Pith. sign in

REVIEW 5 major objections 7 minor 25 references

Multi-Treatment-DML: Causal Estimation for Multi-Dimensional Continuous Treatments with Monotonicity Constraints in Personal Loan Risk Optimization

T0 review · 5 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Multi-Treatment-DML debiases continuous multi-dimensional treatments and makes loan risk monotonic in each component, with a 10% profit gain online.

desk verdict Real problem, plausible pipeline, but the monotonicity guarantee is false as written and the DML debiasing is unproven; worth a major-revision review, not acceptance. read the letter →

arxiv 2508.02183 v2 pith:S6QSE4QY submitted 2025-08-04 cs.LG

Kexin Zhao , Bo Wang , Cuiying Zhao , Tongyao Wan This is my paper
classification cs.LG
keywords multi-dimensionalcontinuoustreatmentsdoublemachinelearningmonotonicityconstraintsupliftmodelingobservationalcausalinferencepersonalloanrisklatentdisentanglementtreatmenteffectestimation
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

Personal loan platforms set credit limits, interest rates, and loan terms from observational data, yet causal effect estimates for such continuous, multi-dimensional treatments are rare and often biased by borrower creditworthiness. The paper proposes Multi-Treatment-DML, a Double Machine Learning framework that residualizes multiple continuous treatments against confounders and adds a linear monotonic module linking treatment changes to predicted outcome changes. The claim is that this design debiases treatment-effect estimates for arbitrary-dimensional continuous treatments while guaranteeing that raising any treatment component never lowers predicted risk, satisfying financial domain constraints. The authors report that the method outperforms existing baselines on public benchmarks and on a large proprietary lending dataset, and that an online deployment improved cumulative profit by 10%.

What carries the argument

The load-bearing object is the linear monotonic model $\hat{Y}_{\mathrm{final}} = \hat{Y} + f_{\mathrm{linear}}(K, \Delta T)$, where $K$ is a sensitivity coefficient predicted per user and $\Delta T$ is the change in the multi-dimensional continuous treatment vector. Because the function is linear in $\Delta T$, the direction of the predicted outcome change is determined by the sign of $K$; the paper asserts this enforces monotonicity by design. This module sits on top of a propensity network that disentangles features into instrumental, confounder, and adjustment embeddings and predicts $\hat{T} = E[T \mid X]$ and $\hat{Y} = E[Y \mid X]$, with a Tweedie loss for zero-inflated long-tailed financial outcomes. A two-fold cross-fitting procedure swaps the roles of the networks to reduce overfitting bias, mirroring the DML residualization principle.

What would settle it

Take the trained model and any held-out user; increase exactly one treatment component T_k while holding all others fixed and record the predicted outcome. If the predicted outcome ever decreases, the monotonicity claim 'by design' is false. A cheaper check: scan the learned K values for any negative entry, since the linear model makes a negative K produce exactly that decrease.

Watch

Extended reading notes

Core claim

On its own terms, the paper claims that the combination of a disentangling propensity network and a sensitivity-aware causal network extends DML to settings with K-dimensional continuous treatments and enforces a monotonic dose-response relationship by construction. The propensity network predicts the treatment vector, the outcome, and a per-user sensitivity coefficient K from features split into instrumental, confounder, and adjustment components. The causal network then computes the final prediction as the baseline outcome plus a linear function of treatment changes scaled by K, so any increase in a treatment component increases the predicted outcome if K is non-negative. The paper treats this as a rigorous guarantee rather than an empirical tendency, and supports it with experiments on Twins, Jobs, Lazada, and a real personal-loan dataset, plus an online A/B test reporting a 10% cumulative profit gain.

Load-bearing premise

The whole debiasing and monotonicity story rests on two unproven premises at once: that the learned I/C/A disentanglement and residualization actually make treatment residuals independent of confounders, and that the predicted sensitivity coefficient K is non-negative for every user.

Editorial extensions

If this is right

  • Credit platforms can estimate counterfactual risk for candidate offers varying in limit, rate, and term simultaneously from observational data, without randomized trials.
  • Monotonicity becomes a structural property of the model, so business teams can rely on the qualitative rule that larger limits or longer terms never appear to reduce risk.
  • The method yields per-user sensitivity coefficients K that could be used to segment customers by risk responsiveness, as the paper's high-credit versus medium-credit analysis illustrates.
  • The framework extends DML to arbitrary treatment dimension K, opening the approach to other continuous multi-treatment domains such as pricing and healthcare dosing.
  • Online deployment results suggest the uplift modeling transfers to production LTV optimization, with cumulative profit gains persisting over longer horizons.

Reading between the lines

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

  • A minimal testable extension the paper does not run is a synthetic-data experiment with a known monotonic response and known confounding; this would directly verify whether the residualization removes bias and whether monotonicity holds out-of-sample.
  • Because the paper never constrains K to be non-negative, the monotonicity guarantee is actually contingent; adding a non-negativity constraint such as a softplus output would make the stated guarantee unconditional.
  • The linear model $\hat{Y}_{\mathrm{final}} = \hat{Y} + K \cdot \Delta T$ assumes additive, component-wise effects with no interactions between treatment components; the framework may understate risk when limit, rate, and term interact.
  • The disentanglement into I, C, A is encouraged by a cosine-similarity penalty on weights, but the paper gives no criterion for verifying conditional independence; an adopter could check this with a residual-independence test on held-out data.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 7 minor

Summary. The manuscript proposes Multi-Treatment-DML, a framework for estimating individual treatment effects of multi-dimensional continuous treatments from observational data, with applications to personal-loan credit decisions. The method combines a propensity network that predicts treatment, outcome, and a per-user sensitivity coefficient K from features, with a causal network whose final prediction is a linear function of the treatment change plus the predicted outcome. The paper claims three contributions: an extension of Double Machine Learning (DML) to multiple continuous treatments, a 'by design' monotonicity guarantee between treatments and outcomes, and validation on public benchmarks, a proprietary lending dataset, and an online deployment.

Significance. Personal-loan credit optimization is a practically important domain, and the paper addresses a genuine gap: most causal effect estimators focus on binary treatments, while credit decisions involve continuous multi-dimensional offers. The authors deserve credit for attempting a real-world deployment and for using a Tweedie loss suited to zero-inflated, long-tailed financial outcomes. However, the central theoretical claims are not established in the current manuscript: the monotonicity guarantee is not enforced by the described architecture, the DML orthogonalization is not derived or even fully specified, and the experiments do not validate the claimed continuous multi-dimensional treatment setting. As it stands, the paper substantially overstates its contributions, though the identified problems are in principle addressable in a major revision.

major comments (5)
  1. [Section 4.2, Eq. (5)] The claimed monotonicity guarantee is false as stated. The final prediction is written as \hat{Y}_final = f_linear(K, \Delta T) + \hat{Y}; for a linear parameterization, the partial derivative with respect to a treatment component T_k equals K_k. The only penalty on K is the L2 regularization in Eq. (5), which does not prevent negative values. No non-negativity constraint, clamping, or reparameterization (e.g., softplus) is described anywhere in Section 4.2 or in the overall loss in Eq. (7). A trained model can therefore produce K_k < 0, in which case increasing T_k decreases the predicted outcome, directly contradicting the 'rigorously enforcing monotonicity by design' claim. The paper should either constrain K to be non-negative and prove that the derivative is non-negative, or withdraw the guarantee.
  2. [Sections 4.2 and 4.4] The DML extension is not actually specified. The paper never defines the residuals (T - E[T|X] and Y - E[Y|X]) or states the moment condition used to estimate K; the phrase 'sequential residualization' is not backed by any equations. Section 4.4 on cross-fitting is garbled and non-reproducible: it begins by describing K-fold cross-fitting, then switches to 'the data is split into two non-overlapping parts,' with no algorithm for how the propensity and causal networks are trained across folds. Without a precise statement of the residualization and cross-fitting procedure, the claimed debiasing/orthogonalization property cannot be assessed.
  3. [Section 4.1, Eq. (4)] The RLO disentanglement does not establish the causal identification needed for the method. Minimizing cosine similarity between weight vectors does not imply that the learned I, C, A satisfy the conditional independence or exclusion restrictions required for instrumental-variable/confounder/adjustment decompositions. Moreover, the method is internally inconsistent with the instrumental-variable interpretation: although I is meant to affect only treatment, the sensitivity coefficient K is predicted from Concat(I,C,A) and enters the final outcome prediction, so I affects Y through K. The paper needs a formal statement of the assumptions under which the learned decomposition removes confounding, and a proof or a clear argument for why the architecture satisfies them.
  4. [Sections 5.1-5.4] The experimental evaluation does not test the claimed contribution. The public benchmarks (Twins, Jobs, Lazada) all involve binary treatments, and Table 4 further binarizes the credit-limit change in the personal-loan data. No experiment compares methods on multi-dimensional continuous treatments with known ground-truth effects. The real-world dataset has no ground truth, and Figure 4 only plots slopes for two groups without confidence intervals or any causal validation. Consequently, the central claim of handling 'arbitrary-dimensional continuous treatments' is unsupported by the evidence presented.
  5. [Section 4.2, Eq. (6)] The quantities \Delta T and \Delta Y are never formally defined. The text says 'we model the change in predicted outcome \Delta Y as a ... function of the change in treatment \Delta T,' and Eq. (6) writes a loss between Y_i and \hat{Y}_final,i, but there is no equation or sentence defining \Delta T for a given observation (e.g., \Delta T = T - \hat{T} or \Delta T = T - T_ref) or how \Delta Y relates to \hat{Y}_final - \hat{Y}. Without these definitions the model is not reproducible, and the claimed reduction to a DML residual regression cannot be verified.
minor comments (7)
  1. [Section 3] Assumptions 2 and 3 are written for binary treatments (using (Y_0, Y_1) and P(T=t|X=x)); in a continuous-treatment setting, positivity should be stated in terms of a density bounded away from zero, and the potential outcomes should be indexed by the continuous treatment value.
  2. [Related Work] The description of DRNet is duplicated verbatim in two consecutive sentences; one copy should be removed.
  3. [Section 5.2] There is a typo: 'expexted Precision in Estimation of Heterogeneous Effect' should be 'expected Precision in Estimation of Heterogeneous Effect'.
  4. [Abstract and Section 3] The outcome variable is described inconsistently: Section 3 defines Y as 'utilized loan amount,' while the abstract and introduction refer to default risk. Please clarify the target outcome(s) and how the Tweedie loss is applied to each.
  5. [Tables 3 and 5] Ablation results in Table 3 are reported as single numbers without standard errors, and Table 5 reports only '+10%' with no experiment duration, sample size, or significance test; please add these details or soften the claims.
  6. [Section 4.4] The sentence 'Next, select the next sample to estimate residuals, and repeat the above steps until all n samples are processed' is incomplete and should be removed or rewritten as a precise algorithmic description.
  7. [Section 5.2] The PCOC metric is defined as 'Predict Click Over Click' and described for click-through rates, but it is applied to loan risk/utilization outcomes; please clarify the outcome being calibrated and why a CTR calibration metric is appropriate.

Circularity Check

2 steps flagged · score 6.0 of 10

Monotonicity 'guarantee' is definitional and the treatment response is a fitted slope, so the central claims reduce to assumptions rather than derived results.

  1. self definitional [Section 4.2, Causal Network; Eq. (5)-(7)]
    "we model the change in predicted outcome ∆Y as a monotonic function of the change in treatment ∆T, parameterized by a sensitivity coefficient K that is dynamically predicted by the Propensity Network. This formulation ensures that any increase in a treatment component Tk cannot result in a decrease in the predicted outcome Y , thereby rigorously enforcing monotonicity by design."

    The monotonicity guarantee is not derived from a constraint; it is just the defining property of f_linear, which the paper calls 'a monotonic function.' The only regularization on K is L2 (Eq. 5), which limits magnitude but not sign. For the linear parameterization used, the partial derivative of the final prediction with respect to T_k equals K_k, so monotonicity holds if and only if K_k >= 0. No non-negativity constraint, clamping, or reparameterization is provided anywhere. Thus 'rigorously enforcing by design' reduces to an unverified premise: if K_k were negative, the same formulation would produce a decrease. The claimed guarantee is the assumption itself, not a consequence of the architecture.

  2. fitted input called prediction [Section 4.2, Eq. (6) and Section 5.4, Figure 4]
    "The loss function can be defined as the mean squared error between the true outcome Y and the sum ˆYfinal: Lfinal = 1/N ∑ ℓ(Yi, ˆYfinal,i) [...] The results demonstrate that as ∆T (credit limit increase) grows, ∆Y (risk change) exhibits a linear upward trend, with the slope for group B being notably steeper."

    The 'predicted' response ∆Y is K·∆T under the linear monotonic model, and K is learned by minimizing Lfinal (Eq. 6) on the same observed outcome Y. The risk-change versus credit-limit-increase plot in Figure 4 is therefore a display of the fitted linear function, not an independent empirical check: the upward trend is forced by the chosen parametric form and the fitted per-user slope K. Treating this fitted slope as a validated causal sensitivity makes the 'prediction' equivalent to its own fitting input.

full rationale

The paper's DML framing is largely imported from Chernozhukov et al. (2017) and does not rely on a load-bearing self-citation, so no uniqueness-imported-from-authors or ansatz-smuggled-via-citation issue appears. The main circularity is in the monotonicity contribution: the guarantee is asserted by defining f_linear as monotonic, while the only stated mechanism, L2 regularization on K, does not prevent negative sensitivities. For a linear model the claim holds exactly when K is non-negative, which is never enforced, so the 'by design' result is the unstated assumption rather than a derived property. Relatedly, the treatment response used in validation is the fitted K·∆T itself, so Figure 4's linear upward trend is a consequence of the training objective, not an external confirmation of monotonicity or debiasing. These two steps make the central claims partially circular, though the rest of the empirical pipeline is independently reported and the DML baseline is externally grounded.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central claim rests on standard causal assumptions plus two ad hoc modeling choices: the linear monotonic response with a per-user sensitivity coefficient, and the validity of the disentangled representations. The only explicit numeric hyperparameter is rho=1.9; the sensitivity K is learned from data.

free parameters (2)
  • Tweedie power parameter rho = 1.9
    Set by hand in Section 4.1 to model zero-inflated, long-tailed outcomes.
  • Per-user sensitivity coefficient K = learned (network output)
    Section 4.2: the slope of the linear monotonic model is predicted by a network and fitted to data; no non-negativity constraint is specified.
assumptions (3)
  • domain assumption SUTVA, unconfoundedness, and positivity (Assumptions 1-3) hold for the observed data
    Standard causal assumptions invoked in Section 3, not verified or tested.
  • ad hoc to paper The outcome change Delta-Y is linear in treatment change Delta-T with sensitivity coefficient K
    Section 4.2: linear parameterization is asserted to enforce monotonicity, but no evidence that the true response is linear.
  • ad hoc to paper Disentanglement of X into instrumental, confounder, and adjustment factors is valid via RLO orthogonality
    Section 4.1: orthogonality of weight vectors is assumed to yield independent latent factors, but there is no theoretical guarantee.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multi-Treatment-DML: Causal Estimation for Multi-Dimensional Continuous Treatments with Monotonicity Constraints in Personal Loan Risk Optimization." pith.science (2026). https://pith.science/paper/S6QSE4QY

@misc{pith2026250802183,
  author       = {Pith},
  title        = {Pith review of: Multi-Treatment-DML: Causal Estimation for Multi-Dimensional Continuous Treatments with Monotonicity Constraints in Personal Loan Risk Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S6QSE4QY}},
  note         = {Machine review of arXiv:2508.02183}
}
read the original abstract

Optimizing credit limits, interest rates, and loan terms is crucial for managing borrower risk and lifetime value (LTV) in personal loan platform. However, counterfactual estimation of these continuous, multi-dimensional treatments faces significant challenges: randomized trials are often prohibited by risk controls and long repayment cycles, forcing reliance on biased observational data. Existing causal methods primarily handle binary/discrete treatments and struggle with continuous, multi-dimensional settings. Furthermore, financial domain knowledge mandates provably monotonic treatment-outcome relationships (e.g., risk increases with credit limit).To address these gaps, we propose Multi-Treatment-DML, a novel framework leveraging Double Machine Learning (DML) to: (i) debias observational data for causal effect estimation; (ii) handle arbitrary-dimensional continuous treatments; and (iii) enforce monotonic constraints between treatments and outcomes, guaranteeing adherence to domain requirements.Extensive experiments on public benchmarks and real-world industrial datasets demonstrate the effectiveness of our approach. Furthermore, online A/B testing conducted on a realworld personal loan platform, confirms the practical superiority of Multi-Treatment-DML in real-world loan operations.

Figures

Figures reproduced from arXiv: 2508.02183 by the authors.

Figure 1
Figure 1. Illustration of the causal inference framework for [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Illustration of the overall architecture of Multi-Treatment DML for multi-dimensional continuous treatments. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Distribution of outcome variable in the Personal [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Trend of ∆Y versus ∆T for the outcome variable. User group etc. domain expertise, the sensitivity of risk to credit limit in￾creases is significantly lower for high-credit customers com￾pared to medium-credit customers. The results demonstrate that as ∆T (credit limit …
Figure 5
Figure 5. Figure 5: Overview of the online deployment framework for [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

25 extracted references · 20 canonical work pages

  1. [1]

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

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn 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.a...

  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 capitalize " " * FUNCT...

  3. [3]

    Y.; and Lee, D

    Almond, D.; Chay, K. Y.; and Lee, D. S. 2005. The costs of low birth weight. The Quarterly Journal of Economics, 120(3): 1031--1083

    work page 2005

  4. [4]

    Cheng, M.; Liao, X.; Liu, Q.; Ma, B.; Xu, J.; and Zheng, B. 2022. Learning disentangled representations for counterfactual regression via mutual information minimization. In Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval, 1802--1806

    work page 2022

  5. [5]

    Chernozhukov, V.; Chetverikov, D.; Demirer, M.; Duflo, E.; Hansen, C.; Newey, W.; and Robins, J. 2017. Double/debiased machine learning for treatment and causal parameters. Technical report

    work page 2017

  6. [6]

    Guelman, L.; Guill \'e n, M.; and P \'e rez-Mar \' n, A. M. 2015. A decision support framework to implement optimal personalized marketing interventions. Decision Support Systems, 72: 24--32

    work page 2015

  7. [7]

    Kitouni, O.; Nolte, N.; and Williams, M. 2023. Expressive monotonic neural networks. arXiv preprint arXiv:2307.07512

    arXiv 2023

  8. [8]

    Kuang, K.; Cui, P.; Li, B.; Jiang, M.; Yang, S.; and Wang, F. 2017. Treatment effect estimation with data-driven variable decomposition. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 31

    work page 2017

Show all 25 references
  1. [9]

    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, 116(10): 4156--4165

    2019

  2. [10]

    LaLonde, R. J. 1986. Evaluating the econometric evaluations of training programs with experimental data. The American economic review, 604--620

    1986

  3. [11]

    Liu, D.; Tang, X.; Gao, H.; Lyu, F.; and He, X. 2023. Explicit feature interaction-aware uplift network for online marketing. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 4507--4515

    2023

  4. [12]

    Ma, J.; Chen, C.; Vullikanti, A.; Mishra, R.; Madden, G.; Borrajo, D.; and Li, J. 2023. A look into causal effects under entangled treatment in graphs: Investigating the impact of contact on mrsa infection. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discover...

    2023

  5. [13]

    V.; Ravi, L.; Chen, H.; Kohli, N.; Khawand, C.; and Mandalapu, D

    Mukerji, A.; More, S.; Kannan, A. V.; Ravi, L.; Chen, H.; Kohli, N.; Khawand, C.; and Mandalapu, D. 2024. Valuing an engagement surface using a large scale dynamic causal model. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 5556--5565

    2024

  6. [14]

    Nie, L.; Ye, M.; Liu, Q.; and Nicolae, D. 2021. Vcnet and functional targeted regularization for learning causal effects of continuous treatments. arXiv preprint arXiv:2103.07861

    2021 arXiv

  7. [15]

    R.; and Rubin, D

    Rosenbaum, P. R.; and Rubin, D. B. 1983. The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1): 41--55

    1983

  8. [16]

    Rubin, D. B. 2005. Causal inference using potential outcomes: Design, modeling, decisions. Journal of the American statistical Association, 100(469): 322--331

    2005

  9. [17]

    Runje, D.; and Shankaranarayana, S. M. 2023. Constrained monotonic neural networks. In International Conference on Machine Learning, 29338--29353. PMLR

    2023

  10. [18]

    M.; and Karlen, W

    Schwab, P.; Linhardt, L.; Bauer, S.; Buhmann, J. M.; and Karlen, W. 2020. Learning counterfactual representations for estimating individual dose-response curves. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, 5612--5619

    2020

  11. [19]

    D.; and Sontag, D

    Shalit, U.; Johansson, F. D.; and Sontag, D. 2017. Estimating individual treatment effect: generalization bounds and algorithms. In International conference on machine learning, 3076--3085. PMLR

    2017

  12. [20]

    Shi, C.; Blei, D.; and Veitch, V. 2019. Adapting neural networks for the estimation of treatment effects. Advances in neural information processing systems, 32

    2019

  13. [21]

    Wehenkel, A.; and Louppe, G. 2019. Unconstrained monotonic neural networks. Advances in neural information processing systems, 32

    2019

  14. [22]

    Wu, A.; Kuang, K.; Yuan, J.; Li, B.; Wu, R.; Zhu, Q.; Zhuang, Y.; and Wu, F. 2020. Learning decomposed representation for counterfactual inference. arXiv preprint arXiv:2006.07040

    2020 arXiv

  15. [23]

    Yao, L.; Chu, Z.; Li, S.; Li, Y.; Gao, J.; and Zhang, A. 2021. A survey on causal inference. ACM Transactions on Knowledge Discovery from Data (TKDD), 15(5): 1--46

    2021

  16. [24]

    Zhang, W.; Liu, L.; and Li, J. 2021. Treatment effect estimation with disentangled latent factors. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, 10923--10930

    2021

  17. [25]

    Zhong, K.; Xiao, F.; Ren, Y.; Liang, Y.; Yao, W.; Yang, X.; and Cen, L. 2022. Descn: Deep entire space cross networks for individual treatment effect estimation. In Proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining, 4612--4620

    2022

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.