Pith. sign in

REVIEW 5 major objections 6 minor 59 references

SurvUnc: A Meta-Model Based Uncertainty Quantification Framework for Survival Analysis

T0 review · 5 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read SurvUnc claims to be the first model-agnostic framework that attaches a predictive-uncertainty score to any survival model, trained only on covariates and anchor-based concordance labels.

desk verdict A genuinely new meta-learning trick for survival UQ, with an evaluation that stops at uncensored samples—worth reviewing, but the central claim needs a censored-case check. read the letter →

arxiv 2505.14803 v2 pith:NT5MFWEA submitted 2025-05-20 cs.LG cs.AIcs.ET

Yu Liu , Weiyao Tao , Tong Xia , Simon Knight , Tingting Zhu This is my paper · ORCID
classification cs.LGcs.AIcs.ET
keywords survivalanalysisuncertaintyquantificationmeta-modelanchor-basedlearningconcordancecensoringselectivepredictionout-of-domaindetection
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

SurvUnc addresses a gap: survival models predict time-to-event, but almost none of them say how much to trust a given prediction. The paper claims to fix this with a post-hoc meta-model that takes only the patient covariates and outputs an uncertainty score, without touching the survival model's architecture or parameters. Because true survival curves are never observed, the authors construct training labels from a concordance principle: among two patients, the one who dies sooner should have a lower survival probability. The meta-model learns to flag cases where the base survival model gets that ordering wrong relative to a set of 'anchor' uncensored patients. Experiments across four datasets and five survival models show the scores improve selective prediction, correlate with per-patient error, and separate in-distribution from out-of-distribution patients, outperforming dropout and ensemble baselines.

What carries the argument

The load-bearing object is the anchor-based concordance label (Eq. 3): a per-patient number between 0 and 1 that measures the fraction of worse-off anchor patients the base model ranks incorrectly. The anchors are randomly sampled uncensored patients, and the pairwise comparison is made at the shorter patient's event time. This converts the unobservable ground-truth survival curve problem into a supervised label-generation step, after which any regression-style model (the paper tests MLP and random forest) can be trained as the meta-model $U(\cdot): \mathbb{R}^d \to \mathbb{R}_{\ge 0}$. The machinery works because it quantifies exactly what clinicians care about—correct risk ordering—rather than predictive variance, which is why the paper argues it beats dropout and ensembles.

What would settle it

Train SurvUnc on a synthetic dataset with known survival times where censoring is informative (for instance, high-risk patients are more likely to be censored), then measure the Pearson correlation between SurvUnc's scores and actual prediction error on the censored test patients; if the completely-at-random censoring assumption is load-bearing, this correlation will drop toward zero, showing the scores do not generalize to the censored cases the framework is meant to cover.

Watch

Extended reading notes

Core claim

The central claim is that predictive uncertainty in survival analysis can be learned as a supervised ranking-reliability problem even though ground-truth survival curves are unobservable. For each uncensored training patient $(\boldsymbol{x}_j, t_j)$, SurvUnc draws $K$ uncensored anchors and counts the fraction of anchors with $t_j < t_{A_k}$ for which the base model assigns survival probability $S(t_j|\boldsymbol{x}_j) \ge S(t_j|\boldsymbol{x}_{A_k})$; that fraction $y^{\mathrm{meta}}_j \in [0,1]$ is the training label. A meta-model $U(\boldsymbol{x})$ trained on these labels outputs the uncertainty score at test time from covariates alone. The paper claims this is the first framework that quantifies prediction uncertainty for any survival model—neural, tree-based, or Bayesian—without modifying the model or accessing its internals, and that the resulting scores outperform MC-Dropout and deep ensembles on selective prediction, misprediction detection, and out-of-domain detection across four datasets and five base models.

Load-bearing premise

The training labels come only from uncensored patients, and the method assumes censoring is completely at random, so the uncensored anchors are representative of the whole population; if censoring is informative, the uncertainty scores may be biased precisely for the censored patients that dominate many real datasets.

Editorial extensions

If this is right

  • Existing survival models—including tree-based models like RSF that cannot use MC-Dropout—can be paired with a lightweight uncertainty scorer without retraining or parameter access.
  • Selective prediction improves: discarding the highest-uncertainty samples raises time-dependent concordance by roughly 5–40% relative to best baselines at 50% discard, depending on model and dataset, and can nearly halve the Brier score.
  • Misprediction detection shows SurvUnc's scores correlate positively with per-patient Brier error across all tested models and datasets, while MC-Dropout and ensembles frequently show zero or negative correlation, so the scores can triage which predictions need expert review.
  • On the SEER breast-cancer versus heart-disease split, SurvUnc raises out-of-domain detection AUROC by roughly 20–30% over baselines, indicating the anchor labels encode domain knowledge.
  • The meta-model choice is flexible: random forest, MLP, and tabular models all fit the same labels, so the anchor-label scheme rather than a specific architecture carries the effect.

Reading between the lines

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

  • Because the label uses only uncensored anchors, I expect the scores to degrade as the censoring rate rises; a natural fix would be inverse-probability-of-censoring weighting of the pairwise terms.
  • The framework deliberately ignores the base model's output at test time, so it cannot detect a base model that silently degrades (for example, one with corrupted weights) when the covariates look normal; adding a cheap fingerprint of the base model's predictions would be a direct ablation.
  • The same anchor-concordance scheme could extend to competing risks by comparing cause-specific survival probabilities, and to time-varying covariates by evaluating the label at multiple event horizons—the paper lists both as future work.
  • A cheap validation anyone can run on the paper's own SAC3 synthetic data is to check whether the high-uncertainty samples consistently recover the longer-survival tail shown in Figure 4; if changing the anchor set flips the ordering of high- and low-uncertainty samples, the stability of the labels with respect to the anchor sample is the thing to test.
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 / 6 minor

Summary. The paper proposes SurvUnc, a post-hoc, model-agnostic uncertainty quantification framework for survival models. The method trains a lightweight meta-model (random forest or MLP) on the base model's covariates alone, using labels derived from an anchor-based concordance-error measure: for each uncensored training sample, the label is the proportion of anchors with longer observed event times for which the base model incorrectly ranks the survival probabilities. The trained meta-model then outputs an uncertainty score for a new sample using only its covariates. The authors evaluate SurvUnc across four datasets and five survival models on three tasks: selective prediction (discarding high-uncertainty samples and recomputing C_td and IBS), misprediction detection (correlation between uncertainty scores and per-sample IBS), and out-of-domain detection (SEER-BC vs. SEER-HD). They report that SurvUnc outperforms MC-Dropout and Deep Ensemble baselines, and also outperforms the native Bayesian uncertainty of BNNSurv.

Significance. If the empirical claims hold, the framework would be practically useful: it is truly post-hoc and model-agnostic, requires no access to base-model parameters, and the implementation is provided. The experiments cover a reasonable spread of survival models (DeepSurv, DeepHit, DSM, RSF, BNNSurv) and datasets, and the authors report bootstrap resampling and Wilcoxon signed-rank tests. However, the paper's central claim that it quantifies uncertainty for 'any survival model' is not supported by the evaluation as currently presented: selective prediction lacks a random-discarding baseline, all main analyses are restricted to uncensored test samples, and the label definition in Eq. (3) is undefined for a non-negligible subset of training samples. These gaps are load-bearing because they directly affect the interpretation of every reported improvement. The work is a reasonable step toward standardized UQ evaluation for survival analysis, but the manuscript needs additional experiments and clarifications before the central claims can be accepted.

major comments (5)
  1. [Section 4.2, Table 2] The selective prediction experiment lacks a random discarding baseline. The paper reports C_td after removing 10%, 30%, and 50% of test samples according to an uncertainty score, but never reports C_td for a random subset of the same size, nor the C_td at 0% discarding for each base model. Without this baseline, the large relative improvements (e.g., 32.5% on FLCHAIN for DeepSurv) cannot be attributed to the uncertainty ordering: C_td on a smaller subset can change mechanically, and the selected subset may differ in composition. Please add a random-discarding curve (or an oracle upper bound) for every model-dataset combination, and report the no-discard C_td so that the improvement is interpretable.
  2. [Section 3.3, Eq. (3)] The denominator in Eq. (3), sum over k of 1[t_j < t_A_k], is zero whenever no sampled anchor has an observed event time greater than t_j. This is likely to occur for the longest-surviving uncensored training samples, since K=50 anchors are drawn at random. The paper does not state whether such samples are excluded from the meta-training set, assigned a default label, or handled in some other way. This is a real specification gap: those samples are exactly the long-term survival cases, and their exclusion or arbitrary labeling changes the meta-model's training distribution. Please define the rule explicitly and report how many training samples have zero denominator in each dataset.
  3. [Section 4.2, Section 4.3, Table 1] All selective prediction and misprediction detection results are computed on uncensored test samples only, as stated in Section 4.2: 'we discard only uncensored samples, and subsequent analyses also focus on uncensored ones.' On datasets with majority censoring (FLCHAIN 69.9%, SEER-BC 73.7%), this means the reported results cover only a minority of test cases, and there is no evidence that SurvUnc provides useful uncertainty scores for censored predictions. The framework's central claim is about any survival model on censored data, so this is a major omission. Please evaluate on the full test set using censoring-adjusted metrics (e.g., IPCW-weighted IBS for misprediction, and C_td computed with all comparable pairs, or a separate reported analysis for censored samples).
  4. [Section 3.3, Section 4.4] The CCR (censoring completely at random) assumption is load-bearing for label construction, because anchors and labeled training instances are drawn exclusively from uncensored samples. The paper invokes the assumption without validation, but informative censoring (e.g., sicker patients lost to follow-up) would bias the concordance-error labels toward the uncensored subpopulation. The fact that the evaluation is also restricted to uncensored test samples masks this problem. Please add a sensitivity analysis: for instance, simulate a survival dataset under informative censoring and show whether SurvUnc's scores still rank censored samples correctly, or compare the covariate distributions of uncensored anchors with those of censored samples in the real datasets.
  5. [Section 4.4, Table 4] The absolute OOD detection performance is low: AUROC values are about 0.62-0.67 and AUPRC values about 0.58-0.63. While these numbers improve over baselines near 0.5, an AUROC of roughly 0.65 is generally considered a weak detector, and the text's claim of a 'significant performance improvement' and 'notable enhancement' overstates the practical capability. Please report confidence intervals for the OOD metrics and discuss whether such AUROC levels would be acceptable for the intended high-stakes deployment scenarios.
minor comments (6)
  1. [Section 4.4] The heading 'ODD detection' appears to be a typo for 'OOD detection'. Also, in the caption of Figure 5, 'SurvUnC-RF' should be 'SurvUnc-RF'.
  2. [Appendix B.5, Table 12] The entry '0701' for TabPFN at 30% discarding appears to be missing the leading '0' and decimal point; it should likely be '0.701'.
  3. [Section 4.2] In the sentence discussing BNNSurv, 'BNNSurve' appears to be a typo for 'BNNSurv'.
  4. [Section 4.1.5 and Table 2] The statement 'p-value<0.001 is true for all results without *' is ambiguous: it does not specify the null hypothesis or the comparison (e.g., SurvUnc-RF vs. best baseline, or vs. MC-Dropout), and it is unclear whether the test is the Wilcoxon signed-rank test across bootstrap resamples or across repeated data splits. Please clarify the statistical test procedure.
  5. [Section 3.3, Eq. (3)] The indicator 1[S(t_j|x_j) >= S(t_j|x_A_k)] counts ties as 'incorrectly ordered pairs', but in survival analysis a tie in predicted survival probability is not necessarily an error. Consider using a strict inequality or a tie-handling convention consistent with the C_td definition.
  6. [Section 1] The claim that SurvUnc is 'the first model-agnostic framework capable of quantifying the uncertainty of predictions in any survival model' should be moderated or supported by a more thorough comparison with existing post-hoc UQ methods for survival analysis, since several Bayesian and ensemble approaches are model-agnostic in practice.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the meta-model is trained on observed-event-time concordance labels and evaluated on held-out samples, with independent OOD and IBS-based checks.

full rationale

SurvUnc's derivation chain is self-contained. The meta-model U(x) is trained on labels y_meta computed in Eq. (3) from observed event times, the base model's survival probabilities, and randomly selected uncensored anchors; the training instances are uncensored samples from the training set. At inference, U(x) receives only covariates and outputs an uncertainty score for a new sample. The headline evaluations are not the same quantity as the training labels: selective prediction uses C_td and IBS computed on held-out test samples after discarding high-uncertainty items, misprediction detection correlates the predicted scores with held-out IBS values, and OOD detection is performed on a distinct disease cohort (SEER-HD). The metric alignment between the concordance-derived training label and C_td is real, but it is not circular by construction because the meta-model must generalize from covariates to unseen samples; it is not fitting the test metric. The SAC3 ground-truth survival curve case study also provides an external check. Self-citations in the paper (e.g., [16], [53], [55]) appear in related-work and future-work contexts and are not load-bearing for the central claim; there is no imported uniqueness theorem or fitted parameter renamed as a prediction. The CCR assumption and uncensored-only evaluation noted in Section 3.3 and Section 4.2 are validity concerns for real-world censored data, but they concern correctness of generalization, not circularity of the method.

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

No new physical or mathematical entities are introduced; the meta-model is a standard learner. The free parameters are the anchor count and the meta-model's fixed hyperparameters. The key unproven inputs are the random-censoring assumption, the concordance proxy, and the covariate-only design.

free parameters (2)
  • Anchor count K = 50
    Hand-chosen in Section 4.1.5; robustness is checked by varying K, so the central results do not depend on a carefully tuned optimum, but K remains a free design choice.
  • Meta-model hyperparameters (RF: n_estimators=100, min_samples_leaf=5, min_samples_split=10; MLP: lr=0.001, hidden… = Fixed values reported in Section 4.1.5
    Fixed without dataset-specific tuning; these choices affect the meta-model fit, but robustness across two architectures is shown.
assumptions (3)
  • domain assumption Censoring occurs completely at random.
    Invoked in Section 3.3 before Eq. (3) to justify using uncensored training samples as anchors and labels.
  • ad hoc to paper The concordance-error rate in Eq. (3) is a valid proxy for predictive uncertainty.
    The paper asserts this proxy without proving it equals expected error; the SAC3 ground-truth validation gives partial empirical support.
  • domain assumption Uncertainty can be predicted from covariates alone at test time, without the base model's output.
    The usage u_i = U(x_i) in Algorithm 1 assumes features contain enough signal about prediction difficulty; this limits expressiveness for sample-specific errors.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SurvUnc: A Meta-Model Based Uncertainty Quantification Framework for Survival Analysis." pith.science (2026). https://pith.science/paper/NT5MFWEA

@misc{pith2026250514803,
  author       = {Pith},
  title        = {Pith review of: SurvUnc: A Meta-Model Based Uncertainty Quantification Framework for Survival Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NT5MFWEA}},
  note         = {Machine review of arXiv:2505.14803}
}
read the original abstract

Survival analysis, which estimates the probability of event occurrence over time from censored data, is fundamental in numerous real-world applications, particularly in high-stakes domains such as healthcare and risk assessment. Despite advances in numerous survival models, quantifying the uncertainty of predictions from these models remains underexplored and challenging. The lack of reliable uncertainty quantification limits the interpretability and trustworthiness of survival models, hindering their adoption in clinical decision-making and other sensitive applications. To bridge this gap, in this work, we introduce SurvUnc, a novel meta-model based framework for post-hoc uncertainty quantification for survival models. SurvUnc introduces an anchor-based learning strategy that integrates concordance knowledge into meta-model optimization, leveraging pairwise ranking performance to estimate uncertainty effectively. Notably, our framework is model-agnostic, ensuring compatibility with any survival model without requiring modifications to its architecture or access to its internal parameters. Especially, we design a comprehensive evaluation pipeline tailored to this critical yet overlooked problem. Through extensive experiments on four publicly available benchmarking datasets and five representative survival models, we demonstrate the superiority of SurvUnc across multiple evaluation scenarios, including selective prediction, misprediction detection, and out-of-domain detection. Our results highlight the effectiveness of SurvUnc in enhancing model interpretability and reliability, paving the way for more trustworthy survival predictions in real-world applications.

Figures

Figures reproduced from arXiv: 2505.14803 by the authors.

Figure 1
Figure 1. Illustration of our proposed framework SurvUnc. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. 𝐶 td of four survival models of (a) DeepSurv, (b) DeepHit, (c) DSM and (d) RSF on SEER-BC dataset with different percentages of samples discarded according to uncertainty scores from different UQ methods. A consistent upward trend is expected as the percentage of discarded samples increases. Error bars are omitted for better visualization [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Predicted uncertainty scores versus IBSs from [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Survival curve comparison of high-uncertainty [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Uncertainty score distribution comparison of Deep [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Varying anchor number 𝐾 on selective prediction performance of SurvUnc-RF with DeepSurv on SEER-BC. 5 Conclusion In this paper, we address the novel challenge of uncertainty quan￾tification in survival analysis and propose SurvUnc, a meta-model based framework that inc…
Figure 7
Figure 7. Figure 7: IBS of four survival models of (a) DeepSurv, (b) DeepHit, (c) DSM and (d) RSF on SEER-BC dataset with different [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: Predicted uncertainty scores versus IBSs from DeepSurv quantified by MC-Dropout across samples on (a) FLCHAIN, [PITH_FULL_IMAGE:figures/full_fig_p015_8.png]
Figure 9
Figure 9. Figure 9: Predicted uncertainty scores versus IBSs from DeepSurv quantified by Ensemble across samples on (a) FLCHAIN, (b) [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Varying anchor number 𝐾 on selective prediction performance (a) IBS: SurvUnc-RF, (b) 𝐶 td: SurvUnc-MLP and (c) IBS: SurvUnc-MLP with DeepSurv on SEER-BC [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 56 canonical work pages

  1. [1]

    Ahmed M Alaa and Mihaela van der Schaar. 2017. Deep Multi-task Gaussian Processes for Survival Analysis with Competing Risks. In Advances in Neural Information Processing Systems. 2326–2334

    work page 2017

  2. [2]

    L Antolini, P Boracchi, and E Biganzoli. 2005. A Time-dependent Discrimination Index for Survival Data. Statistics in Medicine 24, 24 (2005), 3927–3944

    work page 2005

  3. [3]

    Sercan Ö Arik and Tomas Pfister. 2021. Tabnet: Attentive Interpretable Tabular Learning. In Proceedings of the AAAI conference on Artificial Intelligence , Vol. 35. 6679–6687

    work page 2021

  4. [4]

    Tongfei Chen, Jirí Navrátil, Vijay Iyengar, and Karthikeyan Shanmugam. 2019. Confidence Scoring Using Whitebox Meta-models with Linear Classifier Probes. In International Conference on Artificial Intelligence and Statistics . PMLR, 1467– 1475

    work page 2019

  5. [5]

    Alfred F Connors, Neal V Dawson, Norman A Desbiens, William J Fulkerson, Lee Goldman, William A Knaus, Joanne Lynn, Robert K Oye, Frank E Harrell, Russell S Phillips, et al. 1995. A Controlled Trial to Improve Care for Seriously Ill Hospitalized Patients: The Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT). Jou...

    work page 1995

  6. [6]

    David R Cox. 1972. Regression Models and Life-tables. Journal of the Royal Statistical Society: Series B (Methodological) 34, 2 (1972), 187–202

    work page 1972

  7. [7]

    Chang Cui, Yongqiang Tang, and Wensheng Zhang. 2024. Deep Survival Analysis With Latent Clustering and Contrastive Learning. IEEE Journal of Biomedical and Health Informatics 28, 5 (May 2024), 3090–3101

    work page 2024

  8. [8]

    Tanujit Dey, Stuart R Lipsitz, Zara Cooper, Quoc-Dien Trinh, Martin Krzywinski, and Naomi Altman. 2022. Survival Analysis—Time-to-event Data and Censoring. Nature Methods 19, 8 (2022), 903

    work page 2022

Show all 59 references
  1. [9]

    Angela Dispenzieri, Jerry A Katzmann, Robert A Kyle, Dirk R Larson, Terry M Therneau, Colin L Colby, Raynell J Clark, Graham P Mead, Shaji Kumar, L Joseph Melton, et al. 2012. Use of Nonclonal Serum Immunoglobulin Free Light Chains to Predict Overall Survival in the General Po...

    2012

  2. [10]

    Benjamin Elder, Matthew Arnold, Anupama Murthi, and Jiří Navrátil. 2021. Learn- ing Prediction Intervals for Model Performance. In Proceedings of the AAAI Con- ference on Artificial Intelligence, Vol. 35. 7305–7313

    2021

  3. [11]

    Fahimeh Fakour, Ali Mosleh, and Ramin Ramezani. 2024. A Structured Review of Literature on Uncertainty in Machine Learning & Deep Learning. arXiv preprint arXiv:2406.00332 (2024)

    2024 arXiv

  4. [12]

    David Faraggi and Richard Simon. 1995. A Neural Network Model for Survival Data. Statistics in Medicine 14, 1 (1995), 73–82

    1995

  5. [13]

    Tamara Fernández, Nicolás Rivera, and Yee Whye Teh. 2016. Gaussian Processes for Survival Analysis. In Advances in Neural Information Processing Systems . 5021–5029

    2016

  6. [14]

    Yarin Gal and Zoubin Ghahramani. 2016. Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning. In International Conference on Machine Learning. PMLR, 1050–1059

    2016

  7. [15]

    Jakob Gawlikowski, Cedrique Rovile Njieutcheu Tassi, Mohsin Ali, Jongseok Lee, Matthias Humt, Jianxiang Feng, Anna Kruspe, Rudolph Triebel, Peter Jung, Ribana Roscher, et al. 2023. A Survey of Uncertainty in Deep Neural Networks. Artificial Intelligence Review 56 (2023), 1513–1589

    2023

  8. [16]

    Xiao Gu, Yu Liu, Zaineb Mohsin, Jonathan Bedford, Anshul Thakur, Peter Watkin- son, Lei Clifton, Tingting Zhu, and David Clifton. 2025. Are Time Series Founda- tion Models Ready for Vital Sign Forecasting in Healthcare?. InMachine Learning for Health (ML4H). PMLR, 401–419

    2025

  9. [17]

    FE Harrell. 1982. Evaluating the Yield of Medical Tests. Journal of the American Medical Association 247, 18 (1982), 2543–2546. KDD ’25, August 3–7, 2025, Toronto, ON, Canada Yu Liu, Weiyao Tao, Tong Xia, Simon Knight, and Tingting Zhu

    1982

  10. [18]

    Wenchong He and Zhe Jiang. 2023. A Survey on Uncertainty Quantification Methods for Deep Neural Networks: An Uncertainty Source Perspective. arXiv preprint arXiv:2302.13425 (2023)

    2023

  11. [19]

    Noah Hollmann, Samuel Müller, Lennart Purucker, Arjun Krishnakumar, Max Körfer, Shi Bin Hoo, Robin Tibor Schirrmeister, and Frank Hutter. 2025. Accurate Predictions on Small Data with A Tabular Foundation Model. Nature 637, 8045 (2025), 319–326

    2025

  12. [20]

    Eyke Hüllermeier and Willem Waegeman. 2021. Aleatoric and Epistemic Uncer- tainty in Machine Learning: An Introduction to Concepts and Methods. Machine Learning 110, 3 (2021), 457–506

    2021

  13. [21]

    National Cancer Institute. 2024. SEER Incidence Data, 1975-2021. https://seer. cancer.gov/data/

    2024

  14. [22]

    Kogalur, Eugene H

    Hemant Ishwaran, Udaya B. Kogalur, Eugene H. Blackstone, and Michael S. Lauer

  15. [23]

    Vincent Jeanselme, Nikita Agarwal, and Chen Wang. 2024. Review of Language Models for Survival Analysis. In AAAI 2024 Spring Symposium on Clinical Foun- dation Models

    2024

  16. [24]

    Edward L Kaplan and Paul Meier. 1958. Nonparametric Estimation From Incom- plete Observations. J. Amer. Statist. Assoc. 53, 282 (1958), 457–481

    1958

  17. [25]

    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. 2018. DeepSurv: Personalized Treatment Recommender System Using a Cox Proportional Hazards Deep Neural Network. BMC Medical Research Methodology 18, 1 (Feb. 2018)

    2018

  18. [26]

    Håvard Kvamme and Ørnulf Borgan. 2021. Continuous and Discrete-time Survival Prediction with Neural Networks. Lifetime Data Analysis 27, 4 (2021), 710–736

    2021

  19. [27]

    Håvard Kvamme and Ørnulf Borgan. 2023. The Brier Score Under Administrative Censoring: Problems and A Solution. Journal of Machine Learning Research 24, 2 (2023), 1–26

    2023

  20. [28]

    Håvard Kvamme, Ørnulf Borgan, and Ida Scheel. 2019. Time-to-Event Prediction with Neural Networks and Cox Regression. Journal of Machine Learning Research 20, 129 (2019), 1–30

    2019

  21. [29]

    Balaji Lakshminarayanan, Alexander Pritzel, and Charles Blundell. 2017. Sim- ple and Scalable Predictive Uncertainty Estimation using Deep Ensembles. In Advances in Neural Information Processing Systems

    2017

  22. [30]

    Changhee Lee, William Zame, Jinsung Yoon, and Mihaela Van Der Schaar. 2018. DeepHit: A Deep Learning Approach to Survival Analysis With Competing Risks. In Proceedings of the AAAI Conference on Artificial Intelligence . 2314–2321

    2018

  23. [31]

    Christian Marius Lillelund, Martin Magris, and Christian Fischer Pedersen. 2023. Uncertainty Estimation in Deep Bayesian Survival Models. In IEEE EMBS Inter- national Conference on Biomedical and Health Informatics . IEEE, 1–4

    2023

  24. [32]

    Christian Marius Lillelund, Martin Magris, and Christian Fischer Pedersen. 2024. Efficient Training of Probabilistic Neural Networks for Survival Analysis. IEEE Journal of Biomedical and Health Informatics (2024)

    2024

  25. [33]

    Tie-Yan Liu et al. 2009. Learning to Rank for Information Retrieval. Foundations and Trends® in Information Retrieval 3, 3 (2009), 225–331

    2009

  26. [34]

    Christos Louizos and Max Welling. 2017. Multiplicative Normalizing Flows for Variational Bayesian Neural Networks. In International Conference on Machine Learning. PMLR, 2218–2227

    2017

  27. [35]

    Hrushikesh Loya, Pranav Poduval, Deepak Anand, Neeraj Kumar, and Amit Sethi

  28. [36]

    Yingzhou Lu, Tianyi Chen, Nan Hao, Capucine Van Rechem, Jintai Chen, and Tianfan Fu. 2024. Uncertainty Quantification and Interpretability for Clinical Trial Approval Prediction. Health Data Science 4 (Jan. 2024)

    2024

  29. [37]

    Andrey Malinin and Mark Gales. 2018. Predictive Uncertainty Estimation via Prior Networks. Advances in Neural Information Processing Systems 31 (2018)

    2018

  30. [38]

    Xenia Miscouridou, Adler Perotte, Noémie Elhadad, and Rajesh Ranganath. 2018. Deep Survival Analysis: Nonparametrics and Missingness. In Machine Learning for Healthcare Conference. PMLR, 244–256

    2018

  31. [39]

    Chirag Nagpal, Xinyu Li, and Artur Dubrawski. 2021. Deep Survival Machines: Fully Parametric Survival Regression and Representation Learning for Censored Data With Competing Risks. IEEE Journal of Biomedical and Health Informatics 25, 8 (Aug. 2021), 3163–3175

    2021

  32. [40]

    Konstantin Posch, Jan Steinbrener, and Jürgen Pilz. 2019. Variational Infer- ence to Measure Model Uncertainty in Deep Neural Networks. arXiv preprint arXiv:1902.10189 (2019)

    2019 arXiv

  33. [41]

    Rajesh Ranganath, Adler Perotte, Noémie Elhadad, and David Blei. 2016. Deep Survival Analysis. In Machine Learning for Healthcare Conference . PMLR

    2016

  34. [42]

    Kan Ren, Jiarui Qin, Lei Zheng, Zhengyu Yang, Weinan Zhang, Lin Qiu, and Yong Yu. 2019. Deep Recurrent Survival Analysis. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 33. 4798–4805

    2019

  35. [43]

    Murat Sensoy, Lance Kaplan, and Melih Kandemir. 2018. Evidential Deep Learn- ing to Quantify Classification Uncertainty. In Advances in Neural Information Processing Systems. 3183–3193

    2018

  36. [44]

    Maohao Shen, Yuheng Bu, Prasanna Sattigeri, Soumya Ghosh, Subhro Das, and Gregory Wornell. 2023. Post-hoc Uncertainty Learning Using a Dirichlet Meta- Model. In Proceedings of the AAAI Conference on Artificial Intelligence. 9772–9781

    2023

  37. [45]

    Ethan Steinberg, Jason Alan Fries, Yizhe Xu, and Nigam Shah. 2024. MOTOR: A Time-to-Event Foundation Model For Structured Medical Records. InInternational Conference on Learning Representations

    2024

  38. [46]

    Zhaohong Sun, Wei Dong, Jinlong Shi, Kunlun He, and Zhengxing Huang. 2021. Attention-Based Deep Recurrent Model for Survival Prediction.ACM Transactions on Computing for Healthcare 2, 4 (Sept. 2021), 1–18

    2021

  39. [47]

    Dennis Ulmer, Christian Hardmeier, and Jes Frellsen. 2023. Prior and Posterior Networks: A Survey on Evidential Deep Learning Methods For Uncertainty Estimation. Transactions on Machine Learning Research (2023)

    2023

  40. [48]

    Joost Van Amersfoort, Lewis Smith, Yee Whye Teh, and Yarin Gal. 2020. Un- certainty Estimation Using A Single Deep Deterministic Neural Network. In International Conference on Machine Learning . PMLR, 9690–9700

    2020

  41. [49]

    Ping Wang, Yan Li, and Chandan K Reddy. 2019. Machine Learning for Survival Analysis: A Survey. Comput. Surveys 51, 6 (2019), 1–36

    2019

  42. [50]

    Zifeng Wang and Jimeng Sun. 2022. SurvTRACE: Transformers for Survival Anal- ysis with Competing Events. In ACM International Conference on Bioinformatics, Computational Biology and Health Informatics . 1–9

    2022

  43. [51]

    Florian Wenzel, Jasper Snoek, Dustin Tran, and Rodolphe Jenatton. 2020. Hyper- parameter Ensembles for Robustness and Uncertainty Quantification. Advances in Neural Information Processing Systems 33 (2020), 6514–6527

    2020

  44. [52]

    Simon Wiegrebe, Philipp Kopper, Raphael Sonabend, Bernd Bischl, and Andreas Bender. 2024. Deep Learning for Survival Analysis: A Review. Artificial Intelli- gence Review 57, 3 (2024), 65

    2024

  45. [53]

    Tong Xia, Ting Dang, Jing Han, Lorena Qendro, and Cecilia Mascolo. 2024. Uncertainty-aware Health Diagnostics via Class-balanced Evidential Deep Learn- ing. IEEE Journal of Biomedical and Health Informatics 28, 11 (2024), 6417–6428

    2024

  46. [54]

    Xuanlong Yu, Gianni Franchi, Jindong Gu, and Emanuel Aldea. 2024. Discretization-Induced Dirichlet Posterior for Robust Uncertainty Quantification on Regression. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 38. 6835–6843

    2024

  47. [55]

    Yuwei Zhang, Tong Xia, Abhirup Ghosh, and Cecilia Mascolo. 2023. Uncertainty Quantification in Federated Learning for Heterogeneous Health Data. In KDD Workshop on Federated Learning for Distributed Data Mining

    2023

  48. [56]

    Qixian Zhong, Jonas Mueller, and Jane-Ling Wang. 2021. Deep Extended Haz- ard Models for Survival Analysis. In Advances in Neural Information Processing Systems. 15111–15124. A Experiment Setup A.1 Survival Model Details To quantify the uncertainty for survival models, we firs...

    2021

  49. [59]

    While TabPFN achieves marginally better results, the improvements are minimal considering the substantial computational and imple- mentation overhead

    as meta models with selective prediction, as shown in Table|12. While TabPFN achieves marginally better results, the improvements are minimal considering the substantial computational and imple- mentation overhead. Table 12:𝐶td of DeepSurv under different discarding percent- a...

    2025

  50. [2008]

    The Annals of Applied Statistics 2, 3 (Sept

    Random Survival Forests. The Annals of Applied Statistics 2, 3 (Sept. 2008). https://doi.org/10.1214/08-aoas169

    2008 doi

  51. [2020]

    In ICLR Workshop on AI for Affordable Healthcare

    Uncertainty Estimation in Cancer Survival Prediction. In ICLR Workshop on AI for Affordable Healthcare

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.