Pith. sign in

REVIEW 4 major objections 5 minor 74 references

Extended Fiducial Inference for Individual Treatment Effects via Deep Neural Networks

T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A Double-NN method uses deep networks to estimate individual treatment effects with calibrated uncertainty, and proves the model size may grow as O(n^ζ) for any ζ < 1.

desk verdict The paper sells consistency as calibration; the underlying Double-NN method is real and worth refereeing, but the advertised UQ guarantee for O(n^ζ) models is not proven. read the letter →

arxiv 2505.01995 v1 pith:XKNKWVZS submitted 2025-05-04 stat.ML cs.LGmath.STstat.COstat.TH

Sehwan Kim , Faming Liang This is my paper · ORCID
classification stat.MLcs.LGmath.STstat.COstat.TH MSC 62F9962G0868T07
keywords extendedfiducialinferenceindividualtreatmenteffectsdeepneuralnetworksuncertaintyquantificationcausalDouble-NNstochasticgradientMCMClargemodels
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper introduces Double-NN, a method that estimates individual treatment effects and quantifies their uncertainty by combining deep neural networks with extended fiducial inference (EFI). The central claim is that the inverse mapping that recovers the parameters of the treatment and control networks remains a consistent estimator even when the model dimension grows with the sample size $n$ as $O(n^{\zeta})$ for any $0 \leq \zeta < 1$, provided the EFI network's layers are narrow in the sense that $\sum_{l=1}^H d_l \prec n$. If true, this would extend trustworthy uncertainty quantification to large deep models, where classical central-limit-theorem guarantees stop at $\zeta < 1/2$ and Bayesian posterior consistency is currently proved only up to $\zeta < 1/4$. The paper also reports that Double-NN prediction intervals for individual treatment effects are shorter and closer to nominal coverage than conformal quantile regression intervals on simulated and real data.

What carries the argument

The load-bearing object is the estimated inverse mapping $G^*(Y_n,X_n,Z_n)$, approximated by a DNN $\hat{g}(y,x,z,w_n)$ with an output layer matching $\dim(\theta)$. EFI treats the unobserved errors $z_i$ as latent variables, defines the energy $U_n = \sum_i d(y_i,x_i,z_i,\bar{\theta}) + \eta \sum_i \|\hat{\theta}_i - \bar{\theta}\|^2$, and with $\epsilon \downarrow 0$ the extended fiducial density of $\theta$ is obtained by pushing the conditional distribution of $Z_n$ through $G^*$. To prove consistency for large models, the paper replaces the EFI network by an auxiliary stochastic neural network (StoNet) with small injected noises; each layer then forms a low-dimensional sparse regression, and a high-dimensional regression lemma plus an imputation-regularized optimization convergence argument yield consistency of $w^*_n$. The contractivity of the resulting iteration mapping $M(w_n)$, stated as Assumption 6, is what turns per-layer estimation error into a global consistent inverse mapping.

What would settle it

Estimate the largest singular value of $\partial M(w_n)/\partial w_n$ by finite differences during the IRO/EFI iterations on actual datasets from the paper's Examples 1 and 2; if the largest value reaches or exceeds 1 at any iterated $w_n$, the contraction premise behind Theorem 3.1 fails for that setting. Separately, run the method with $\dim(\theta) = O(n^{0.9})$ and heavy-tailed or dependent errors, and check whether the fiducial intervals keep nominal coverage; a coverage breakdown would show the stated large-model guarantee does not extend beyond the Gaussian-noise setting.

Watch

Extended reading notes

Core claim

The paper's central claim is that, under assumptions stated in the supplement, the EFI inverse mapping $G^*(Y_n,X_n,Z_n) = \frac{1}{n}\sum_{i=1}^n \hat{g}(y_i,x_i,z_i,w^*_n)$ is a consistent estimator of the inverse function $\theta = G(Y_n,X_n,Z_n)$ when $\sum_{l=1}^H d_l \prec n$ and the model dimension is $\dim(\theta) = d_H = O(n^{\zeta})$ for $0 \leq \zeta < 1$. From this consistency, the extended fiducial density propagates uncertainty from the imputed latent errors $Z_n$ to $\theta$, so fiducial samples of $\bar{\theta}$ produce prediction intervals for the predictive individual treatment effect $\tilde{\tau}_i$. The paper therefore claims a rigorous framework for uncertainty quantification of deep neural networks whose size follows a neural scaling law $\dim(\theta) \prec n$, and supports the claim with simulations in which Double-NN intervals are shorter and better calibrated than conformal quantile regression intervals.

Load-bearing premise

The proof hinges on Assumption 6, which says the iteration map $M(w_n)$ is a strict contraction with largest singular value bounded below 1 uniformly; the paper calls this 'reasonable to assume' and cites earlier work but does not numerically verify it, and the theory also assumes the true data-generating process is a neural network while all simulations use non-neural functions.

Editorial extensions

If this is right

  • If Theorem 3.1 is correct, fiducial-based uncertainty quantification remains valid for deep models whose parameter count grows like $O(n^{\zeta})$ for every $\zeta < 1$, not just the $\zeta < 1/2$ allowed by classical CLT arguments.
  • The Double-NN construction gives prediction intervals for predictive individual treatment effects in all three test-set situations: control-observed, treatment-observed, and covariates-only subjects.
  • The same inverse-mapping consistency supports inference for the conditional average treatment effect function $\tau(x)$, not only point estimates, because each fiducial $\bar{\theta}$ sample yields a $\tau$-network estimate.
  • The method targets the zero-energy manifold as $\epsilon \downarrow 0$, and the reported simulations recover the true latent errors even when the true model is not a neural network; the paper presents this as empirical robustness, not a proved theorem.
  • Under the empirical neural scaling law $\dim(\theta) = O(n^{\zeta})$ with $\zeta \in [0.5,1)$, the theory gives a formal uncertainty-quantification framework for large-scale DNNs, provided a matching depth growth $H = O(n^{\beta})$ with $\beta < 1 - \zeta$ is chosen.

Reading between the lines

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

  • A direct test of the paper's main premise would be to compute the largest singular value of $\partial M(w_n)/\partial w_n$ on the simulation datasets; if it is not below 1 along the algorithm's trajectory, the consistency theorem's Assumption 6 is not satisfied in practice.
  • Because the EFI uncertainty is carried by imputed errors rather than by the approximating function class, the same Double-NN construction could be extended to average treatment effects and to propensity-score-adjusted estimators by changing only the function $b(\theta)$ in the extended fiducial density integral; the paper notes this extension but does not implement it.
  • The paper's link to conformal prediction through generalized fiducial inference suggests EFI intervals might eventually be made distribution-free by an appropriate calibration step, though that is not established here.
  • For over-parameterized models with $\dim(\theta) \succeq n$, imposing sparsity on $\theta$ and applying post-selection EFI is a natural next step; the paper lists it as future work, so any such guarantee would be an extension, not a claim of this paper.
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

4 major / 5 minor

Summary. The paper proposes the Double-NN method, an extended fiducial inference (EFI) procedure for individual treatment effect (ITE) estimation, in which deep neural networks model the control response and treatment effect functions and an additional neural network approximates the inverse mapping from data and imputed latent errors to the model parameters. The central theoretical claim is that the dimension of the target DNN model may grow as O(n^ζ) for any 0 ≤ ζ < 1 while the method still provides proper uncertainty quantification, extending beyond the classical O(n^{1/2}) CLT range. The paper proves consistency of the learned inverse mapping under stated assumptions, presents an illustrative linear example, compares prediction intervals against conformal quantile regression on two simulations and two real datasets, and makes code publicly available.

Significance. If the advertised result were fully established, it would be a notable contribution to uncertainty quantification for large neural-network models and would position EFI as a theoretically grounded alternative to conformal methods for ITE inference. The paper includes reproducible code, detailed experimental settings, and a sensible benchmark against CQR, which are strengths. However, the main significance claim—proper uncertainty quantification for models of dimension O(n^ζ) with ζ approaching 1—is not actually derived in the theorems: Theorem 3.1 establishes only consistency of the inverse mapping, and the bridge to coverage guarantees is an informal Slutsky argument. The proof also relies on an unverified contractive-mapping assumption. These gaps currently limit the significance of the theoretical contribution.

major comments (4)
  1. [Section 3.1, Theorem 3.1; Section 6; Supplement §1.1] The abstract and Section 6 claim that the paper theoretically proves proper quantification of uncertainty for model dimension O(n^ζ), but Theorem 3.1 only proves that G*(Y_n,X_n,Z_n) is a consistent estimator of the inverse function. The supplement's §1.1 invokes Slutsky's theorem to propagate uncertainty from Z_n to θ, but consistency of a point estimator does not imply that the empirical fiducial measure (S7) yields intervals with nominal coverage. No theorem in the paper provides a CLT or concentration result for the fiducial samples in the growing-dimension regime, and no theorem bounds the coverage error of the §5.1 prediction intervals. This is a load-bearing gap between the formal results and the central advertised contribution.
  2. [Supplement §1.2.2, Assumption 6] The proof of Theorem S2, and therefore of Theorem 3.1, requires that the imputation-regularized optimization mapping M(w_n) be differentiable and contractive with largest singular value bounded by λ* < 1 uniformly over W_n for sufficiently large n and almost every dataset. The paper states that this is 'reasonable to assume' and cites Nielsen (2000) and Liang et al. (2018a), but it does not verify the contractivity condition for the DNN/EFI setting. If this assumption fails, the convergence of \hat{w}_n^{(t)} to \tilde{w}_n^* is not established. A proof or a concrete sufficient condition for Assumption 6 is needed.
  3. [Section 3.2 and Section 5.2] The theoretical development assumes that the true data-generating model is a neural network, thereby omitting approximation error. In contrast, all simulation studies use non-neural functions: Example 1 uses a linear control response with a sigmoid-transformed treatment effect, and Example 2 uses a rational control response. The paper asserts robustness to approximation error without proof. As a result, the numerical evidence does not directly validate the conditions of Theorem 3.1, and the claim of broad applicability requires either a misspecification analysis or a clear statement that the theory covers only correctly specified neural-network models.
  4. [Supplement, Lemma S3 proof] For the ReLU activation function, the proof of Lemma S3 relies on the informal assertion that a neuron belonging to the true neuron set cannot be constantly zero over all samples, and on the existence of a threshold q_min for the expected squared gradient. This threshold is not stated as a formal condition in Assumptions 1–6 or in Theorem S1/S2, leaving another unverified step in the chain of arguments leading to the consistency of \hat{w}_n^*.
minor comments (5)
  1. [Supplement §1.2.1, after equation (S10)] The text says convergence holds if max{σ_1,...,σ_H} ≺ τ(d_1,...,d_H) and "ϵ→∞"; this should read "ϵ is sufficiently small" or "ϵ↓0" to be consistent with the rest of the paper.
  2. [Assumption 5] The phrase "priro distribution" should be corrected to "prior distribution".
  3. [Section 5.3.2] The heading "NSLM" should be "NLSM" to match the dataset name used in the text and the reference to the National Study of Learning Mindsets.
  4. [Section 5.3.1 and Figure 4] The spelling "LaLonde" is used in the section heading while "Lalonde" appears in the body text; please standardize. In addition, Figure 4 is based on a single train/test split, so the comparison shown there would be more informative with repeated splits or error bars.
  5. [Section 5.1] The text sometimes refers to the intervals as "confidence intervals" and at other times as "prediction intervals"; since the target is the predictive ITE, the terminology should be made consistent.

Circularity Check

1 steps flagged · score 4.0 of 10

One load-bearing self-citation (unverified contractivity) conditions the large-model consistency result; the coverage claim is an inference gap rather than a circular reduction.

  1. self citation load bearing [Supplement §1.2.2 (Assumption 6 and Theorem S2), used by Theorem 3.1]
    "As argued in Liang et al. (2018a) and Nielsen (2000), it is reasonable to assume that the mapping is contractive. ... Assumption 6. The mapping M(w_n) is differentiable. ... There exists a number λ∗ < 1 such that λmax(M_{w_n}) ≤ λ∗ for all w_n ∈ W_n ... Theorem S2 ... Proof: This lemma directly follows from Theorem 4 of Liang et al. (2018a) that the estimator \hat w^{(t)}_n is consistent when both n and t are sufficiently large."

    The advertised large-model consistency result (Theorem 3.1) is routed through Theorem S2, whose proof is a one-line appeal to Theorem 4 of Liang et al. (2018a), a prior paper by the same author. The key contraction condition is not verified in the present paper; it is asserted as 'reasonable to assume' and attributed to that same self-authored work plus Nielsen (2000). Thus the O(n^ζ) consistency guarantee is not derived from first principles here but is conditioned on an imported assumption from a self-citation chain. This is load-bearing, although it is not an equation-level identity, and the rest of the ITE methodology still has independent numerical content.

full rationale

The central point-estimation result, Theorem 3.1, is a genuine consistency theorem whose proof invokes several previously published results: the StoNet equivalence (Liang et al., 2022b), IRO convergence (Liang et al., 2018a), sparse deep learning theory (Sun et al., 2022), and the high-dimensional regression lemma (Song and Liang, 2023). Citing these peer-reviewed results is not circular by itself. The numerical comparisons against CQR also provide independent external benchmarks. The main circularity-adjacent concern is Assumption 6: the contractivity of M(w_n) is assumed on the authority of self-authored prior work and not verified, and Theorem S2's proof is just a citation to that same work, making the large-model consistency guarantee conditional. Separately, the advertised uncertainty-quantification guarantee does not follow from consistency alone: Supplement §1.1 states 'by Slutsky's theorem, the uncertainty of Z_n can be propagated to θ', but no theorem bounds the coverage gap, gives a CLT or concentration result in the growing-dimension regime, or treats the non-neural true models used in the simulations. These are inference gaps rather than circular reductions, so I do not count them as circular steps for scoring.

Assumptions & free parameters 3 free parameters · 7 assumptions · 0 invented entities

The central theorem rests on a chain of assumptions: causal identifiability, an additive-noise data generator with known error distribution, the neural-network form of the true model, compactness and identifiability of the parameter space, sparsity-inducing mixture priors, and a contractive IRO mapping. The most fragile entries are the neural-network truth assumption and the unverified contractivity condition. Experimental hyperparameters η, ϵ, and prior sparsity settings are chosen by hand and affect the numerical results.

free parameters (3)
  • EFI energy temperature ϵ = 0.1
    Set to 0.1 in all experiments; the theory requires ϵ to be sufficiently small for the zero-energy manifold argument.
  • EFI regularization coefficient η = 10 (ITE examples), 500 (ATE example)
    Chosen by hand to balance fitting error and inverse-mapping regularity; affects interval lengths and coverage in the numerical studies.
  • Mixture Gaussian prior sparsity parameters (ρ, σ0, σ1) = ρ=0.01, σ0=0.01, σ1=1
    Set by hand for the experiments; the theory only specifies asymptotic growth conditions, not these exact values.
assumptions (7)
  • domain assumption Strong ignorability and overlap for causal identifiability
    Section 1, used to define ITE and justify estimation from observational data.
  • domain assumption Additive noise data-generating model with known error distribution π0
    Equation (1) in Section 1; the EFI derivation assumes the standardized random errors z_i follow a known distribution (e.g., N(0,1)).
  • ad hoc to paper True data-generating model is a neural network
    Section 3.2: 'we actually assume that the true data-generating model Y=f(X,Z,θ) is a neural network'; required for Theorem 3.1.
  • ad hoc to paper Assumption 6: IRO mapping M(w_n) is contractive with λ_max ≤ λ* < 1
    Supplement §1.2.2; required for Theorem S2 and Theorem 3.1; the paper calls it 'reasonable to assume' but does not verify it.
  • domain assumption Asymptotic equivalence between DNN and auxiliary StoNet
    Supplement §1.2.1, equation (S10), citing Liang et al. (2022b); used to transfer consistency from the stochastic network to the EFI network.
  • standard math DNN universal approximation capability
    Section 1, citing Hornik et al. (1989) and Kidger and Lyons (2020); used to justify approximating c(x) and τ(x) by DNNs.
  • ad hoc to paper Compact parameter space W_n and identifiability up to loss-invariant transformations
    Assumptions 1-2, Supplement §1.1; needed for posterior consistency and for the uniqueness claims used in Lemma S1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Extended Fiducial Inference for Individual Treatment Effects via Deep Neural Networks." pith.science (2026). https://pith.science/paper/XKNKWVZS

@misc{pith2026250501995,
  author       = {Pith},
  title        = {Pith review of: Extended Fiducial Inference for Individual Treatment Effects via Deep Neural Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XKNKWVZS}},
  note         = {Machine review of arXiv:2505.01995}
}
abstract

Individual treatment effect estimation has gained significant attention in recent data science literature. This work introduces the Double Neural Network (Double-NN) method to address this problem within the framework of extended fiducial inference (EFI). In the proposed method, deep neural networks are used to model the treatment and control effect functions, while an additional neural network is employed to estimate their parameters. The universal approximation capability of deep neural networks ensures the broad applicability of this method. Numerical results highlight the superior performance of the proposed Double-NN method compared to the conformal quantile regression (CQR) method in individual treatment effect estimation. From the perspective of statistical inference, this work advances the theory and methodology for statistical inference of large models. Specifically, it is theoretically proven that the proposed method permits the model size to increase with the sample size $n$ at a rate of $O(n^{\zeta})$ for some $0 \leq \zeta<1$, while still maintaining proper quantification of uncertainty in the model parameters. This result marks a significant improvement compared to the range $0\leq \zeta < \frac{1}{2}$ required by the classical central limit theorem. Furthermore, this work provides a rigorous framework for quantifying the uncertainty of deep neural networks under the neural scaling law, representing a substantial contribution to the statistical understanding of large-scale neural network models.

Figures

Figures reproduced from arXiv: 2505.01995 by the authors.

Figure 1
Figure 1. Illustration of the EFI network (Liang et al., 2024), where the orange nodes and [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Demonstration of the Double-NN method for a dataset simulated from (24): (left) [PITH_FULL_IMAGE:figures/full_fig_p016_2.png] view at source ↗
Figure 3
Figure 3. Demonstration of the Double-NN method for a dataset simulated from (26): (left) [PITH_FULL_IMAGE:figures/full_fig_p017_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Comparison of prediction intervals resulting from Double-NN (labeled as EFI) and [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]
Figure 5
Figure 5. Figure 5: Comparison of the average length of intervals obtained by the Double-NN (labeled as [PITH_FULL_IMAGE:figures/full_fig_p020_5.png]
Figure 6
Figure 6. Figure 6: Fractions of the intervals obtained by Double-NN (labeled as EFI) and CQR (labeled [PITH_FULL_IMAGE:figures/full_fig_p021_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

74 extracted references · 69 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address author booktitle chapter edition editor howpublished institution journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.all := #1 'mid.sentence := #2 '...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in " " * FUNCTION format....

  3. [3]

    and Robins, J

    Bang, H. and Robins, J. M. (2005), Doubly Robust Estimation in Missing Data and Causal Inference Models, Biometrics, 61, 962--972

    work page 2005

  4. [4]

    (2019), Optimal approximation with sparsely connected deep neural networks, SIAM Journal on Mathematics of Data Science, 1, 8--45

    Bolcskei, H., Grohs, P., Kutyniok, G., and Petersen, P. (2019), Optimal approximation with sparsely connected deep neural networks, SIAM Journal on Mathematics of Data Science, 1, 8--45

    work page 2019

  5. [5]

    (1998), Arcing classifier (with discussion and a rejoinder by the author), Annals of Statistics, 26, 801--849

    Breiman, L. (1998), Arcing classifier (with discussion and a rejoinder by the author), Annals of Statistics, 26, 801--849

    work page 1998

  6. [6]

    --- (2001), Random forests, Machine learning, 45, 5--32

    work page 2001

  7. [7]

    Caron, A., Baio, G., and Manolopoulou, I. (2022), Estimating Individual Treatment Effects using Non-Parametric Regression Models: a Review , Journal of the Royal Statistical Society Series A: Statistics in Society, 185, 1115--1149

    work page 2022

  8. [8]

    M., Feller, A., Murray, J., Woody, S., and Yeager, D

    Carvalho, C. M., Feller, A., Murray, J., Woody, S., and Yeager, D. S. (2019), Assessing Treatment Effect Variation in Observational Studies: Results from a Data Challenge, Observational Studies

    work page 2019

Show all 74 references
  1. [9]

    A., and Burgette, L

    Cefalu, M., Ridgeway, G., McCaffrey, D., Morral, A., Griffin, B. A., and Burgette, L. (2021), Package `twang': Toolkit for Weighting and Analysis of Nonequivalent Groups, R Package

    2021

  2. [10]

    (2014), Stochastic gradient hamiltonian monte carlo, in International conference on machine learning, pp

    Chen, T., Fox, E., and Guestrin, C. (2014), Stochastic gradient hamiltonian monte carlo, in International conference on machine learning, pp. 1683--1691

    2014

  3. [11]

    A., George, E

    Chipman, H. A., George, E. I., and McCulloch, R. E. (2010), BART: Bayesian Additive Regression Trees, The Annals of Applied Statistics, 4, 266--298

    2010

  4. [12]

    (2019), An adaptive empirical Bayesian method for sparse deep learning, Advances in neural information processing systems, 32

    Deng, W., Zhang, X., Liang, F., and Lin, G. (2019), An adaptive empirical Bayesian method for sparse deep learning, Advances in neural information processing systems, 32

    2019

  5. [13]

    J., and Maass, P

    Dittmer, S., King, E. J., and Maass, P. (2018), Singular Values for ReLU Layers, IEEE Transactions on Neural Networks and Learning Systems, 31, 3594--3605

    2018

  6. [14]

    and Hill, J

    Dorie, V. and Hill, J. L. (2020), Bartcause: Causal Inference using Bayesian Additive Regression Trees [R package bartCause version 1.0-4], R Package

    2020

  7. [15]

    (2021), Deep Neural Networks for Estimation and Inference, Econometrica, 89, 181--213

    Farrell, M., Liang, T., and Misra, S. (2021), Deep Neural Networks for Estimation and Inference, Econometrica, 89, 181--213

    2021

  8. [16]

    Fisher, R. A. (1935), The fiducial argument in statistical inference, Annals of Eugenics, 6, 391--398

    1935

  9. [17]

    C., Taylor, J

    Foster, J. C., Taylor, J. M., and Ruberg, S. J. (2011), Subgroup identification from randomized clinical trial data, Statistics in Medicine, 30

    2011

  10. [18]

    Fraser, D. A. S. (1966), Structural probability and a generalization, Biometrika, 53, 1--9

    1966

  11. [19]

    --- (1968), The Structure of Inference, New York-London-Sydney: John Wiley & Sons

    1968

  12. [20]

    and Yang, S

    Guan, Q. and Yang, S. (2019), A Unified Framework for Causal Inference with Multiple Imputation Using Martingale, arXiv: Methodology

    2019

  13. [21]

    R., Murray, J

    Hahn, P. R., Murray, J. S., and Carvalho, C. M. (2020), Bayesian Regression Tree Models for Causal Inference: Regularization, Confounding, and Heterogeneous Effects (with Discussion) , Bayesian Analysis, 15, 965 -- 2020

    2020

  14. [22]

    (2009), On generalized fiducial inference, Statistica Sinica, 19, 491--544

    Hannig, J. (2009), On generalized fiducial inference, Statistica Sinica, 19, 491--544

    2009

  15. [23]

    Hannig, J., Iyer, H., Lai, R. C. S., and Lee, T. C. M. (2016), Generalized Fiducial Inference: A Review and New Results, Journal of the American Statistical Association, 111, 1346--1361

    2016

  16. [24]

    F., Jun, H., Kianinejad, H., Patwary, M

    Hestness, J., Narang, S., Ardalani, N., Diamos, G. F., Jun, H., Kianinejad, H., Patwary, M. M. A., Yang, Y., and Zhou, Y. (2017), Deep Learning Scaling is Predictable, Empirically, ArXiv, abs/1712.00409

    2017 arXiv

  17. [25]

    Higham, N. J. and Cheng, S. H. (1998), Modifying the inertia of matrices arising in optimization, Linear Algebra and its Applications, 261--279

    1998

  18. [26]

    Hill, J. L. (2011), Bayesian Nonparametric Modeling for Causal Inference, Journal of Computational and Graphical Statistics, 20, 217--240

    2011

  19. [27]

    (1991), Approximation capabilities of multilayer feedforward networks, Neural Networks, 4, 251--257

    Hornik, K. (1991), Approximation capabilities of multilayer feedforward networks, Neural Networks, 4, 251--257

    1991

  20. [28]

    (1989), Multilayer feedforward networks are universal approximators, Neural Networks, 2, 359--366

    Hornik, K., Stinchcombe, M., and White, H. (1989), Multilayer feedforward networks are universal approximators, Neural Networks, 2, 359--366

    1989

  21. [29]

    (2004), Nonparametric Estimation of Average Treatment Effects Under Exogeneity: A Review, The Review of Economics and Statistics, 86, 4--29

    Imbens, G. (2004), Nonparametric Estimation of Average Treatment Effects Under Exogeneity: A Review, The Review of Economics and Statistics, 86, 4--29

    2004

  22. [30]

    Imbens, G. W. and Rubin, D. B. (2015), Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction, USA: Cambridge University Press

    2015

  23. [31]

    and Montanari, A

    Javanmard, A. and Montanari, A. (2014), Confidence iReferntervals and hypothesis testing for high-dimensional regression, Journal of Machine Learning Research, 15, 2869--2909

    2014

  24. [32]

    (2007), Bayesian variable selection for high dimensional generalized linear models: convergence rates of the fitted densities, The Annals of Statistics, 35, 1487--1511

    Jiang, W. (2007), Bayesian variable selection for high dimensional generalized linear models: convergence rates of the fitted densities, The Annals of Statistics, 35, 1487--1511

    2007

  25. [33]

    and Lyons, T

    Kidger, P. and Lyons, T. (2020), Universal Approximation with Deep Narrow Networks, Proceedings of Machine Learning Research, 125, 1--22

    2020

  26. [34]

    (2023), Minimum width for universal approximation using ReLU networks on compact domain, ArXiv, abs/2309.10402

    Kim, N., Min, C., and Park, S. (2023), Minimum width for universal approximation using ReLU networks on compact domain, ArXiv, abs/2309.10402

    2023 arXiv

  27. [35]

    R., Sekhon, J

    K \"u nzel, S. R., Sekhon, J. S., Bickel, P. J., and Yu, B. (2019), Metalearners for estimating heterogeneous treatment effects using machine learning, Proceedings of the National Academy of Sciences of the United States of America, 116, 4156 -- 4165

    2019

  28. [36]

    (2016), Exact post-selection inference, with application to the Lasso, , 44, 907--927

    Lee, J., Sun, D., Sun, Y., and Taylor, J. (2016), Exact post-selection inference, with application to the Lasso, , 44, 907--927

    2016

  29. [37]

    and Candès, E

    Lei, L. and Candès, E. J. (2021), Conformal Inference of Counterfactuals and Individual Treatment Effects , Journal of the Royal Statistical Society Series B: Statistical Methodology, 83, 911--938

    2021

  30. [38]

    (2018 a ), An imputation--regularized optimization algorithm for high dimensional missing data problems and beyond, , 80, 899--926

    Liang, F., Jia, B., Xue, J., Li, Q., and Luo, Y. (2018 a ), An imputation--regularized optimization algorithm for high dimensional missing data problems and beyond, , 80, 899--926

    2018

  31. [39]

    (2024), Exended Fiducial Inference: Toward an Automated Process of Statistical Inference, , in press

    Liang, F., Kim, S., and Sun, Y. (2024), Exended Fiducial Inference: Toward an Automated Process of Statistical Inference, , in press

    2024

  32. [40]

    (2018 b ), Bayesian Neural Networks for Selection of Drug Sensitive Genes, , 113, 955--972

    Liang, F., Li, Q., and Zhou, L. (2018 b ), Bayesian Neural Networks for Selection of Drug Sensitive Genes, , 113, 955--972

    2018

  33. [41]

    (2022 a ), Markov neighborhood regression for high-dimensional inference, , 117, 1200--1214

    Liang, F., Xue, J., and Jia, B. (2022 a ), Markov neighborhood regression for high-dimensional inference, , 117, 1200--1214

    2022

  34. [42]

    (2022 b ), Nonlinear Sufficient Dimension Reduction with a Stochastic Neural Network, NeurIPS 2022

    Liang, S., Sun, Y., and Liang, F. (2022 b ), Nonlinear Sufficient Dimension Reduction with a Stochastic Neural Network, NeurIPS 2022

    2022

  35. [43]

    J., and Ishwaran, H

    Lu, M., Sadiq, S., Feaster, D. J., and Ishwaran, H. (2018), Estimating Individual Treatment Effect in Observational Data Using Random Forest Methods, Journal of Computational and Graphical Statistics, 27, 209 -- 219

    2018

  36. [44]

    (2023), Fiducial inference viewed through a possibility-theoretic inferential model lens, , 215, 299--310

    Martin, R. (2023), Fiducial inference viewed through a possibility-theoretic inferential model lens, , 215, 299--310

    2023

  37. [45]

    C., Hannig, J., and Williams, J

    Murph, A. C., Hannig, J., and Williams, J. P. (2022), Generalized Fiducial Inference on Differentiable Manifolds, arXiv:2209.15473

    2022 arXiv

  38. [46]

    (2000), The stochastic EM algorithm: Estimation and asymptotic results, Bernoulli, 6, 457--489

    Nielsen, S. (2000), The stochastic EM algorithm: Estimation and asymptotic results, Bernoulli, 6, 457--489

    2000

  39. [47]

    (2020), Minimum Width for Universal Approximation, ArXiv, abs/2006.08859

    Park, S., Yun, C., Lee, J., and Shin, J. (2020), Minimum Width for Universal Approximation, ArXiv, abs/2006.08859

    2020 arXiv

  40. [48]

    and Voigtlaender, F

    Petersen, P. and Voigtlaender, F. (2018), Optimal approximation of piecewise smooth functions using deep ReLU neural networks, Neural Networks, 108, 296--330

    2018

  41. [49]

    (1986), On the central limit theorem in R ^p when p , Probability Theory and Related Fields, 73, 571--583

    Portnoy, S. (1986), On the central limit theorem in R ^p when p , Probability Theory and Related Fields, 73, 571--583

    1986

  42. [50]

    --- (1988), Asymptotic behavior of likelihood methods for exponential families when the number of parameters tend to infinity, , 16, 356--366

    1988

  43. [51]

    and Monro, S

    Robbins, H. and Monro, S. (1951), A Stochastic Approximation Method, Annals of Mathematical Statistics, 22, 400--407

    1951

  44. [52]

    M., Rotnitsky, A., and Zhao, L

    Robins, J. M., Rotnitsky, A., and Zhao, L. P. (1994), Estimation of regression coefficients when some regressors are not always observed, Journal of the American Statistical Association, 89, 846--866

    1994

  45. [53]

    Romano, Y., Patterson, E., and Cand \`e s, E. J. (2019), Conformalized Quantile Regression, in Neural Information Processing Systems

    2019

  46. [54]

    Rosenbaum, P. R. (1987), Model-Based Direct Adjustment, Journal of the American Statistical Association, 82, 387--394

    1987

  47. [55]

    --- (2002), Observational Studies (2nd edition), New York: Springer

    2002

  48. [56]

    Rubin, D. B. (1974), Estimating causal effects of treatments in randomized and nonrandomized studies. Journal of Educational Psychology, 66, 688--701

    1974

  49. [57]

    Schapire, R. E. (1990), The Strength of Weak Learnability, Machine Learning, 5, 197--227

    1990

  50. [58]

    (2020), Nonparametric regression using deep neural networks with ReLU activation function, The Annals of Statistics, 48, 1875--1897

    Schmidt-Hieber, J. (2020), Nonparametric regression using deep neural networks with ReLU activation function, The Annals of Statistics, 48, 1875--1897

    2020

  51. [59]

    G., Lopez, R., Mohan, R

    Sethuraman, M. G., Lopez, R., Mohan, R. V., Fekri, F., Biancalani, T., and Hutter, J.-C. (2023), NODAGS-Flow: Nonlinear Cyclic Causal Structure Learning, in International Conference on Artificial Intelligence and Statistics

    2023

  52. [60]

    and Vovk, V

    Shafer, G. and Vovk, V. (2008), A Tutorial on Conformal Prediction, J. Mach. Learn. Res., 9, 371–421

    2008

  53. [61]

    D., and Sontag, D

    Shalit, U., Johansson, F. D., and Sontag, D. A. (2017), Estimating individual treatment effect: generalization bounds and algorithms , in Proceedings of the 34th International Conference on Machine Learning , PMLR , pp. 3076--3085

    2017

  54. [62]

    and Liang, F

    Song, Q. and Liang, F. (2023), Nearly optimal Bayesian Shrinkage for high dimensional regression, China Science Mathematics, 66, 409--442

    2023

  55. [63]

    (2020), Extended Stochastic Gradient MCMC for Large-Scale Bayesian Variable Selection, Biometrika, 107, 997--1004

    Song, Q., Sun, Y., Ye, M., and Liang, F. (2020), Extended Stochastic Gradient MCMC for Large-Scale Bayesian Variable Selection, Biometrika, 107, 997--1004

    2020

  56. [64]

    and Liang, F

    Sun, Y. and Liang, F. (2022), A kernel-expanded stochastic neural network, Journal of the Royal Statistical Society Series B, 84, 547--578

    2022

  57. [65]

    (2022), Consistent Sparse Deep Learning: Theory and Computation, , 117, 1981--1995

    Sun, Y., Song, Q., and Liang, F. (2022), Consistent Sparse Deep Learning: Theory and Computation, , 117, 1981--1995

    2022

  58. [66]

    J., Foygel Barber, R., Candes, E., and Ramdas, A

    Tibshirani, R. J., Foygel Barber, R., Candes, E., and Ramdas, A. (2019), Conformal Prediction Under Covariate Shift, in Advances in Neural Information Processing Systems, eds. Wallach, H., Larochelle, H., Beygelzimer, A., d Alch\' e -Buc, F., Fox, E., and Garnett, R., Curran A...

    2019

  59. [67]

    (2014), On asymptotically optimal confidence regions and tests for high-dimensional models, Ann

    van de Geer, S., B\"uhlmann, P., Ritov, Y., and Dezeure, R. (2014), On asymptotically optimal confidence regions and tests for high-dimensional models, Ann. Statist., 42, 1166--1202

    2014

  60. [68]

    (2005), Algorithmic Learning in a Random World, Springer

    Vovk, V., Gammerman, A., and Shafer, G. (2005), Algorithmic Learning in a Random World, Springer

    2005

  61. [69]

    and Athey, S

    Wager, S. and Athey, S. (2018), Estimation and Inference of Heterogeneous Treatment Effects using Random Forests, Journal of the American Statistical Association, 113, 1228--1242

    2018

  62. [70]

    Williams, J. P. (2023), Model-free generalized fiducial inference,

    2023

  63. [71]

    S., Hanselman, P., Walton, G

    Yeager, D. S., Hanselman, P., Walton, G. M., Murray, J. S., Crosnoe, R., Muller, C., Tipton, E., Schneider, B., Hulleman, C. S., Hinojosa, C. P., Paunesku, D., Romero, C., Flint, K., Roberts, A., Trott, J., Iachan, R., Buontempo, J., Yang, S. M., Carvalho, C. M., Hahn, P. R., ...

    2019

  64. [72]

    Zabell, S. L. (1992), R. A. Fisher and Fiducial Argument, Statistical Science, 7, 369--387

    1992

  65. [73]

    and Sjölander, A

    Zetterqvist, J. and Sjölander, A. (2015), Doubly Robust Estimation with the R Package drgee, Epidemiologic Methods, 4, 69--86

    2015

  66. [74]

    and Zhang, S

    Zhang, C.-H. and Zhang, S. S. (2014), Confidence intervals for low dimensional parameters in high dimensional linear models, Journal of the Royal Statistical Society: Series B (Statistical Methodology), 76, 217--242

    2014

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.