Pith. sign in

REVIEW 5 major objections 6 minor 32 references

TV-SurvCaus: Dynamic Representation Balancing for Causal Survival Analysis

T0 review · 5 major / 6 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 TV-SurvCaus claims that balancing sequence-level representations across treatment histories makes counterfactual survival functions consistently estimable, with TV-PEHE converging to zero under stabilized weighting and an MMD regularizer.

desk verdict A plausible architecture with honest limitations, but no actual results: the experiments are explicit placeholders and the main consistency theorem assumes what it claims to prove. read the letter →

arxiv 2505.01785 v1 pith:WWI32CHL submitted 2025-05-03 stat.ML cs.LG

Ayoub Abraich This is my paper · ORCID
classification stat.MLcs.LG MSC 62D2062N0162N0268T07
keywords time-varyingtreatmentscausalsurvivalanalysisrepresentationbalancingtreatment-confounderfeedbackpotentialoutcomesmaximummeandiscrepancyTV-PEHEmarginalstructuralmodels
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

TV-SurvCaus sets out to show that representation balancing—learning a latent history representation that looks similar across different treatment sequences—extends to causal survival analysis with time-varying treatments and censored outcomes. The paper argues that estimating counterfactual survival curves under dynamic treatment regimes is feasible and provably consistent when a survival loss weighted by stabilized inverse-probability weights is combined with an MMD balancing penalty. It backs this with a generalization bound on the time-varying precision in estimating heterogeneous effects (TV-PEHE), a consistency theorem, convergence-rate discussion, and a bias decomposition. If correct, the framework would give medical and policy analysts a principled way to estimate individual treatment effects of evolving treatment protocols from longitudinal observational data, including real clinical records.

What carries the argument

The load-bearing object is the balanced representation $z = \phi(\psi(X,T))$, built by a sequence encoder (an LSTM or GRU) followed by a representation network, with maximum mean discrepancy (MMD) as the integral probability metric measuring distributional balance between treatment sequences. The training objective combines a discrete-time survival negative log-likelihood weighted by stabilized sequential inverse-probability weights, the MMD balancing penalty $\alpha L_{bal}$, and an $L_2$ regularization term. The main theoretical tool is a domain-adaptation-style bound on counterfactual risk; the bound identifies representation imbalance as the term the balancing loss must drive to zero, and the consistency theorem rests on that driving condition.

What would settle it

On synthetic data with a known true survival function, induce strong time-varying confounding and misspecify the propensity model used to build the stabilized weights; if the MMD between learned representations stays bounded away from zero and TV-PEHE does not shrink with increasing sample size, the consistency claim is refuted.

Watch

Extended reading notes

Core claim

The central discovery is that the counterfactual survival risk under an alternative treatment sequence is bounded by the factual risk plus an integral-probability-metric discrepancy between representation distributions; minimizing that discrepancy and using bounded stabilized inverse-probability weights makes the TV-PEHE converge in probability to zero (Theorem 4.6). In the paper's own terms, the estimated potential survival function converges to the true function for sequences represented in the data, and experiments on synthetic, semi-synthetic, and real clinical data show lower TV-PEHE and better discrimination and calibration than marginal structural models, G-formula, and deep sequence baselines.

Load-bearing premise

The consistency proof requires that the balancing penalty can actually drive the distributional discrepancy between treatment groups to zero as the sample grows, and that the stabilized inverse-probability weights are correctly specified or consistently estimated and bounded; if either condition fails, the claimed convergence of the estimator and of TV-PEHE to zero does not follow.

Editorial extensions

If this is right

  • Counterfactual survival curves for arbitrary treatment sequences can be estimated from observational longitudinal data without fitting a separate outcome model per regime, since one balanced representation plus a sequence-conditioned prediction head serves all sequences.
  • In nonlinear data-generating processes and longer treatment horizons, the combination of stabilized weighting and representation balancing gives progressively larger reductions in TV-PEHE relative to marginal structural models and G-formula.
  • The balancing regularizer absorbs part of the bias from treatment-confounder feedback, so the estimator can remain competitive when the propensity model used for weighting is mildly misspecified.
  • Consistent TV-PEHE estimates make downstream dynamic treatment regime optimization possible, because the learned counterfactual survival models supply the per-sequence outcomes such methods require.

Reading between the lines

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

  • In finite samples the practical value probably hinges on the balancing strength $\alpha$; a testable prediction is that as $\alpha$ shrinks to zero, TV-SurvCaus degrades toward a plain sequence survival model, and monitoring validation MMD should track residual confounding bias.
  • The same bounding structure suggests extensions to continuous treatment doses and competing-risk endpoints, provided the IPM and survival head are redefined; the representation-balancing argument itself would carry over.
  • The theory's reliance on asymptotic balance implies that overlap violations are the real threat in applications: when some histories make treatment nearly deterministic, neither weighting nor representation balancing can recover the counterfactual, and reported causal effects in such strata should be downweighted or omitted.
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

5 major / 6 minor

Summary. The paper introduces TV-SurvCaus, a neural architecture for estimating the causal effect of time-varying treatments on survival outcomes. The method combines an RNN sequence encoder, a representation network trained with an MMD-based balancing penalty, and a discrete-time survival prediction head, with stabilized inverse probability weights in the loss. The theoretical sections claim a generalization bound connecting TV-PEHE to representation imbalance, variance analysis for sequential stabilized weights, consistency of the estimator, convergence rates under temporal dependence, and a bias decomposition for treatment-confounder feedback. The empirical section reports results on synthetic, semi-synthetic (SUPPORT, TCGA, METABRIC), and MIMIC-III data, claiming consistent improvements over Cox-MSM, G-formula, DeepSurv-L, RSF-MSM, and RMSN.

Significance. If the claims were established, the paper would offer a useful extension of representation balancing to dynamic treatment regimes with censored survival data, a setting that is indeed under-explored. The conceptual framing is relevant and the proposed architecture is plausible. However, the current manuscript provides no verifiable theoretical proof: the central consistency theorem is conditioned on essentially the conclusion, the convergence rate is an unspecified decomposition, and the bias theorem is explicitly labeled a conceptual decomposition. The empirical pillar is also missing: Section 7 opens with the placeholder text '(Assuming tables contain actual results now)', the synthetic data generator is never described, and no training details, hyperparameters, code, or data-processing steps are given. Consequently, the claimed significance cannot be assessed beyond the proposal level, and the paper in its present form does not support its central assertions.

major comments (5)
  1. [Section 4.3, Theorem 4.6, Condition (iv)] The consistency claim TV-PEHE → 0 is conditioned on the assumption that the balancing penalty drives the IPM discrepancy d_H(P^φ_a, P^φ_a') to zero as n → ∞. This is essentially the conclusion the theorem purports to prove. The proof sketch does not show that the minimizer of the empirical objective in Eq. (10) achieves this; it merely cites Condition (iv) in the combining step. No argument is given for how α must scale, what model capacity suffices, or how optimization error is overcome. As stated, the theorem is circular and does not establish consistency of the TV-SurvCaus estimator.
  2. [Section 7 (opening sentence and Tables 1-11)] The empirical section is not a real evaluation: the text explicitly says '(Assuming tables contain actual results now)'. The tables appear to be placeholders and cannot be independently checked. In addition, the synthetic data generator is described as 'described previously' but is never actually described anywhere in the manuscript, and no details are provided for the semi-synthetic constructions, the MIMIC-III cohort, hyperparameter selection, or experimental protocol. The central empirical claim that TV-SurvCaus outperforms baselines is therefore unsubstantiated in this submission.
  3. [Section 4.4, Theorem 4.7] The claimed convergence rate is not derived. Equation (4) is a generic error decomposition into statistical error, approximation error, and balancing error, where the statistical rate R_stat(n, F, β) is left unspecified. The proof sketch does not prove any of the three displayed terms, and no explicit rate is given for the time-varying survival setting. Since the paper lists 'convergence rates for representation learning with temporal dependencies' as a contribution, this is a load-bearing gap rather than a minor omission.
  4. [Section 4.5, Theorem 4.8] The proof sketch is described as 'primarily a conceptual decomposition'. This is not a theorem with a proof; it is a qualitative list of bias sources. The statement does not define the bias components formally, does not state assumptions under which the decomposition holds, and provides no bound on the total bias. As such, Theorem 4.8 does not provide the 'refined analysis of bias' promised in the introduction.
  5. [Section 4.1, Theorem 4.1 and Corollary 4.2] The main theoretical motivation for the balancing objective is not established. Theorem 4.1 is stated as an adaptation of existing domain-adaptation bounds, but no proof is given that the IPM discrepancy between representation distributions across treatment sequences controls counterfactual survival risk under the weighted discrete-time survival loss with censoring. Corollary 4.2 handwaves the step 'relating risk to MSE (e.g., via properties of the survival loss)', which is precisely the step that would connect the abstract bound to the TV-PEHE metric. Without this step, the bound does not directly motivate the objective in Eq. (10).
minor comments (6)
  1. [Section 3.1] The notation T_i(−1) = empty is used in the setup, but Equation (2) uses T_i(k−1) for k = 0; please make the convention for the empty history explicit at the point where the weights are defined.
  2. [Section 3.3, Definition 3.4] The notation 'TV-CATES(X, a, a′, τ; ·) − TV-CATES(X, a, a′, τ; ·∗)' is confusing because the first argument uses the estimator and the second uses the true function; clarify by writing F̂ and F* explicitly.
  3. [Section 4.2, Theorem 4.3] Theorem 4.3 is a definition of stabilized weights, not a theorem; it should be labeled as a definition or equation rather than a numbered theorem.
  4. [Section 6.1] The synthetic data section says the data is 'Generated as described previously', but no previous description exists in the paper; the full generator including the treatment assignment mechanism, covariate dynamics, outcome generation, censoring, and feedback strength β must be specified.
  5. [Section 7.3, Table 8] The table reports '28-day mortality risk reduction', but the method section does not explain how this estimand is computed from the survival curves; state the formula used.
  6. [Section 8.3] The limitations section appropriately acknowledges assumptions and computational cost, but the paper should also state explicitly that no code or data are released, which is relevant for reproducibility.

Circularity Check

1 steps flagged · score 6.0 of 10

Theorem 4.6's consistency claim is conditional on Condition (iv) already asserting the balancing-driven dH→0 that the proof uses to conclude TV-PEHE→0; the empirical section is explicitly a placeholder.

  1. self definitional [Section 4.3, Theorem 4.6, Conditions (iii)-(iv), Eq. (3); training objective Eq. (10) in Section 5.5]
    "Assume: ... (iii) Stabilized weights wstab(K) are correctly specified (or consistently estimated) and bounded. (iv) The balancing penalty αLbal successfully drives the chosen IPM discrepancy dH towards zero as n→∞. ... Then, under Assumptions 3.1-3.5, the estimated potential survival function converges in probability to the true function ... Consequently, TV-PEHE ... p→ 0 (3). Proof Sketch. ... 2. Counterfactual Generalization via Balancing: The balancing term αLbal (Equation (7)) penalizes discrepancies dH(Pφa,Pφa′) ..."

    The only mechanism offered for counterfactual generalization is Condition (iv), which is exactly the assertion that the learned representation balances the treatment-sequence distributions to zero IPM discrepancy. The proof then says the objective 'forces this discrepancy towards zero (Condition iv)' and invokes Theorem 4.1, whose bound contains dH(Pφa,Pφa′) as the discrepancy term. No argument is given that minimizing Eq. (10) with finite α, finite model class, and optimization error actually achieves dH→0; Condition (iii) similarly assumes away weight misspecification. Thus the stated consistency reduces to assuming the convergence it claims to establish: if Condition (iv) fails, the TV-PEHE bound in Eq. (4) retains a nonzero balancing error O(dH) and no consistency follows.

full rationale

The theoretical centerpiece Theorem 4.6 is conditional on Conditions (iii) and (iv) that already contain the desired conclusion: Condition (iv) assumes the balancing penalty drives the IPM discrepancy dH to zero, and the proof sketch then cites that condition as the reason the counterfactual risk is controlled and TV-PEHE→0. This is a self-definitional circularity because the convergence of the estimator is made an assumption of the theorem rather than derived. The paper's own Section 7 admits the experimental tables are placeholders ('(Assuming tables contain actual results now)'), and Section 4.8 labels the bias decomposition a 'conceptual decomposition.' These internal acknowledgements strengthen the finding that neither the theoretical nor empirical pillar provides independent evidence. The generalization bound Theorem 4.1 is a standard adaptation of Ben-David et al. and Johansson et al., so it is not itself circular with respect to the observed data; the circularity is concentrated in Theorem 4.6's consistency claim. The empirical tables, even if filled, would not rescue the theoretical circularity because the experiments are not used to verify Condition (iv). Hence the central claim of guaranteed convergence reduces by construction to an assumption of that convergence, warranting a score of 6.

Assumptions & free parameters 4 free parameters · 9 assumptions · 0 invented entities

The causal identification rests entirely on the five standard assumptions (3.1 to 3.5), which the authors acknowledge are strong and untestable. The theoretical claims add no new axioms beyond standard statistical learning results, but Theorem 4.6 introduces an ad hoc condition that the balancing loss drives the discrepancy to zero, which essentially assumes the conclusion. The method also relies on estimated propensity and stabilization models whose fitted parameters are nuisance quantities.

free parameters (4)
  • Balancing penalty alpha = Tuned on validation (no fixed value)
    Chosen by hand; controls the trade-off between the survival loss and the MMD balancing loss in Equation (10). It is central to the method's claimed improvements, but no principled selection rule is given.
  • MMD kernel bandwidth sigma = Tuned per experiment
    Hyperparameter of the Gaussian RBF kernel in Equation (7); the authors note performance depends on IPM parameters (Section 8.3).
  • Number of discrete time intervals m = Chosen by hand
    Discretization granularity for the survival prediction head in Equation (8); not specified for any experiment, a modeling choice that affects results.
  • Propensity score and stabilization model parameters = Estimated from data
    Parameters of models for e_k and p_k used to build stabilized weights in Equation (2); misspecification of these nuisance models biases results (Theorem 4.8).
assumptions (9)
  • domain assumption Sequential exchangeability (Assumption 3.1): {Y(a), C(a)} is independent of T(k) given X(k) and T(k-1)
    Central identifying assumption requiring no unmeasured time-varying confounders; the authors note it is strong and untestable (Remark 3.1).
  • domain assumption Positivity (Assumption 3.2): 0 < e_k(t|history) < 1 for all histories
    Required for IPTW weights and for meaningful overlap in representation balancing.
  • domain assumption Consistency (Assumption 3.3): observed outcomes equal potential outcomes under the observed treatment sequence
    Standard connection between observed data and potential outcomes.
  • domain assumption Conditional non-informative censoring (Assumption 3.4): Y(a) independent of C(a) given full history
    Censoring provides no additional information about survival beyond the history.
  • domain assumption No anticipation (Assumption 3.5): potential outcomes do not depend on future treatment components
    Standard assumption tying potential outcomes to the sequence up to the event time.
  • ad hoc to paper Condition (iv) of Theorem 4.6: the balancing penalty drives the IPM discrepancy to zero as n tends to infinity
    The consistency proof assumes the balancing regularizer achieves perfect asymptotic balance, which is essentially the convergence property being proved.
  • domain assumption Beta-mixing temporal process with rate O(m^-b), b > 1 (Theorem 4.7)
    Assumed to derive statistical rates under temporal dependence; no verification is provided for the real data.
  • standard math Standard domain adaptation bound (Ben-David et al. 2010) invoked in Theorem 4.1
    Restated rather than proved in this paper; a known result from the prior literature.
  • standard math M-estimator regularity conditions (Van der Vaart 2000) invoked in Theorem 4.6
    Standard conditions for consistency of the weighted survival loss as an M-estimator.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TV-SurvCaus: Dynamic Representation Balancing for Causal Survival Analysis." pith.science (2026). https://pith.science/paper/WWI32CHL

@misc{pith2026250501785,
  author       = {Pith},
  title        = {Pith review of: TV-SurvCaus: Dynamic Representation Balancing for Causal Survival Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WWI32CHL}},
  note         = {Machine review of arXiv:2505.01785}
}
read the original abstract

Estimating the causal effect of time-varying treatments on survival outcomes is a challenging task in many domains, particularly in medicine where treatment protocols adapt over time. While recent advances in representation learning have improved causal inference for static treatments, extending these methods to dynamic treatment regimes with survival outcomes remains under-explored. In this paper, we introduce TV-SurvCaus, a novel framework that extends representation balancing techniques to the time-varying treatment setting for survival analysis. We provide theoretical guarantees through (1) a generalized bound for time-varying precision in estimation of heterogeneous effects, (2) variance control via sequential balancing weights, (3) consistency results for dynamic treatment regimes, (4) convergence rates for representation learning with temporal dependencies, and (5) a formal bound on the bias due to treatment-confounder feedback. Our neural architecture incorporates sequence modeling to handle temporal dependencies while balancing time-dependent representations. Through extensive experiments on both synthetic and real-world datasets, we demonstrate that TV-SurvCaus outperforms existing methods in estimating individualized treatment effects with time-varying covariates and treatments. Our framework advances the field of causal inference by enabling more accurate estimation of treatment effects in dynamic, longitudinal settings with survival outcomes.

Figures

Figures reproduced from arXiv: 2505.01785 by the authors.

Figure 1
Figure 1. TV-SurvCaus Architecture: A sequence encoder [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 27 canonical work pages

  1. [1]

    P. C. Austin. The use of propensity score methods with survival or time-to-event outcomes: reporting measures of effect similar to those used in randomized experiments. Statistics in Medicine, 33 0 (7): 0 1242--1258, 2014

    work page 2014

  2. [2]

    Ben-David, J

    S. Ben-David, J. Blitzer, K. Crammer, A. Kulesza, F. Pereira, and J. W. Vaughan. A theory of learning from different domains. Machine Learning, 79: 0 151--175, 2010

    work page 2010

  3. [3]

    Chakraborty and E

    B. Chakraborty and E. E. Moodie. Dynamic treatment regimes. Annual Review of Statistics and Its Application, 1: 0 447--464, 2014

    work page 2014

  4. [4]

    Chapfuwa, P

    P. Chapfuwa, P. Zeng, L. Carin, and M. Udell. Survival causal transport: Estimating causal effects under non-independent censorings. In H. Daum \'e III and A. Singh, editors, Proceedings of the 37th International Conference on Machine Learning, pages 1340--1349. PMLR, 2020

    work page 2020

  5. [5]

    K. Cho, B. Van Merri \"e nboer, C. Gulcehre, D. Bahdanau, F. Bougares, H. Schwenk, and Y. Bengio. Learning phrase representations using rnn encoder-decoder for statistical machine translation. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1724--1734, 2014

    work page 2014

  6. [6]

    D. R. Cox. Regression models and life-tables. Journal of the Royal Statistical Society: Series B (Methodological), 34 0 (2): 0 187--202, 1972

    work page 1972

  7. [7]

    M. F. Gensheimer and B. Narasimhan. Nnet-survival : a deep learning model for survival analysis. PeerJ, 7: 0 e7134, 2019

    work page 2019

  8. [8]

    Hatt and S

    T. Hatt and S. Feuerriegel. Estimating the effects of continuous-valued interventions using generalized propensity score. In M. Ranzato, A. Beygelzimer, Y. Dauphin, P. Liang, and J. W. Vaughan, editors, Advances in Neural Information Processing Systems, volume 34, pages 16861--16873, 2021

    work page 2021

Show all 32 references
  1. [9]

    Hochreiter and J

    S. Hochreiter and J. Schmidhuber. Long short-term memory. Neural computation, 9 0 (8): 0 1735--1780, 1997

    1997

  2. [10]

    Ishwaran, U

    H. Ishwaran, U. B. Kogalur, E. H. Blackstone, and M. S. Lauer. Random survival forests. The Annals of Applied Statistics, 2 0 (3): 0 841--860, 2008

    2008

  3. [11]

    A. E. Johnson, T. J. Pollard, L. Shen, L.-w. H. Lehman, M. Feng, M. Ghassemi, B. Moody, P. Szolovits, L. A. Celi, and R. G. Mark. MIMIC-III , a freely accessible critical care database. Scientific Data, 3 0 (1): 0 160035, 2016

    2016

  4. [12]

    F. D. Johansson, U. Shalit, and D. Sontag. Generalization bounds and representation learning for estimation of potential outcomes and causal effects. Journal of Machine Learning Research, 21 0 (119): 0 1--60, 2020

    2020

  5. [13]

    J. L. Katzman, U. Shaham, A. Cloninger, J. Bates, T. Jiang, and Y. Kluger. DeepSurv : personalized treatment recommender system using a Cox proportional hazards deep neural network. BMC Medical Research Methodology, 18 0 (1): 0 24, 2018

    2018

  6. [14]

    D. P. Kingma and J. Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014

    2014 arXiv

  7. [15]

    Kontorovich and K

    A. Kontorovich and K. Ramanan. Concentration inequalities for dependent variables via martingale decomposition. In 2008 IEEE Information Theory Workshop, pages 475--479. IEEE, 2008

    2008

  8. [16]

    Kreif, L

    N. Kreif, L. Tran, R. Grieve, B. De Stavola, R. C. Tasker, and M. Petersen. Estimating the comparative effectiveness of feeding interventions in the pediatric intensive care unit: a demonstration of longitudinal targeted maximum likelihood estimation. American Journal of Epide...

    2017

  9. [17]

    C. Lee, W. R. Zame, J. Yoon, and M. Van der Schaar. Deephit : A deep learning approach to survival analysis with competing risks. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 32, pages 2788--2795, 2018

    2018

  10. [18]

    F. Li, K. L. Morgan, and A. M. Zaslavsky. Balancing propensity score for causal inference effects: A tutorial and survey. In Statistical Methods in Medical Research, volume 27, pages 2355--2371. Sage Publications Sage UK: London, England, 2018

    2018

  11. [19]

    B. Lim, A. M. Alaa, and M. Van der Schaar. Forecasting treatment responses over time using recurrent marginal structural networks. In S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett, editors, Advances in Neural Information Processing Systems, ...

    2018

  12. [20]

    Mohri, A

    M. Mohri, A. Rostamizadeh, and A. Talwalkar. Foundations of machine learning. MIT Press, Second edition, 2018

    2018

  13. [21]

    Nagpal, X

    C. Nagpal, X. Li, and A. Dubrawski. Deep survival machines: Fully parametric survival regression and representation learning for censored data with competing risks. IEEE Journal of Biomedical and Health Informatics, 25: 0 3163--3175, 2021

    2021

  14. [22]

    J. Robins. A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect. Mathematical Modelling, 7 0 (9-12): 0 1393--1512, 1986

    1986

  15. [23]

    J. M. Robins. Estimation of the causal effect of a time-varying exposure on the marginal mean of a repeated binary outcome. Journal of the American Statistical Association, 87 0 (419): 0 1013--1025, 1992

    1992

  16. [24]

    J. M. Robins, M. \'A . Hern \'a n, and B. Brumback. Marginal structural models and causal inference in epidemiology. Epidemiology, 11 0 (5): 0 550--560, 2000

    2000

  17. [25]

    D. B. Rubin. Estimating causal effects of treatments in randomized and nonrandomized studies. Journal of Educational Psychology, 66 0 (5): 0 688--701, 1974

    1974

  18. [26]

    Schulam and S

    P. Schulam and S. Saria. Reliable decision support using counterfactual models. In I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, editors, Advances in Neural Information Processing Systems, volume 30, pages 1697--1708, 2017

    2017

  19. [27]

    Schuler, G

    A. Schuler, G. Zerbini, and R. Ranganath. Estimating causal effects under data scarcity with deep generate-and-check. arXiv preprint arXiv:2306.04557, 2023

    2023 arXiv

  20. [28]

    Shalit, F

    U. Shalit, F. D. Johansson, and D. Sontag. Estimating individual treatment effect: generalization bounds and algorithms. In D. Precup and Y. W. Teh, editors, Proceedings of the 34th International Conference on Machine Learning, volume 70, pages 3076--3085. PMLR, 2017

    2017

  21. [29]

    A. W. Van der Vaart. Asymptotic statistics, volume 3. Cambridge university press, 2000

    2000

  22. [30]

    J. Yoon, J. Jordon, and M. van der Schaar. Ganite: Estimation of individualized treatment effects using generative adversarial nets. In International Conference on Learning Representations, 2018

    2018

  23. [31]

    B. Yu. Rates of convergence for empirical processes of mixing sequences. The Annals of Probability, 22 0 (1): 0 94--116, 1994

    1994

  24. [32]

    Abraich, A

    A. Abraich, A. Guilloux, and B. Hanczar. Survcaus : Representation balancing for survival causal inference. arXiv preprint arXiv:2203.15672, 2022. URL https://arxiv.org/abs/2203.15672

    2022 arXiv

Pith tools

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