- Rating
- 4(+1)
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
ProxMAR extends neural proximal causal inference to settings where outcome-inducing proxies are missing at random. It integrates conditional-mean imputation, inverse probability weighting, and augmented weighting into NMMR and DFPV, addressing missingness in bridge learning and causal-mean evaluation. The paper establishes conditional consistency guarantees and evaluates the resulting estimators on synthetic and semi-synthetic datasets.
Critical strengths
The paper addresses an important practical limitation of proximal estimators through a unified framework that corrects estimator-specific contributions without requiring explicit imputation of high-dimensional proxy values. Its treatment of both bridge learning and population averaging, together with an explicit product-remainder characterization of double robustness, provides a coherent methodological foundation. The experiments demonstrate substantial improvements over complete-case estimation in many settings, particularly at higher missingness rates.
Critical weaknesses
-
It seems that several consistency assumptions are not connected to the actual implementation. Verifying these conditions or characterizing the potential persistent bias when this assumption fails would be helpful for better transparency. Specifically:
- C.5: The paper provides sufficient conditions linking vanishing moment risk to causal-mean consistency, but does not establish that these conditions hold for the implemented neural class.
- C.6 and C.9: The required uniform control over candidate bridges is not established for the implemented kernel regressions. In particular, the fixed bandwidth grid may leave persistent smoothing bias.
- C.4: The theory assumes that the candidate neural class contains an exactly valid bridge, without verifying this for the implemented architecture.
- D.6: The theory requires learning errors to vanish sufficiently quickly relative to the inverse-stability factor. No concrete regime is established in which network capacity, regularization, and nuisance estimation jointly satisfy this requirement.
-
For the high-dimensional experiment, although has 100–300 coordinates, it is generated from only two latent factors through a linear Gaussian measurement model. The true bridge is additive and linear in , and nuisance regressions always condition on just four variables, . All three main experimental designs also omit observed covariates . This design provides limited evidence for complex settings. Maybe the authors can try varying latent dimension, proxy informativeness, nonlinear measurement structure, and the dimension of to better test the scalability claim.
Questions for authors
-
Appendix E.1.2 acknowledges that the Demand benchmark’s observation probabilities are not uniformly bounded away from zero, whereas Assumption 2 requires . Could the authors clarify which guarantees remain applicable to this benchmark and whether consistency can be established under weaker overlap or inverse-probability moment conditions?
-
For the continuous-treatment experiments, what properties of the neural class and DGP ensure Assumption D.7(ii), and how can the growth of be controlled?
-
Tables 1–5 compare ProxMAR primarily with complete-case estimation and simple mean imputation within DFPV and NMMR. Could the authors include a conditional distribution-based or multiple-imputation baseline for given , to assess the claimed advantage of recovering estimator-specific functions directly in terms of accuracy and computational cost?
-
For table1-5, it seems that AIPW is frequently worse than IMP or IPW, sometimes substantially so, particularly for DFPV. What could be the explanation? Under what practical conditions does its additional robustness become beneficial?
-
Could the authors explain the settings where complete-case estimation outperforms all corrections? For example, Demand NMMR at 30% missingness (Table 1) and Demand DFPV at (n=1000) (Table 2)
-
Also in tables 2 and 5 it seems that the corrected estimators worsen as sample size increases. What explains this pattern?
8 comments
Critical Weaknesses 1.2 and 1.3: Uniform Nuisance Control and Neural Bridge Specification
1.2. Assumptions C.6/C.9: Uniform nuisance control
**Bandwidth selection.**Fixed bandwidth grids may retain smoothing bias; cross-validation (CV) over a fixed grid alone cannot establish C.6 or C.9. Original Demand/Gaussian benchmarks retain fixed bandwidths for finite-sample comparison.
The nonlinear high-dimensional study (Appendix F.2) and MNAR sensitivity analysis (Appendix I) use five-fold CV over 30 random hyperparameter configurations (Appendix F, Table 10). For IMP/AIPW, candidate bandwidths depend on sample size and conditioning dimension : with .
For fixed , this grid shrinks with and meets the usual bandwidth/effective-sample-size scaling, but this alone does not establish uniform nuisance consistency.
C.6 uniformity. In NMMR, the Nadaraya-Watson (NW) residual smoother uses shared nonnegative weights summing to one. Bandwidth selection uses initial training residuals; the weights stay fixed during bridge optimization (Appendix C.1.3). For bridges in the common compact class , these weights imply
where vhat is the NW regression of the envelope v(W,A,X) with those weights. If NW is L2-consistent for fixed residual responses and fitted envelope regressions are uniformly L2-bounded in probability, finite covers give uniform consistency over K; cross-fitting transfers this to C.6's held-out empirical errors (Appendix C.4.1).
C.9's two requirements. First, under a fitted propensity floor , the AIPW product remainder is bounded by
It vanishes if either nuisance error vanishes and the other remains bounded.
Second, C.9 separately requires a centered uniform empirical-process condition. Appendix C.4.3 gives sufficient conditions through cross-fitting, a bounded kernel, suitable envelopes and finite covers; the oracle-class complexity condition is separate. These arguments do not claim that the original fixed numerical configurations satisfy all asymptotic assumptions.
Revisions for 1.2. Appendix E separates fixed finite-sample bandwidths from asymptotic guarantees; Appendix F/Table 10 lists data-driven candidates for F.2/I; Appendices C.4.1 and C.4.3 give sufficient uniform-control arguments.
1.3. Assumption C.4: Approximate neural bridge specification
We replaced exact bridge inclusion in the implemented network class with approximate specification for growing classes, motivated by neural approximation theory (e.g., Hornik, 1991; Schmidt-Hieber, 2020; Farrell et al., 2021). For a valid bridge and bounded MMR kernel,
Thus approximation can drive risk to zero even if is never exactly in . Revised C.4 asks for a deterministic comparison sequence in with
Both approximation and the penalty matter: increasing width alone is insufficient. Appendix C.3, Eq. (61), gives for each outer fold and correction :
where is the uniform corrected-criterion deviation and the last term includes approximate-minimization error. Together with uniform estimation/optimization, C.4 yields vanishing risk; C.5 and held-out evaluation conditions then imply causal-mean consistency.
Connection to implementation. The original Demand NMMR uses a fixed three-hidden-layer width-80 network and penalty across sample sizes; we retain it but do not claim it satisfies the asymptotic conditions.
Since its valid bridge is known, fitting the same architecture to on simulated full data and measuring held-out error could upper-bound its population moment risk, although we did not run this diagnostic.
For growing networks with parameters bounded coordinatewise by , suffices for the penalty contribution; approximation, uniform estimation and optimization still need control.
F.2 and I instead tune 30 randomly sampled configurations by five-fold CV, including architecture, learning rates and regularization; finite-candidate tuning also does not prove C.4.
Revisions for 1.3. Section 3.1/Appendix C.2 introduce growing classes and approximate specification; Theorem 1/Appendix C.3 incorporate approximation and penalty errors. Appendix E documents fixed configurations, while Appendix F/Table 10 reports the expanded tuning protocol.
Critical Weakness 1.4 & Question 3.2: Inverse stability and fixed-treatment evaluation
The original Assumptions D.6 and D.7(ii) are strong and are not guaranteed by a neural architecture or hyperparameter tuning. In particular, controlling the inverse-stability factor and the fixed-treatment transfer factor requires additional restrictions on the candidate bridge class.
Alternative local dual-bridge proof. We revised Theorem 2 to establish an alternative route to DFPV causal-mean consistency without D.6 or D.7(ii). The two-stage analysis bounds the projected error
Appendix D.4 separates neural approximation, nuisance estimation, empirical-risk deviations, regularization, and optimization errors.
Assumption D.8 instead uses a local dual treatment bridge satisfying, for almost every near the target ,
and . This directly controls causal-mean error through the projected bridge error rather than strong convergence of the entire outcome bridge.
Continuous treatment and pointwise consistency. Convergence integrated over the factual treatment distribution does not alone imply consistency at a fixed continuous treatment value. D.8 therefore also requires local treatment mass and local -Hölder continuity of the fitted causal-mean errors, with factor . Theorem D.1 (Eq. 106) gives
where and controls Stage 1 and evaluation errors. If , any vanishing projected error suffices for this local transfer step. More generally, for and , the sufficient rate condition is .
Appendix D.3.4 bounds the neural mean-curve Hölder factor using the coefficient norm, treatment-feature regularity, and proxy/covariate-feature moments; controlling treatment-network layer spectral norms is sufficient for the Lipschitz contribution.
Comparison of sufficient conditions. We retain D.6 and D.7(ii) as an alternative. For projected error , that route requires ; the local dual route instead requires the above Hölder-rate condition. The latter avoids these two inverse-stability/transfer factors but introduces substantive assumptions on dual-bridge existence, integrability, local treatment mass, and regularity. Neither route universally dominates, and neither is guaranteed by the implemented architectures.
Revisions. Section 4 (Theorem 2) states both routes; Appendix D.3.4 gives the dual-bridge, local mass, neural Hölder and rate conditions; Appendix D.4 proves the fixed-treatment bound and consistency.
Response to Critical Weakness 1.1
Critical Weakness 1.1: Assumption C.5 -- Target-mean Separation
We have revised the NMMR theory to accommodate growing neural bridge classes and clarified the identification, approximation, complexity, and nuisance recovery conditions required for consistency.
Assumption C.5 links vanishing population NMMR risk to causal-mean consistency. For the neural bridge class Hn={hθ:θ∈Θn}, define
ba(θ)=E[hθ(W,a,X)]
ca,n(ε)=inf{Rκ(θ):θ∈Θn,∣ba(θ)−β(a)∣≥ε}.
The revised condition requires liminfn→∞ca,n(ε)>0 for every ε>0.
Sufficient conditions. Appendix C.2 establishes three conditions for this uniform separation:
(i) Identification. Every zero-risk function in the common function-class closure K must identify the target causal mean: Rκ(h)=0 implies E[h(W,a,X)]=β(a).
This follows from proximal identification when zero-risk candidates satisfy the required outcome-bridge and integrability conditions. For continuous treatment, bridge validity at the specified treatment value and the pointwise conditions in Appendix A.1 are additionally needed. Bridge uniqueness is not required.
(ii) Continuity. Both the population risk Rκ(h) and target mean ba(h) must be continuous on K in the same topology. A weighted supremum norm provides one sufficient setting under appropriate moment conditions and a bounded NMMR kernel. For a fixed finite ReLU network with bounded parameters, a linear-growth envelope and finite input moments can provide the required continuity.
(iii) Common compactness. The closure K of the union of the neural classes Hn must be compact in the chosen function topology. For fixed architectures, bounded parameter sets and continuity of the network imply compactness of the corresponding function class. For growing classes, however, compactness of each parameter space Θn alone is insufficient; uniform control across widths is required. On compact input domains, uniform boundedness and equicontinuity provide sufficient conditions through the Arzelà–Ascoli theorem.
Under these three properties, the set of functions in K whose target means differ from β(a) by at least ε is compact and contains no zero-risk function. By continuity, the population risk attains a strictly positive minimum on this set whenever it is nonempty. This minimum uniformly bounds ca,n(ε) away from zero, establishing C.5.
Connection to implementation. Prescribed parameter-box constraints can be enforced by projection after each optimizer update (Appendix C.1.3 and Algorithm 1). Such constraints make each finite-dimensional parameter space compact but do not themselves guarantee common compactness for growing-width networks.
The reported benchmarks use finite, fixed-width architectures. We do not claim that their numerical configurations verify all the above sufficient conditions. Instead, the revised theory characterizes conditions under which neural classes satisfy C.5, while the experiments assess finite-sample performance. Approximation requirements for growing classes are addressed separately under Assumption C.4.
Revisions. Section 3.1 introduces sample-size-dependent neural bridge classes; Section 4 states the revised target-mean separation condition. Appendix C.2 establishes sufficient conditions based on common compactness, continuity, and identification of zero-risk bridges. Appendix C.1.3 and Algorithm 1 describe projection onto prescribed parameter constraints and distinguish these constraints from the approximate-minimization requirement.