Pith. sign in

REVIEW 52 references

A Generative Framework for Causal Estimation via Importance-Weighted Diffusion Distillation

T0 review · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Distilling a pretrained diffusion model under a randomized covariate–treatment measure yields a one-step generator that estimates potential outcomes and treatment effects from observational data.

desk verdict The method (randomized marginal sampling inside diffusion distillation for one-step causal estimation) is genuinely new and the empirical direction is promising, but Theorem 1's variance-reduction claim is false—a simple counterexample with π=0.9 breaks the proof—so the paper's headline theoretical advantage is currently unsupported. read the letter →

arxiv 2505.11444 v1 pith:F5RZUJXT submitted 2025-05-16 cs.LG stat.APstat.MEstat.ML

Xinran Song , Tianyu Chen , Mingyuan Zhou This is my paper · ORCID
classification cs.LGstat.APstat.MEstat.ML
keywords causalinferencepotentialoutcomesdiffusionmodelsscoredistillationimportanceweightinginverseprobabilitytreatmenteffectestimationone-stepgeneration
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

IWDD claims that a one-step conditional generator can estimate potential outcomes and treatment effects from observational data by doing importance weighting inside diffusion distillation rather than through an explicit propensity-score reweighting. The key identity is that the inverse-probability-weighted loss equals an unweighted divergence evaluated under the product of the covariate marginal and an independent Bernoulli(1/2) treatment draw, so sampling $x$ and $z$ independently performs the RCT-style correction implicitly. The paper proves that this randomized estimator has no larger gradient variance than explicit IPW, avoids weight clipping and trimming, and yields samples in a single network pass. On synthetic shift data and benchmark datasets, the generator improves out-of-sample potential-outcome and CATE accuracy, with the highest win rates among compared baselines on ACIC 2018. If the claim holds, individualized treatment recommendations from observational data could be produced by a fast one-step generative model with no propensity network.

What carries the argument

The central object is the randomized-divergence loss $L^{\mathrm{IWDD}}_\theta = \mathbb{E}_{x\sim p_{\mathrm{data}}(x),\, z\sim \mathrm{Bernoulli}(0.5)}[D(q_\theta(y\mid x,z), p(y\mid x,z))]$, which is proved equivalent to the IPW objective while requiring no explicit propensity estimate. It is optimized by score identity distillation: an alternating scheme that trains a 'fake score' network on observational pairs to track the generator's noisy output distribution, and updates the generator from the difference between the teacher and fake scores on randomized pairs, using the Fisher divergence rather than the KL divergence. The randomization adjustment, shuffling covariates and sampling treatments from Bernoulli(0.5), is what breaks the confounding dependence between $x$ and $z$, and the variance comparison $w^2-w = 1/(4\pi(x)(1-\pi(x)))-1 \ge 0$ is what guarantees the gradient estimate is no noisier than explicit IPW.

What would settle it

On a synthetic benchmark with known potential outcomes where the training treatment rule leaves some $(x, z=1)$ region nearly empty, compare the teacher's out-of-sample RMSE for $Y(1)$ with IWDD's after training both on identical data; if the one-step generator does not beat the teacher in that low-density stratum, the claim that randomization-based distillation corrects low-density confounding is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that causal estimation can be cast as a divergence between a generator and the true outcome distribution under a randomized reference measure. Starting from the standard IPW loss $L^{\mathrm{IPW}}_\theta = \mathbb{E}_{(x,z)\sim p_{\mathrm{data}}(x,z)}[w(x,z)\,D(q_\theta(y\mid x,z), p(y\mid x,z))]$ with $w(x,z)=p_{\mathrm{data}}(x)p_{\mathrm{rct}}(z)/p_{\mathrm{data}}(x,z)$, Lemma 1 rewrites it as $L^{\mathrm{IWDD}}_\theta = \mathbb{E}_{x\sim p_{\mathrm{data}}(x),\, z\sim \mathrm{Bernoulli}(0.5)}[D(q_\theta(y\mid x,z), p(y\mid x,z))]$, so no propensity network is needed. Theorem 1 then shows the gradient covariance under this marginal-sampling estimator is dominated by that of the explicit-IPW estimator, because $w^2-w = 1/(4\pi(x)(1-\pi(x)))-1 \ge 0$ where $\pi(x)$ is the propensity score. The generator is trained by Fisher-divergence score distillation from a pretrained covariate- and treatment-conditional diffusion teacher, with the teacher queried on randomized pairs $(\tilde{x},\tilde{z})$ while a fake score network is trained on observational pairs. Empirical results on synthetic covariate-shift data and on ACIC 2016, ACIC 2018, and IHDP claim improved out-of-sample prediction of $Y(0)$, $Y(1)$, and CATE, with the largest win rates on ACIC 2018.

Load-bearing premise

The pretrained teacher diffusion model already yields accurate conditional outcome estimates in the $(x,z)$ regions that the randomized distillation loss emphasizes; if the teacher is wrong in low-density regions, distillation can only inherit and sharpen those errors.

Editorial extensions

If this is right

  • After distillation, potential outcomes for both treatment arms are produced in a single forward pass, replacing the teacher's iterative reverse-diffusion sampling and sharply reducing inference cost.
  • The randomized sampling scheme removes the need to fit a propensity network, so the arbitrariness of weight clipping and sample trimming disappears.
  • Theorem 1 implies the stochastic gradient under the IWDD loss has covariance no larger than under the explicit-IPW loss, making training more stable, especially when propensities are extreme.
  • Because the distillation target is the RCT-style reference distribution, the generator is pushed to generalize to underrepresented $(x,z)$ pairs, improving out-of-sample potential-outcome and CATE accuracy.

Reading between the lines

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

  • The Lemma 1 identity is not diffusion-specific: any truncated divergence $D$ between a generator and a teacher can be debiased by replacing $p_{\mathrm{data}}(x,z)$ with $p_{\mathrm{data}}(x)p_{\mathrm{rct}}(z)$, so the trick could transfer to other generative causal estimators such as GANs or normalizing flows.
  • The teacher sets the ceiling: on the paper's own synthetic experiment the teacher's out-of-sample $Y(1)$ RMSE is 3.07 and IWDD only reduces it to 2.76, far from the in-sample 0.12, suggesting IWDD corrects imbalance but cannot invent outcome structure the teacher never learned.
  • The Bernoulli(0.5) choice encodes a target population with balanced treatment; the same derivation would carry through with any reference distribution over $z$, so the framework offers a tunable way to target different policy populations.
  • A directly testable corollary is that the gradient-variance reduction should appear empirically as smoother loss curves under marginal sampling than under explicit IPW with the same compute budget; the paper proves the inequality but does not report a direct variance measurement.
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.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: the IWDD objective is an exact change-of-measure identity, and the causal claims are benchmarked externally; only minor self-citation in distillation provenance.

full rationale

The paper's central derivation chain is not circular. Lemma 1 (Equation 5) is an exact identity obtained by substituting the definition w(x,z)=pdata(x)prct(z)/pdata(x,z) into Equation 4; the IWDD loss is the same expectation rewritten under the product of marginals, not a prediction derived from a fitted quantity. Theorem 1 compares two unbiased estimators of the same population gradient, so even if the variance-reduction proof contained an algebraic error, that would be a correctness or soundness issue, not a circularity issue: neither estimator is fitted to the quantity being predicted. The randomization-based adjustment (Algorithm 1) changes the sampling distribution for the generator while keeping the teacher and fake-score networks fixed, and the empirical evaluation is against external benchmarks (ACIC 2016, ACIC 2018, IHDP) plus the pretrained teacher, so the state-of-the-art claim is not forced by the construction of the loss. The main self-citation is the adoption of SiD (Zhou et al., 2024) and denoising score distillation (Chen et al., 2025) as the distillation machinery in Section 3.2.2; this is method provenance from published prior work, including one shared author, but the causal-application claim and the comparison to independent baselines do not reduce to those citations. Accordingly, there is no load-bearing circular step and no fitted input is renamed as a prediction.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The method rests on standard causal identification assumptions, the reliability of the pretrained teacher in low-density regions, and SiD's alternating-distillation correctness. Free parameters are limited to hyperparameters alpha and sigma_init; no invented entities are introduced. The variance-reduction theorem is exact, but the teacher-reliability premise is the weakest link.

free parameters (2)
  • alpha (generator loss blending coefficient) = 0.7 (default; tuned over {0.3, 0.4, 0.5, 0.6, 0.7, 1.0, 1.2} in Appendix C)
    Balances the two terms in the generator loss Eq. 6. The ablation shows performance varies with alpha (e.g., IHDP RMSE improves from 1.197 to 1.056 when alpha goes from 0.3 to 0.7), so the reported results depend on this hand-chosen hyperparameter.
  • sigma_init (generator initial noise level) = 2.5
    Sets the noise scale for the one-step generator's input and is fixed throughout distillation and evaluation. No justification or sensitivity analysis is given for this choice.
assumptions (4)
  • domain assumption Consistency, unconfoundedness, and overlap (Assumption 1).
    Identifiability of potential outcome conditionals from observational data, stated in Section 3. If unmeasured confounding exists, the teacher model estimates the wrong conditional and IWDD propagates the bias.
  • domain assumption The pretrained diffusion model f_phi accurately approximates p(y|x,z) across the support of pdata(x) x {0,1}.
    The distillation loss Eq. 5 evaluates the teacher at randomized (x,z) pairs, many of which may be low-density in the training data. The synthetic experiment in Section 4.1 shows the teacher's out-of-sample Y(1) RMSE is 3.07, directly challenging this premise.
  • standard math The Fisher divergence objective, optimized via SiD (Zhou et al., 2024), provides an unbiased or sufficiently accurate gradient for the generator loss.
    The method adopts SiD's alternating optimization and score identities without re-deriving them, relying on the correctness of that prior framework.
  • standard math The importance weight identity pdata(x,z) = pdata(x) pdata(z|x) holds.
    Definitional; used in Lemma 1 and Appendix A.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Generative Framework for Causal Estimation via Importance-Weighted Diffusion Distillation." pith.science (2026). https://pith.science/paper/F5RZUJXT

@misc{pith2026250511444,
  author       = {Pith},
  title        = {Pith review of: A Generative Framework for Causal Estimation via Importance-Weighted Diffusion Distillation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F5RZUJXT}},
  note         = {Machine review of arXiv:2505.11444}
}
read the original abstract

Estimating individualized treatment effects from observational data is a central challenge in causal inference, largely due to covariate imbalance and confounding bias from non-randomized treatment assignment. While inverse probability weighting (IPW) is a well-established solution to this problem, its integration into modern deep learning frameworks remains limited. In this work, we propose Importance-Weighted Diffusion Distillation (IWDD), a novel generative framework that combines the pretraining of diffusion models with importance-weighted score distillation to enable accurate and fast causal estimation-including potential outcome prediction and treatment effect estimation. We demonstrate how IPW can be naturally incorporated into the distillation of pretrained diffusion models, and further introduce a randomization-based adjustment that eliminates the need to compute IPW explicitly-thereby simplifying computation and, more importantly, provably reducing the variance of gradient estimates. Empirical results show that IWDD achieves state-of-the-art out-of-sample prediction performance, with the highest win rates compared to other baselines, significantly improving causal estimation and supporting the development of individualized treatment strategies. We will release our PyTorch code for reproducibility and future research.

Figures

Figures reproduced from arXiv: 2505.11444 by the authors.

Figure 1
Figure 1. Overview of IWDD. We first pretrain a conditional diffusion model fϕ(y | x, z) to approximate the true conditional distribution p(y | x, z) over observational data. In the distillation stage, we train a generator qθ(y | x, z) using marginal sampling, which implicitly applies importance weighting without requiring explicit propensity estimation. We apply a randomization-based sampling adjustment: covariates x are shu… view at source ↗
Figure 2
Figure 2. Marginal distributions of y in training and testing sets. Due to the shift in treatment as￾signment, the induced distribution of y differs across domains. We designed a synthetic data experiment to visualize the effective￾ness of IWDD in addressing out-of-sample estimation challenges under distribution shift. The covariate distribution p(x) ∼ N (0, 1) and outcome model y = f(x, z) + ϵ, ϵ ∼ N (0, 1) remain fixed acro… view at source ↗
Figure 3
Figure 3. Synthetic data example: estimated potential outcome distributions Y (0) and Y (1) from different models. The pretrained diffusion model performs well in-sample and for Y (0) out-of-sample, but struggles with Y (1). IWDD improves estimation for out-of-sample Y (1) while maintaining performance elsewhere. training distribution, with an RMSE of 3.07. In contrast, IWDD (Row 3) improves estimation, with substantial gain … view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 20 canonical work pages

  1. [1]

    Interventional and Counterfactual Inference with Diffusion Models , June 2023

    Patrick Chao, Patrick Blöbaum, and Shiva Prasad Kasiviswanathan. Interventional and Counterfactual Inference with Diffusion Models , June 2023. URL http://arxiv.org/abs/2302.00860. arXiv:2302.00860 [cs, stat]

    arXiv 2023

  2. [2]

    Denoising score distillation: From noisy diffusion pretraining to one-step high-quality generation

    Tianyu Chen, Yasi Zhang, Zhendong Wang, Ying Nian Wu, Oscar Leong, and Mingyuan Zhou. Denoising score distillation: From noisy diffusion pretraining to one-step high-quality generation. arXiv preprint arXiv:2503.07578, 2025

    arXiv 2025

  3. [3]

    On inductive biases for heterogeneous treatment effect estimation, 2021 a

    Alicia Curth and Mihaela van der Schaar. On inductive biases for heterogeneous treatment effect estimation, 2021 a . URL https://arxiv.org/abs/2106.03765

    arXiv 2021

  4. [4]

    Nonparametric estimation of heterogeneous treatment effects: From theory to learning algorithms

    Alicia Curth and Mihaela van der Schaar. Nonparametric estimation of heterogeneous treatment effects: From theory to learning algorithms. In Proceedings of the 24th International Conference on Artificial Intelligence and Statistics (AISTATS). PMLR, 2021 b

    work page 2021

  5. [5]

    A first course in causal inference, 2023

    Peng Ding. A first course in causal inference, 2023. URL https://arxiv.org/abs/2305.18793

    arXiv 2023

  6. [6]

    Tweedie’s formula and selection bias

    Bradley Efron. Tweedie’s formula and selection bias. Journal of the American Statistical Association, 106 0 (496): 0 1602--1614, 2011

    2011

  7. [7]

    Ruth T. Gross. Infant health and development program (ihdp): Enhancing the outcomes of low birth weight, premature infants in the united states, 1985-1988, 2024. URL https://doi.org/10.3886/ICPSR09795.v2

    doi:10.3886/icpsr09795.v2 1985

  8. [8]

    Card: Classification and regression diffusion models

    Xizewen Han, Huangjie Zheng, and Mingyuan Zhou. Card: Classification and regression diffusion models. In Thirty-Sixth Conference on Neural Information Processing Systems, 2022

    work page 2022

Show all 52 references
  1. [9]

    Qijia He, Fei Gao, Oliver Dukes, Sinead Delany-Moretlwe, and Bo Zhang. Generalizing the intention-to-treat effect of an active control against placebo from historical placebo-controlled trials to an active-controlled trial: A case study of the efficacy of daily oral tdf/ftc in...

    2023 arXiv

  2. [10]

    Denoising diffusion probabilistic models, 2020

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models, 2020. URL https://arxiv.org/abs/2006.11239

    2020 arXiv

  3. [11]

    Paul W. Holland. Statistics and causal inference. Journal of the American Statistical Association, 81 0 (396): 0 945--960, 1986. ISSN 01621459, 1537274X. URL http://www.jstor.org/stable/2289064

    1986

  4. [12]

    Imbens and Donald B

    Guido W. Imbens and Donald B. Rubin. Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction. Cambridge University Press, New York, 2015

    2015

  5. [13]

    D eep M atch: Balancing deep covariate representations for causal inference using adversarial training

    Nathan Kallus. D eep M atch: Balancing deep covariate representations for causal inference using adversarial training. In Hal Daumé III and Aarti Singh, editors, Proceedings of the 37th International Conference on Machine Learning, volume 119 of Proceedings of Machine Learning...

    2020

  6. [14]

    Elucidating the design space of diffusion-based generative models

    Tero Karras, Miika Aittala, Timo Aila, and Samuli Laine. Elucidating the design space of diffusion-based generative models. In Proc. NeurIPS, 2022

    2022

  7. [15]

    Edward H. Kennedy. Towards optimal doubly robust estimation of heterogeneous causal effects, 2023. URL https://arxiv.org/abs/2004.14497

    2023 arXiv

  8. [16]

    Causal Diffusion Autoencoders : Toward Counterfactual Generation via Diffusion Probabilistic Models , May 2024

    Aneesh Komanduri, Chen Zhao, Feng Chen, and Xintao Wu. Causal Diffusion Autoencoders : Toward Counterfactual Generation via Diffusion Probabilistic Models , May 2024. URL http://arxiv.org/abs/2404.17735. arXiv:2404.17735 [cs, stat]

    2024 arXiv

  9. [17]

    Künzel, Jasjeet S

    Sören R. Künzel, Jasjeet S. Sekhon, Peter J. Bickel, and Bin Yu. Metalearners for estimating heterogeneous treatment effects using machine learning. Proceedings of the National Academy of Sciences, 116 0 (10): 0 4156–4165, February 2019. ISSN 1091-6490. doi:10.1073/pnas.180459...

    2019 doi

  10. [18]

    Variance reduction in the inverse probability weighted estimators for the average treatment effect using the propensity score

    Jiangang Liao and Charles Rohde. Variance reduction in the inverse probability weighted estimators for the average treatment effect using the propensity score. Biometrics, 78 0 (2): 0 660--667, 2022. doi:10.1111/biom.13454

    2022 doi

  11. [19]

    Causal modeling with stationary diffusions

    Lars Lorch, Andreas Krause, and Bernhard Sch\" o lkopf. Causal modeling with stationary diffusions. In Sanjoy Dasgupta, Stephan Mandt, and Yingzhen Li, editors, Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, volume 238 of Proceeding...

    1927

  12. [20]

    Diff-I nstruct: A universal approach for transferring knowledge from pre-trained diffusion models

    Weijian Luo, Tianyang Hu, Shifeng Zhang, Jiacheng Sun, Zhenguo Li, and Zhihua Zhang. Diff-I nstruct: A universal approach for transferring knowledge from pre-trained diffusion models. In Thirty-seventh Conference on Neural Information Processing Systems, 2023. URL https://open...

    2023

  13. [21]

    Diff-instruct: A universal approach for transferring knowledge from pre-trained diffusion models, 2024

    Weijian Luo, Tianyang Hu, Shifeng Zhang, Jiacheng Sun, Zhenguo Li, and Zhihua Zhang. Diff-instruct: A universal approach for transferring knowledge from pre-trained diffusion models, 2024. URL https://arxiv.org/abs/2305.18455

    2024 arXiv

  14. [22]

    Diff PO : A causal diffusion model for learning distributions of potential outcomes

    Yuchen Ma, Valentyn Melnychuk, Jonas Schweisthal, and Stefan Feuerriegel. Diff PO : A causal diffusion model for learning distributions of potential outcomes. In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. URL https://openreview.net/foru...

    2024

  15. [23]

    Empirical analysis of model selection for heterogeneous causal effect estimation, 2024

    Divyat Mahajan, Ioannis Mitliagkas, Brady Neal, and Vasilis Syrgkanis. Empirical analysis of model selection for heterogeneous causal effect estimation, 2024. URL https://arxiv.org/abs/2211.01939

    2024 arXiv

  16. [24]

    Diffusion based causal representation learning, 2023

    Amir Mohammad Karimi Mamaghan, Andrea Dittadi, Stefan Bauer, Karl Henrik Johansson, and Francesco Quinzan. Diffusion based causal representation learning, 2023. URL https://arxiv.org/abs/2311.05421

    2023 arXiv

  17. [25]

    Causality: Models, Reasoning and Inference

    Judea Pearl. Causality: Models, Reasoning and Inference. Cambridge University Press, USA, 2nd edition, 2009. ISBN 052189560X

    2009

  18. [26]

    Trygve haavelmo and the emergence of causal calculus

    Judea Pearl. Trygve haavelmo and the emergence of causal calculus. Econometric Theory, 31 0 (1): 0 152–179, 2015. doi:10.1017/S0266466614000231

    2015 doi

  19. [27]

    Barron, and Ben Mildenhall

    Ben Poole, Ajay Jain, Jonathan T. Barron, and Ben Mildenhall. Dreamfusion: Text-to-3d using 2d diffusion, 2022. URL https://arxiv.org/abs/2209.14988

    2022 arXiv

  20. [28]

    An empirical B ayes approach to statistics

    Herbert E Robbins. An empirical B ayes approach to statistics. In Breakthroughs in Statistics: Foundations and basic theory, pages 388--394. Springer, 1992

    1992

  21. [29]

    Robins, Andrea Rotnitzky, and Lue Ping Zhao

    James M. Robins, Andrea Rotnitzky, and Lue Ping Zhao. Estimation of regression coefficients when some regressors are not always observed. Journal of the American Statistical Association, 89 0 (427): 0 846--866, 1994. doi:10.2307/2290910

    1994 doi

  22. [30]

    Rosenbaum and Donald B

    Paul R. Rosenbaum and Donald B. Rubin. The central role of the propensity score in observational studies for causal effects. Biometrika, 70 0 (1): 0 41--55, April 1983. ISSN 0006-3444. doi:10.1093/biomet/70.1.41. URL https://doi.org/10.1093/biomet/70.1.41

    1983 doi

  23. [31]

    Donald B. Rubin. Causal inference using potential outcomes: Design, modeling, decisions. Journal of the American Statistical Association, 100 0 (469): 0 322--331, 2005. doi:10.1198/016214504000001880. URL https://doi.org/10.1198/016214504000001880

    2005 doi

  24. [32]

    Tsaftaris

    Pedro Sanchez and Sotirios A. Tsaftaris. Diffusion causal models for counterfactual estimation, 2022. URL https://arxiv.org/abs/2202.10166

    2022 arXiv

  25. [33]

    O'Neil, and Sotirios A

    Pedro Sanchez, Xiao Liu, Alison Q. O'Neil, and Sotirios A. Tsaftaris. Diffusion Models for Causal Discovery via Topological Ordering , June 2023. URL http://arxiv.org/abs/2210.06201. arXiv:2210.06201 [cs]

    2023 arXiv

  26. [34]

    Johansson, and David Sontag

    Uri Shalit, Fredrik D. Johansson, and David Sontag. Estimating individual treatment effect: generalization bounds and algorithms. In Doina Precup and Yee Whye Teh, editors, Proceedings of the 34th International Conference on Machine Learning, volume 70 of Proceedings of Machin...

    2017

  27. [35]

    Diffusion Model in Causal Inference with Unmeasured Confounders , December 2023

    Tatsuhiro Shimizu. Diffusion Model in Causal Inference with Unmeasured Confounders , December 2023. URL http://arxiv.org/abs/2308.03669. arXiv:2308.03669 [cs, stat]

    2023 arXiv

  28. [36]

    Weiss, Niru Maheswaranathan, and Surya Ganguli

    Jascha Sohl-Dickstein, Eric A. Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsupervised learning using nonequilibrium thermodynamics, 2015. URL https://arxiv.org/abs/1503.03585

    2015 arXiv

  29. [37]

    Generative M odeling by E stimating G radients of the D ata D istribution

    Yang Song and Stefano Ermon. Generative M odeling by E stimating G radients of the D ata D istribution. In Advances in Neural Information Processing Systems, pages 11918--11930, 2019

    2019

  30. [38]

    Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole

    Yang Song, Jascha Sohl - Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations. CoRR, abs/2011.13456, 2020. URL https://arxiv.org/abs/2011.13456

    2011 arXiv

  31. [39]

    Elizabeth A. Stuart. Matching methods for causal inference: A review and a look forward. Statistical Science, 25 0 (1): 0 1--21, 2010

    2010

  32. [40]

    Covariate shift adaptation by importance weighted cross validation

    Masashi Sugiyama, Matthias Krauledat, and Klaus-Robert M \"u ller. Covariate shift adaptation by importance weighted cross validation. Journal of Machine Learning Research, 8 0 (5), 2007

    2007

  33. [41]

    General identifiability and achievability for causal representation learning

    Burak Varici, Emre Acart\" u rk, Karthikeyan Shanmugam, and Ali Tajer. General identifiability and achievability for causal representation learning. In Sanjoy Dasgupta, Stephan Mandt, and Yingzhen Li, editors, Proceedings of The 27th International Conference on Artificial Inte...

    2024

  34. [42]

    A connection between score matching and denoising autoencoders

    Pascal Vincent. A connection between score matching and denoising autoencoders. Neural computation, 23 0 (7): 0 1661--1674, 2011

    2011

  35. [43]

    Diffusion-GAN : Training GANs with diffusion

    Zhendong Wang, Huangjie Zheng, Pengcheng He, Weizhu Chen, and Mingyuan Zhou. Diffusion-GAN : Training GANs with diffusion. International Conference on Learning Representations (ICLR), 2022

    2022

  36. [44]

    Prolific D reamer: High-fidelity and diverse text-to- 3D generation with variational score distillation, 2023 a

    Zhengyi Wang, Cheng Lu, Yikai Wang, Fan Bao, Chongxuan Li, Hang Su, and Jun Zhu. Prolific D reamer: High-fidelity and diverse text-to- 3D generation with variational score distillation, 2023 a

    2023

  37. [45]

    Prolificdreamer: High-fidelity and diverse text-to-3d generation with variational score distillation, 2023 b

    Zhengyi Wang, Cheng Lu, Yikai Wang, Fan Bao, Chongxuan Li, Hang Su, and Jun Zhu. Prolificdreamer: High-fidelity and diverse text-to-3d generation with variational score distillation, 2023 b . URL https://arxiv.org/abs/2305.16213

    2023 arXiv

  38. [46]

    Semi-implicit variational inference

    Mingzhang Yin and Mingyuan Zhou. Semi-implicit variational inference. In International Conference on Machine Learning, pages 5660--5669, 2018

    2018

  39. [47]

    One-step diffusion with distribution matching distillation

    Tianwei Yin, Micha \"e l Gharbi, Richard Zhang, Eli Shechtman, Fr \'e do Durand, William T Freeman, and Taesung Park. One-step diffusion with distribution matching distillation. In CVPR, 2024

    2024

  40. [48]

    GANITE : Estimation of individualized treatment effects using generative adversarial nets

    Jinsung Yoon, James Jordon, and Mihaela van der Schaar. GANITE : Estimation of individualized treatment effects using generative adversarial nets. In International Conference on Learning Representations, 2018. URL https://openreview.net/forum?id=ByKWUeWA-

    2018

  41. [49]

    Hierarchical semi-implicit variational inference with application to diffusion model acceleration

    Longlin Yu, Tianyu Xie, Yu Zhu, Tong Yang, Xiangyu Zhang, and Cheng Zhang. Hierarchical semi-implicit variational inference with application to diffusion model acceleration. In Thirty-seventh Conference on Neural Information Processing Systems, 2023. URL https://openreview.net...

    2023

  42. [50]

    Treatment effect estimation with disentangled latent factors

    Weijia Zhang, Lin Liu, and Jiuyong Li. Treatment effect estimation with disentangled latent factors. Proceedings of the AAAI Conference on Artificial Intelligence, 35 0 (12): 0 10923--10930, May 2021. doi:10.1609/aaai.v35i12.17304. URL https://ojs.aaai.org/index.php/AAAI/artic...

    2021 doi

  43. [51]

    Truncated diffusion probabilistic models

    Huangjie Zheng, Pengcheng He, Weizhu Chen, and Mingyuan Zhou. Truncated diffusion probabilistic models. arXiv preprint arXiv:2202.09671, 2022

    2022 arXiv

  44. [52]

    Score identity distillation: Exponentially fast distillation of pretrained diffusion models for one-step generation

    Mingyuan Zhou, Huangjie Zheng, Zhendong Wang, Mingzhang Yin, and Hai Huang. Score identity distillation: Exponentially fast distillation of pretrained diffusion models for one-step generation. In International Conference on Machine Learning, 2024. URL https://arxiv.org/abs/2404.04057

    2024 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.