the relationships between the covariates and the response variables. For instance, fasting glucose levels are known to be higher in individuals with diabetes (ElSayed et al....

work page 2025

[4] [4]

Thus, we have 1 n B(Z − ˜Z)(Z − ˜Z)T BT − B E h 1 n (Z − ˜Z)(Z − ˜Z)T i BT F = Op M p r K n

Under Assumptions 1 - 4 and the regime of K < n, we have the following bound: E 1 n B(Z − ˜Z)(Z − ˜Z)T BT − B E h 1 n (Z − ˜Z)(Z − ˜Z)T i BT F ≤ C · M p · r K n for some fixed constant C >0. Thus, we have 1 n B(Z − ˜Z)(Z − ˜Z)T BT − B E h 1 n (Z − ˜Z)(Z − ˜Z)T i BT F = Op M p r K n . Proof. Note that for a K × 1 centered sub-Gaussian random vector V , if ...

work page 2017

[5] [5]

true” effect of “Income

We assume the latter model shows the “true” effect of “Income” on “Hypertension”, “Diabetes”, and “Arthritis” after accounting for the effect of “Age”. Comparing the coefficient values corresponding to “Income” on “Arthritis” between Table 4a and Table 4b, it is apparent that G-hive’s 0 .14 is closer to the “true” value of 0 .07 compared to the more posit...

work page 2017

[6] [6]

≥ λp and ˆλ1 ≥

Let Σ, ˆΣ ∈ Rp×p be symmetric with eigenvalues λ1 ≥ ... ≥ λp and ˆλ1 ≥ ... ≥ ˆλp. Fix 1 ≤ r ≤ s ≤ p and assume that min(λr−1 − λr, λs − λs+1) > 0 where we define λ0 = ∞ and λp+1 = −∞. Let d = s − r + 1 and let V = [vr, vr+1, ..., vs] ∈ Rp×d and ˆV = [ˆvr, ˆvr+1, ...,ˆvs] ∈ Rp×d have orthonormal columns satisfying 46 Σvj = λjvj and ˆΣˆvj = ˆλj ˆvj for j = ...

work page 2015

[7] [7]

ERIC. Ning, Y. and Liu, H. (2017). A general theory of hypothesis tests and confidence regions for sparse high dimensional models. The Annals of Statistics 45 158–195. Ouyang, J. , Tan, K. M. and Xu, G. (2023). High-dimensional inference for generalized linear models with hidden confounding. The Journal of Machine Learning Research 24 14030–14090. Parikh,...

work page 2017

[8] [9]

Then, for every t ≥ 0, we have P ( nX i=1 aiXi ≥ t ) ≤ 2 exp − cB · min t2 V 2 · ||a||2 2 , t V · ||a||max where V = max 1≤i≤n ||Xi||ψ1

(Bernstein Inequality) Let X1, ..., Xn be independent, mean-zero sub-exponential random vari- ables, and let a = (a1, ..., an) ∈ Rn. Then, for every t ≥ 0, we have P ( nX i=1 aiXi ≥ t ) ≤ 2 exp − cB · min t2 V 2 · ||a||2 2 , t V · ||a||max where V = max 1≤i≤n ||Xi||ψ1 . The proof for this lemma can be found in Vershynin (2018). Note that if you setT = min...

work page 2018