Pith. sign in

REVIEW 3 major objections 4 minor 63 references

SurvDiff: A Diffusion Model for Generating Synthetic Data in Survival Analysis

T0 review · 3 major / 4 minor · 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 SurvDiff, an end-to-end diffusion model, jointly generates covariates, event times, and right-censoring and, the paper reports, outperforms existing generators on fidelity and downstream survival metrics across three medical datasets.

desk verdict Solid application paper that overclaims: the Cox auxiliary loss does not by itself preserve censoring, and several tables contradict the 'consistently outperforms' framing. read the letter →

arxiv 2509.22352 v3 pith:CDQUNS4L submitted 2025-09-26 cs.LG cs.AI

Marie Brockschmidt , Maresa Schröder , Stefan Feuerriegel This is my paper · ORCID
classification cs.LGcs.AI
keywords survivalanalysissyntheticdatagenerationdiffusionmodelsright-censoringtime-to-eventtabularCoxproportionalhazardsclinicalmachinelearning
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

Survival data are hard to synthesize because event times are often right-censored, and a synthetic cohort must reproduce both the covariate distribution and the censoring mechanism. SurvDiff is a diffusion model that generates covariates, event times, and the censoring indicator together in one end-to-end pass, with an added loss term that ranks patients by survival risk using a weighted Cox partial likelihood. On the AIDS, GBSG2, and METABRIC medical datasets, the paper reports that SurvDiff matches real covariate distributions more closely than existing tabular and survival generators and yields better or equal downstream survival-model performance, with the largest gains on heavily censored data. The paper positions SurvDiff as the first end-to-end diffusion model designed for generating synthetic survival data.

What carries the argument

The central object is the combined objective $L_{\text{total}} = L_{\text{diff}} + \lambda_{\text{surv}} L_{\text{surv}}$. SurvDiff uses a variance-exploding Gaussian diffusion for continuous covariates and event times, plus a masked multinomial diffusion for discrete covariates and the event indicator, following the TabDiff architecture. A denoising transformer reconstructs original values from noise, and a small survival head maps the denoised covariates to a scalar risk score. The survival loss is a weighted Cox partial negative log-likelihood in which only observed events contribute to the numerator while censored patients remain in the risk sets, with an exponential decay weight for late events to reduce instability from small risk sets; an adaptively scaled $\lambda_{\text{surv}}$ balances this loss against the diffusion reconstruction loss.

What would settle it

On a simulated survival dataset with known event and censoring time generators, fit SurvDiff and compare the conditional distribution of the censoring indicator and censoring times in synthetic versus real samples; any material divergence—for example, a covariate that predicts censoring in the truth but not in the synthetic data—would falsify the censoring-preservation claim.

Watch

Extended reading notes

Core claim

The central claim is that a single diffusion model can jointly generate mixed-type covariates, event times, and right-censoring indicators, and that a survival-tailored loss—an extension of the Cox partial likelihood with exponential decay weighting that downweights late events—is what makes censoring-aware generation work. The paper contrasts this end-to-end design with SurvivalGAN, which generates covariates first and then samples event times conditionally, arguing that the staged approach propagates errors and loses censoring fidelity. In experiments across three clinical cohorts, SurvDiff attains lower Jensen–Shannon and Wasserstein distances for covariates, better C-indices and Brier scores in train-on-synthetic/test-on-real evaluation, and competitive Kaplan–Meier and RMST metrics; the authors conclude that the survival loss and joint generation are responsible for the gains, not the diffusion backbone alone.

Load-bearing premise

The method rests on the assumption that a survival-risk loss computed on noisy, partially denoised versions of the data teaches the generator to reproduce the real censoring mechanism, even though no part of the training objective directly checks that mechanism.

Editorial extensions

If this is right

  • Survival models trained on SurvDiff's synthetic cohorts and tested on real patients achieve better C-index and Brier scores than models trained on other generators' synthetic data, the paper's main downstream test.
  • The largest gains occur on the most heavily censored datasets, so the survival-tailored loss is doing the work rather than the diffusion backbone alone.
  • Joint end-to-end generation avoids the error propagation of staged generators, so covariates, event times, and censoring indicators come from one model instead of separately fitted pieces.
  • In downsampled cohorts, SurvDiff's advantage over TabDiff grows, indicating the method transfers to small-sample clinical settings.

Reading between the lines

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

  • A direct test the paper does not report is whether the synthetic censoring-time distribution matches the real one: comparing censoring Kaplan-Meier curves or the covariate-conditional censoring rate would settle the strongest claim more cleanly than the reported downstream metrics.
  • Because the survival head is built on a Cox partial likelihood, SurvDiff's generated event-time structure may inherit the proportional-hazards assumption; on data with crossing hazards, a non-proportional survival head would be a natural variant to test.
  • The same joint-generation recipe—diffusion reconstruction loss plus a loss that scores what downstream models will predict—appears portable to competing risks or interval censoring, since only the outcome head and risk-set definition would change.
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

3 major / 4 minor

Summary. The paper proposes SurvDiff, a diffusion model for generating synthetic survival data. It jointly generates mixed-type covariates, event times, and a right-censoring indicator via a masked-diffusion process for discrete variables and a Gaussian diffusion process for continuous variables, with an added Cox partial-likelihood loss on risk scores computed from denoised covariates. The authors claim that this survival-tailored loss reproduces event-time distributions and preserves the censoring mechanism, and they report experiments on AIDS, GBSG2, and METABRIC comparing against NFlow, TVAE, CTGAN, TabDiff, and SurvivalGAN. The main empirical claims are that SurvDiff consistently outperforms baselines in covariate fidelity and downstream survival metrics.

Significance. If the empirical claims held, SurvDiff would be a useful contribution as the first diffusion model tailored to synthetic survival data, with the practical benefit of end-to-end joint generation. The paper includes a broad benchmark across three medical datasets, five generative baselines, five downstream survival models, and several fidelity metrics, and it reports means and standard deviations over ten seeds. The idea of augmenting a diffusion objective with a Cox-based loss is plausible and relevant. However, the significance is currently limited by the gap between the strong abstract claim of consistent outperformance and the mixed results in the tables, and by the unsupported claim that the Cox loss preserves the censoring mechanism.

major comments (3)
  1. [Abstract and Section 5.1, Table 3] The abstract claims that SurvDiff 'consistently outperforms state-of-the-art generative baselines' in distributional fidelity and survival metrics; this is contradicted by the reported tables. In Table 2, TVAE has a lower Wasserstein distance on AIDS (0.0779 vs 0.0869) and TabDiff has a lower Wasserstein distance on METABRIC (0.0539 vs 0.0547); in Table 3, TabDiff has a slightly higher C-index on METABRIC (0.6118 vs 0.6117); and in Table 7, TabDiff has a lower RMST gap on AIDS (0.0066 vs 0.0134), SurvivalGAN has a lower KM MSE on AIDS (0.0002 vs 0.0003), and TabDiff has a lower KM MSE on GBSG2 (0.0005 vs 0.0062). The claim of consistent outperformance should be qualified, or the experiments and abstract should be revised to match the observed results.
  2. [Section 4.3, Eq. (10)] The central claim that the survival-tailored loss 'preserves the censoring mechanism' is not supported by the formulation. L_surv in Eq. (10) is a Cox partial likelihood on risk scores r_i computed from denoised covariates, with risk sets based on observed times; it does not model the censoring distribution, and a Cox partial likelihood is valid under independent censoring regardless of the censoring mechanism, so it cannot by itself encode or preserve that mechanism. The generated event time T and event indicator E are supervised only by the generic diffusion terms Lcont and Ldisc, which a plain tabular diffusion baseline also has. To support the claim, the paper should either provide a mechanism by which L_surv shapes the censoring distribution or report direct tests of censoring (e.g., censoring-time distributions or conditional censoring probabilities).
  3. [Supplement D, Table 7] The survival-fidelity evidence is mixed and the paper itself acknowledges that the RMST gap can mask compensating deviations. In Table 7, SurvDiff does not dominate: on AIDS, TabDiff has a lower RMST gap (0.0066 vs 0.0134) and SurvivalGAN has a lower KM MSE (0.0002 vs 0.0003); on GBSG2, TabDiff has a lower KM MSE (0.0005 vs 0.0062). The paper describes these results as 'broadly comparable performance,' which is inconsistent with the abstract's 'consistently outperforms.' The authors should either temper the contribution statement or add additional survival metrics that directly assess the censoring distribution.
minor comments (4)
  1. [Section 4, Figure 2] The text refers to 'Figure 4.3' when describing the three components; this should be 'Figure 2'.
  2. [Section 6, Conclusion] The sentence 'SurvDiff generating reliable synthetic datasets' is missing a verb; it should be 'SurvDiff generates reliable synthetic datasets.'
  3. [Appendix B, Tables 4 and 6] Several hyperparameter entries lack a space, e.g., 'No. Epochs4000', 'No. Iterations1500'; this should be fixed for readability.
  4. [Abstract] The abstract contains a duplicated phrase: 'in both distributional fidelity and survival model evaluation metrics across multiple medical datasets' and later 'across multiple datasets'; consider removing one occurrence.

Circularity Check

0 steps flagged · score 1.0 of 10

No material circularity: SurvDiff's contribution is an empirical benchmark, and the Cox survival loss is a training signal rather than a fitted prediction; self-citations are not load-bearing.

full rationale

SurvDiff derives no result by definitional identity. The central claim is evaluated by benchmarking synthetic data against external baselines (NFlow, TVAE, CTGAN, TabDiff, SurvivalGAN) using covariate-fidelity and train-on-synthetic/test-on-real survival metrics. The survival-tailored loss L_surv (Eq. 10) is a weighted Cox partial likelihood computed on observed event times and event indicators; it is an auxiliary training objective that shapes denoised covariates, but it does not define the generated event times or censoring indicators, which are supervised by the standard diffusion terms Lcont and Ldisc (Eqs. 6, 8, 9). No parameter is fitted to a subset of data and then reported as a prediction: the sparsity weights (Eq. 11) and adaptive lambda (Eq. 13) are training stabilizers/hyperparameters, not outputs. There is no load-bearing self-citation: references to Frauen et al. (2025), Ma et al. (2024, 2025), and Amad et al. (2025) are related-work context, and the architecture/noise schedules are credited to external TabDiff work (Shi et al., 2024b). The abstract's stronger wording that L_surv 'preserves the censoring mechanism' is not supported by the equations (a Cox partial likelihood does not model the censoring distribution), but that is an empirical-validity concern, not circularity. One minor non-circular note: because L_surv shares an objective family with the downstream C-index evaluation, some TSTR gain is expected from objective alignment rather than from a closed-form reduction. Overall, no circular step can be exhibited, so the score is 1 rather than 0.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central method rests on standard diffusion assumptions plus the ad hoc assumption that the Cox auxiliary loss transfers to faithful event-time and censoring generation. No new physical entities are introduced. Several loss-controlling hyperparameters are either fit adaptively or unreported, which limits reproducibility.

free parameters (5)
  • lambda_surv (survival loss weight) = adaptive, computed via Eq 13; not reported in final form
    Controls the balance between diffusion loss and survival loss; chosen by calibration to target a fraction of the total objective.
  • alpha_surv = 0.3
    Target fraction of survival loss in the total objective; a fixed hyperparameter that shapes the final model.
  • tau (late-event weighting threshold) = not reported
    Threshold in Eq 11 separating events that receive full weight from those that are exponentially downweighted; value missing from Table 4.
  • alpha (late-event decay rate) = not reported
    Exponential decay rate in Eq 11 controlling how quickly late events are downweighted; value missing from Table 4.
  • lambda_max = not reported
    Upper bound on lambda_surv in Eq 13; not specified in the hyperparameter table.
assumptions (4)
  • standard math The learned score function mu_theta approximates the true score sufficiently to generate samples from P_X,E,T.
    Standard diffusion model assumption in Section 3.1, required for reverse SDE sampling.
  • ad hoc to paper A Cox partial-likelihood loss evaluated on denoised covariates at training time steers the generated joint distribution to preserve time-to-event and censoring structure.
    Core novelty in Section 4.3; no theorem or experiment isolates this transfer, so it is assumed to make the method work.
  • ad hoc to paper Downweighting late events via Eq 11 improves stability without distorting the event-time distribution.
    The weighting alters the training objective; its effect on the generative distribution is not measured, and Table 7 shows mixed event-time metrics.
  • domain assumption Masked diffusion for the discrete event indicator and Gaussian diffusion for event time can represent the joint dependency between censoring and event time.
    The model adds no explicit coupling between E and T beyond the shared latent and the survival loss; the assumption is that the diffusion process captures this dependency.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SurvDiff: A Diffusion Model for Generating Synthetic Data in Survival Analysis." pith.science (2026). https://pith.science/paper/CDQUNS4L

@misc{pith2026250922352,
  author       = {Pith},
  title        = {Pith review of: SurvDiff: A Diffusion Model for Generating Synthetic Data in Survival Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CDQUNS4L}},
  note         = {Machine review of arXiv:2509.22352}
}
read the original abstract

Survival analysis is a cornerstone of clinical research by modeling time-to-event outcomes such as metastasis, disease relapse, or patient death. Unlike standard tabular data, survival data often come with incomplete event information due to dropout, or loss to follow-up. This poses unique challenges for synthetic data generation, where it is crucial for clinical research to faithfully reproduce both the event-time distribution and the censoring mechanism. In this paper, we propose SurvDiff an end-to-end diffusion model specifically designed for generating synthetic data in survival analysis. SurvDiff is tailored to capture the data-generating mechanism by jointly generating mixed-type covariates, event times, and right-censoring, guided by a survival-tailored loss function. The loss encodes the time-to-event structure and directly optimizes for downstream survival tasks, which ensures that SurvDiff (i) reproduces realistic event-time distributions and (ii) preserves the censoring mechanism. Across multiple datasets, we show that SurvDiff consistently outperforms state-of-the-art generative baselines in both distributional fidelity and survival model evaluation metrics across multiple medical datasets. To the best of our knowledge, SurvDiff is the first end-to-end diffusion model explicitly designed for generating synthetic survival data.

Figures

Figures reproduced from arXiv: 2509.22352 by the authors.

Figure 1
Figure 1. SURVDIFF for generating synthetic survival data. Our SURVDIFF generates synthetic samples that retain the structure of the original data, including high-fidelity covariate distributions and faithful event-time distributions while pre￾serving the censoring mechanism. The synthetic dataset can then be used to train downstream survival models without direct access to the original patient-level data. Survival analysis i… view at source ↗
Figure 2
Figure 2. Overview of our SURVDIFF. SURVDIFF consisting of A forward diffusion, the B backward diffusion and the C novel survival-focused loss. Importantly, we distinguish the role of E (event indicator; binary) and T (time-to￾event; continuous), which progress along different noising schemes due to the different variable types. This choice stabilizes the balance between diffusion and survival signals. The fixed calibrated we… view at source ↗
Figure 3
Figure 3. t-SNE visualization of covariate fidelity of real and synthetic data on METABRIC. ⇒ Takeaway: Synthetic samples from SURVDIFF are well aligned with the original data. SURVDIFF achieves high covariate fidelity. • Covariate distribution fidelity: We report the covariate diversity in Ta￾ble 2. We observe the following: SURVDIFF consistently outperforms all other methods in terms of the JS distance across all datasets. … view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: t-SNE visualization of covariate fidelity of real and synthetic data on GBSG2. • Downstream performance: In Ta￾ble 3, we evaluate the performance of all models on downstream survival tasks. We observe that (1) SURVDIFF consistently achieves large improve￾ments over Sur…
Figure 5
Figure 5. Figure 5: t-SNE visualization of covariate fidelity on the AIDS dataset [PITH_FULL_IMAGE:figures/full_fig_p016_5.png]
Figure 6
Figure 6. Figure 6: t-SNE visualization and KM curves on the AIDS dataset. 16 [PITH_FULL_IMAGE:figures/full_fig_p016_6.png]
Figure 7
Figure 7. Figure 7: t-SNE visualization and KM curves on the GBSG2 dataset [PITH_FULL_IMAGE:figures/full_fig_p017_7.png]
Figure 8
Figure 8. Figure 8: t-SNE visualization and KM curves on the METABRIC dataset. 17 [PITH_FULL_IMAGE:figures/full_fig_p017_8.png]
Figure 9
Figure 9. Figure 9: Training dynamics of SURVDIFF. Shown are the discrete/categorical diffusion loss Ldisc, the continuous diffusion loss Lcont, the Cox survival loss Lsurv, and the total objective Ltotal = Ldiff + λsurvLsurv. 20 [PITH_FULL_IMAGE:figures/full_fig_p020_9.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

63 extracted references · 53 canonical work pages

  1. [1]

    Improving the generation and evaluation of synthetic data for downstream medical causal inference

    Harry Amad, Zhaozhi Qian, Dennis Frauen, Julianna Piskorz, Stefan Feuerriegel, and Mihaela van der Schaar. Improving the generation and evaluation of synthetic data for downstream medical causal inference. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2025

    work page 2025

  2. [2]

    C. T. C. Arsene and P. J. G. Lisboa. Chapter 8 - Artificial neural networks used in the survival analysis of breast cancer patients: A node-negative study. In Outcome Prediction in Cancer, pp.\ 191--239. Elsevier, 2007

    work page 2007

  3. [3]

    Johnson, Jonathan Ho, Daniel Tarlow, and Rianne van den Berg

    Jacob Austin, Daniel D. Johnson, Jonathan Ho, Daniel Tarlow, and Rianne van den Berg . Structured denoising diffusion models in discrete state-spaces. In Neural Information Processing Systems ( NeurIPS ) , 2021

    work page 2021

  4. [4]

    Peter C. Austin. Generating survival times to simulate cox proportional hazards models with time-varying covariates. Statistics in Medicine, 31 0 (29): 0 3946--3958, 2012

    work page 2012

  5. [5]

    A general machine learning framework for survival analysis

    Andreas Bender, David R \"u gamer, Fabian Scheipl, and Bernd Bischl. A general machine learning framework for survival analysis. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases , 2021

    work page 2021

  6. [6]

    Generating survival times to simulate cox proportional hazards models

    Ralf Bender, Thomas Augustin, and Maria Blettner. Generating survival times to simulate cox proportional hazards models. Statistics in Medicine, 24 0 (11): 0 1713--1723, 2005

    work page 2005

  7. [7]

    Statistics review 12: Survival analysis

    Viv Bewick, Liz Cheek, and Jonathan Ball. Statistics review 12: Survival analysis. Critical Care, 8 0 (5): 0 389--394, 2004

    work page 2004

  8. [8]

    Glenn W. Brier. Verification of forecasts expressed in terms of probability. Monthly Weather Review, 78 0 (1): 0 1--3, 1950

    1950

Show all 63 references
  1. [9]

    An overview of diffusion models: Applications , guided generation, statistical rates and optimization

    Minshuo Chen, Song Mei, Jianqing Fan, and Mengdi Wang. An overview of diffusion models: Applications , guided generation, statistical rates and optimization. arXiv preprint, arXiv:2404.07771, 2024

    2024 arXiv

  2. [10]

    XGBoost : A scalable tree boosting system

    Tianqi Chen and Carlos Guestrin. XGBoost : A scalable tree boosting system. In International Conference on Knowledge Discovery and Data Mining ( KDD ) , 2016

    2016

  3. [11]

    D. R. Cox. Regression models and life-tables. Journal of the Royal Statistical Society. Series B (Methodological), 34 0 (2): 0 187--220, 1972

    1972

  4. [12]

    Practical lessons from generating synthetic healthcare data with bayesian networks

    Juan de Benedetti , Namir Oues, Zhenchen Wang, Puja Myles, and Allan Tucker. Practical lessons from generating synthetic healthcare data with bayesian networks. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases Workshops , 2020

    2020

  5. [13]

    Diffusion models beat GANs on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat GANs on image synthesis. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2021

    2021

  6. [14]

    Hyland, and Gunnar R \"a tsch

    Crist \'o bal Esteban, Stephanie L. Hyland, and Gunnar R \"a tsch. Real-valued (medical) time series generation with recurrent conditional GANs . arXiv preprint, arXiv:1706.02633, 2017

    2017 arXiv

  7. [15]

    Fay, Erica H

    Michael P. Fay, Erica H. Brittain, and Michael A. Proschan. Pointwise confidence intervals for a survival distribution with small samples or heavy censoring. Biostatistics, 14 0 (4): 0 723--736, 2013

    2013

  8. [16]

    Orthogonal survival learners for estimating heterogeneous treatment effects from time-to-event data

    Dennis Frauen, Maresa Schr \"o der, Konstantin Hess, and Stefan Feuerriegel. Orthogonal survival learners for estimating heterogeneous treatment effects from time-to-event data. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2025

    2025

  9. [17]

    Mauro Giuffr \`e and Dennis L. Shung. Harnessing the power of synthetic data in healthcare: Innovation , application, and privacy. npj Digital Medicine, 6 0 (1): 0 1--8, 2023

    2023

  10. [18]

    Testing generated distributions in GANs to penalize mode collapse

    Yanxiang Gong, Zhiwei Xie, Mei Xie, and Xin Ma. Testing generated distributions in GANs to penalize mode collapse. In International Conference on Artificial Intelligence and Statistics ( AISTATS ) , 2024

    2024

  11. [19]

    Goodfellow, Jean Pouget-Abadie , Mehdi Mirza, Bing Xu, David Warde-Farley , Sherjil Ozair, Aaron Courville, and Yoshua Bengio

    Ian J. Goodfellow, Jean Pouget-Abadie , Mehdi Mirza, Bing Xu, David Warde-Farley , Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2014

    2014

  12. [20]

    Hammer, Kathleen E

    Scott M. Hammer, Kathleen E. Squires, Michael D. Hughes, Janet M. Grimes, Lisa M. Demeter, Judith S. Currier, Joseph J. Eron, Judith E. Feinberg, Henry H. Balfour, Lawrence R. Deyton, Jeffrey A. Chodakewitz, Margaret A. Fischl, John P. Phair, Louise Pedneault, Bach-Yen Nguyen,...

    1997

  13. [21]

    Harrell, Jr, Robert M

    Frank E. Harrell, Jr, Robert M. Califf, David B. Pryor, Kerry L. Lee, and Robert A. Rosati. Evaluating the yield of medical tests. JAMA, 247 0 (18): 0 2543--2546, 1982

    1982

  14. [22]

    Denoising diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2020

    2020

  15. [23]

    Kogalur, Eugene H

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

    2008

  16. [24]

    E. L. Kaplan and Paul Meier. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association, 53 0 (282): 0 457--481, 1958

    1958

  17. [25]

    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 Conference on Neural Information Processing Systems ( NeurIPS ) , 2022

    2022

  18. [26]

    Katzman, Uri Shaham, Alexander Cloninger, Jonathan Bates, Tingting Jiang, and Yuval Kluger

    Jared L. Katzman, Uri Shaham, Alexander Cloninger, Jonathan Bates, Tingting Jiang, and Yuval 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

  19. [27]

    Restricted mean survival time as a measure to interpret clinical trial results

    Dae Hyun Kim, Hajime Uno, and Lee-Jen Wei. Restricted mean survival time as a measure to interpret clinical trial results. JAMA Cardiology, 2 0 (11): 0 1179--1180, 2017

    2017

  20. [28]

    Kingma and Max Welling

    Diederik P. Kingma and Max Welling. Auto-encoding variational bayes. In International Conference on Learning Representations ( ICLR ) , 2013

    2013

  21. [29]

    A. V. Konstantinov, S. R. Kirpichenko, and L. V. Utkin. Generating survival interpretable trajectories and data. Doklady Mathematics, 110 0 (1): 0 S75--S86, 2024

    2024

  22. [30]

    DeepPAMM : Deep piecewise exponential additive mixed models for complex hazard structures in survival analysis

    Philipp Kopper, Simon Wiegrebe, Bernd Bischl, Andreas Bender, and David R \"u gamer. DeepPAMM : Deep piecewise exponential additive mixed models for complex hazard structures in survival analysis. In Pacific- Asia Conference on Knowledge Discovery and Data Mining ( PAKDD ) , 2022

    2022

  23. [31]

    DeepHit : A deep learning approach to survival analysis with competing risks

    Changhee Lee, William Zame, Jinsung Yoon, and Mihaela van der Schaar. DeepHit : A deep learning approach to survival analysis with competing risks. In AAAI Conference on Artificial Intelligence , 2018

    2018

  24. [32]

    DivideMix : Learning with noisy labels as semi-supervised learning

    Junnan Li, Richard Socher, and Steven C H Hoi. DivideMix : Learning with noisy labels as semi-supervised learning. In International Conference on Learning Representations ( ICLR ) , 2020

    2020

  25. [33]

    DiffPO : A causal diffusion model for learning distributions of potential outcomes

    Yuchen Ma, Valentyn Melnychuk, Jonas Schweisthal, and Stefan Feuerriegel. DiffPO : A causal diffusion model for learning distributions of potential outcomes. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2024

    2024

  26. [34]

    A Diffusion-Based Method for Learning the Multi-Outcome Distribution of Medical Treatments

    Yuchen Ma, Jonas Schweisthal, Hengrui Zhang, and Stefan Feuerriegel. A Diffusion-Based Method for Learning the Multi-Outcome Distribution of Medical Treatments . In International Conference on Knowledge Discovery and Data Mining ( KDD ) , 2025

    2025

  27. [35]

    Survival Analysis: A Practical Approach

    David Machin, Yin Bun Cheung, and Mahesh Kb Parmar. Survival Analysis: A Practical Approach . Wiley, 1st edition, 2006

    2006

  28. [36]

    Deep survival analysis: Nonparametrics and missingness

    Xenia Miscouridou, Adler Perotte, Noemie Elhadad, and Rajesh Ranganath. Deep survival analysis: Nonparametrics and missingness. In Machine Learning for Healthcare Conference , 2018

    2018

  29. [37]

    SurvivalGAN : Generating time-to-event data for survival analysis

    Alexander Norcliffe, Bogdan Cebere, Fergus Imrie, Pietro Li \'o , and Mihaela van der Schaar . SurvivalGAN : Generating time-to-event data for survival analysis. In International Conference on Artificial Intelligence and Statistics ( AISTATS ) , 2023

    2023

  30. [38]

    Normalizing flows for probabilistic modeling and inference

    George Papamakarios, Eric Nalisnick, Danilo Jimenez Rezende, Shakir Mohamed, and Balaji Lakshminarayanan. Normalizing flows for probabilistic modeling and inference. Journal of Machine Learning Research, 22 0 (57): 0 1--64, 2021

    2021

  31. [39]

    Rueda, Hans-Kristian Moen Vollan, Elena Provenzano, Helen A

    Bernard Pereira, Suet-Feung Chin, Oscar M. Rueda, Hans-Kristian Moen Vollan, Elena Provenzano, Helen A. Bardwell, Michelle Pugh, Linda Jones, Roslin Russell, Stephen-John Sammut, Dana W. Y. Tsui, Bin Liu, Sarah-Jane Dawson, Jean Abraham, Helen Northen, John F. Peden, Abhik Muk...

    2016

  32. [40]

    The effectiveness of data augmentation in image classification using deep learning

    Luis Perez and Jason Wang. The effectiveness of data augmentation in image classification using deep learning. arXiv preprint, arXiv:1712.04621, 2017

    2017 arXiv

  33. [41]

    Deep survival analysis

    Rajesh Ranganath, Adler Perotte, No \'e mie Elhadad, and David Blei. Deep survival analysis. In Machine Learning for Healthcare Conference , 2016

    2016

  34. [42]

    Patrick Royston and Mahesh K. B. Parmar. The use of restricted mean survival time to estimate the treatment effect in randomized clinical trials when the proportional hazards assumption is in doubt. Statistics in Medicine, 30 0 (19): 0 2409--2421, 2011

    2011

  35. [43]

    Chiu, Alexander M

    Subham Sekhar Sahoo, Marianne Arriola, Yair Schiff, Aaron Gokaslan, Edgar Mariano Marroquin, Justin T. Chiu, Alexander M. Rush, and Volodymyr Kuleshov. Simple and effective masked diffusion language models. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2024

    2024

  36. [44]

    Generative adversarial networks ( GANs ): Challenges , solutions, and future directions

    Divya Saxena and Jiannong Cao. Generative adversarial networks ( GANs ): Challenges , solutions, and future directions. ACM Computing Surveys, 54 0 (3): 0 63:1--63:42, 2021

    2021

  37. [45]

    Randomized 2 x 2 trial evaluating hormonal treatment and the duration of chemotherapy in node-positive breast cancer patients

    M Schumacher, G Bastert, H Bojar, K H \"u bner, M Olschewski, W Sauerbrei, C Schmoor, C Beyerle, R L Neumann, and H F Rauschecker. Randomized 2 x 2 trial evaluating hormonal treatment and the duration of chemotherapy in node-positive breast cancer patients. German breast cance...

    1994

  38. [46]

    Cancer care treatment attrition in adults: Measurement approaches and inequities in patient dropout rates -- A rapid review

    Jenny Shand, Elizabeth Stovold, Lucy Goulding, and Kate Cheema. Cancer care treatment attrition in adults: Measurement approaches and inequities in patient dropout rates -- A rapid review. BMC Cancer, 24 0 (1): 0 1345, 2024

    2024

  39. [47]

    Simplified and generalized masked diffusion for discrete data

    Jiaxin Shi, Kehang Han, Zhe Wang, Arnaud Doucet, and Michalis Titsias. Simplified and generalized masked diffusion for discrete data. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2024 a

    2024

  40. [48]

    TabDiff : A mixed-type diffusion model for tabular data generation

    Juntong Shi, Minkai Xu, Harper Hua, Hengrui Zhang, Stefano Ermon, and Jure Leskovec. TabDiff : A mixed-type diffusion model for tabular data generation. In International Conference on Learning Representations ( ICLR ) , 2024 b

    2024

  41. [49]

    A comprehensive survey of synthetic tabular data generation

    Ruxue Shi, Yili Wang, Mengnan Du, Xu Shen, Yi Chang, and Xin Wang. A comprehensive survey of synthetic tabular data generation. arXiv preprint, arXiv:2504.16506, 2025

    2025 arXiv

  42. [50]

    Deep unsupervised learning using nonequilibrium thermodynamics

    Jascha Sohl-Dickstein , Eric Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsupervised learning using nonequilibrium thermodynamics. In International Conference on Machine Learning ( ICML ) , 2015

    2015

  43. [51]

    Ladder variational autoencoders

    Casper Kaae S nderby, Tapani Raiko, Lars Maal e, S ren Kaae S nderby, and Ole Winther. Ladder variational autoencoders. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2016

    2016

  44. [52]

    Generative modeling by estimating gradients of the data distribution

    Yang Song and Stefano Ermon. Generative modeling by estimating gradients of the data distribution. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2019

    2019

  45. [53]

    Score-based generative modeling through stochastic differential equations

    Yang Song, Jascha Sohl-Dickstein , Diederik P Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations. In International Conference on Learning Representations ( ICLR ) , 2021

    2021

  46. [54]

    DECAF : Generating fair synthetic data using causally-aware generative networks

    Boris van Breugel, Trent Kyono, Jeroen Berrevoets, and Mihaela van der Schaar. DECAF : Generating fair synthetic data using causally-aware generative networks. In Conference on Neural Information Processing Systems ( NeurIPS ) , 2021

    2021

  47. [55]

    A statistical distribution function of wide applicability

    Waloddi Weibull. A statistical distribution function of wide applicability. Journal of Applied Mechanics, 18 0 (3): 0 293--297, 1951

    1951

  48. [56]

    Deep learning for survival analysis: A review

    Simon Wiegrebe, Philipp Kopper, Raphael Sonabend, Bernd Bischl, and Andreas Bender. Deep learning for survival analysis: A review. Artificial Intelligence Review, 57 0 (3): 0 65, 2024

    2024

  49. [57]

    Modeling tabular data using conditional GAN

    Lei Xu, Maria Skoularidou, Alfredo Cuesta-Infante , and Kalyan Veeramachaneni. Modeling tabular data using conditional GAN . In Conference on Neural Information Processing Systems ( NeurIPS ) , 2019

    2019

  50. [58]

    Mooney, and Bradley A

    Chao Yan, Yao Yan, Zhiyu Wan, Ziqi Zhang, Larsson Omberg, Justin Guinney, Sean D. Mooney, and Bradley A. Malin. A multifaceted benchmarking of synthetic electronic health record generation models. Nature Communications, 13 0 (1): 0 7609, 2022

    2022

  51. [59]

    Mixed-type tabular data synthesis with score-based diffusion in latent space

    Hengrui Zhang, Jiani Zhang, Balasubramaniam Srinivasan, Zhengyuan Shen, Xiao Qin, Christos Faloutsos, Huzefa Rangwala, and George Karypis. Mixed-type tabular data synthesis with score-based diffusion in latent space. In International Conference on Learning Representations ( IC...

    2024

  52. [60]

    Deep generative survival analysis: Nonparametric estimation of conditional survival function

    Xingyu Zhou, Wen Su, Changyu Liu, Yuling Jiao, Xingqiu Zhao, and Jian Huang. Deep generative survival analysis: Nonparametric estimation of conditional survival function. arXiv preprint, arXiv:2205.09633, 2022

    2022 arXiv

  53. [61]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  54. [62]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  55. [63]

    Survival metrics over different datasets (reported: mean s.d

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

    2015

Pith tools

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