Pith. sign in

REVIEW 3 major objections 2 minor 1 cited by

Multiply Robust Conformal Risk Control with Coarsened Data

T0 review · 3 major / 2 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read This paper claims that conformal prediction can still give distribution-free valid prediction intervals when training data are coarsened—missing outcomes, missing covariates, or censored time-to-event outcomes—by deriving the efficient infl

desk verdict Promising but unverifiable from the abstract alone; the 'distribution-free' claim needs qualification about the conformal step's exchangeability. read the letter →

arxiv 2508.15489 v1 pith:OUKM67AI submitted 2025-08-21 math.ST stat.MEstat.MLstat.TH

Manit Paul , Arun Kumar Kuchibhotla , Eric J. Tchetgen Tchetgen This is my paper
classification math.STstat.MEstat.MLstat.TH MSC 62G0562G1562N01
keywords conformalpredictioncoarseneddatamissingcensoredoutcomessemiparametricefficiencyefficientinfluencefunctiondistribution-freeinferenceriskcontrol
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

The paper tackles a practical gap: conformal prediction usually assumes the training sample is fully observed, but real data often come with missing outcomes, missing covariates, or censored follow-up times. It claims that valid distribution-free prediction intervals can still be constructed from such coarsened data, provided the coarsening is modelled semiparametrically. The key move is to derive the efficient influence function of the quantile being predicted and use it as a bias-correcting score inside a conformal risk control wrapper. If correct, this unifies missing-data adjustment and distribution-free predictive inference, and it extends naturally to covariate shift and multiply robust prediction sets. A sympathetic reader would care because the method promises valid intervals without assuming a parametric outcome model or requiring complete-case analysis.

What carries the argument

The efficient influence function (EIF) of the target quantile under a coarsened-data semiparametric model, combined with a conformal risk-control wrapper. The EIF is the bias-correcting score that accounts for missingness and censoring; the conformal wrapper converts the corrected scores into finite-sample risk control guarantees.

What would settle it

Simulate coarsened training data where the probability of observing the outcome depends on the unobserved outcome itself, so missingness is not coarsening at random, and check the empirical coverage of the proposed intervals across many replications; if coverage falls materially below the nominal level, the central claim is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that one can obtain distribution-free prediction regions with guaranteed risk control even when the training sample is coarsened, meaning some outcomes are missing, some covariates are missing, or time-to-event outcomes are censored. The construction derives the efficient influence function of the quantile being predicted under a semiparametric model for the coarsened data, then feeds this correction into a new conformal risk control procedure. The result is prediction intervals that keep their nominal coverage without distributional assumptions on the outcome, while the semiparametric formulation allows the nuisance functions to be estimated by flexible machine

Load-bearing premise

The coarsening mechanism is correctly specified as a semiparametric model under which the efficient influence function is well-defined; if missingness is informative or the model is misspecified, the adjustment can be biased and the guarantee can fail.

Editorial extensions

If this is right

  • Prediction intervals can be built directly from training sets with partially observed outcomes and covariates while keeping the stated coverage guarantee.
  • Time-to-event training data with censored outcomes no longer force the user to discard or impute; the censoring is folded into the adjustment.
  • Because nuisance functions can be learned by flexible machine learning, the method scales to high-dimensional or complex data without parametric outcome models.
  • Under covariate shift, the same construction yields stronger coverage than standard conformal intervals.
  • In monotone missingness settings, the framework produces multiply robust prediction sets that remain valid if some of several candidate models are misspecified.

Reading between the lines

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

  • If the construction is as general as it appears, the same EIF-plus-risk-control recipe should extend to other target functionals such as conditional quantiles or expected loss, not just marginal quantiles.
  • A practical litmus test would be a survival-analysis benchmark with heavy administrative censoring: intervals should stay calibrated even when censoring rates vary across covariate groups.
  • The multiply robust property suggests an ensemble interpretation: one could deliberately fit several candidate coarsening models and let the conformal step choose which correction to trust, potentially improving robustness beyond the paper's stated scenarios.
  • The framework's dependence on the EIF implies that its coverage guarantee is only as good as the semiparametric model; users who cannot justify coarsening at random should expect the method to degrade in a measurable, quantifiable way.
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

3 major / 2 minor

Summary. The paper proposes a general framework for conformal prediction with coarsened training data, in which the outcome or some covariates may be missing or censored. The central idea is to derive the efficient influence function (EIF) of the quantile of the outcome under a semiparametric coarsening model and to combine the resulting adjusted scores with a conformal risk control procedure. The abstract claims distribution-free valid prediction regions, improved coverage under covariate shift, multiply robust prediction sets in monotone missingness scenarios, and supporting simulation studies. However, the submitted full text is largely an unreadable encoded placeholder: no equations, theorem statements, proofs, or numeric results are legible, so the claims cannot currently be audited.

Significance. If the results are correct, the paper addresses a practically important gap in conformal prediction: handling missing or censored training data without giving up finite-sample distribution-free guarantees. Deriving EIFs for the target quantile is a principled way to incorporate flexible nuisance estimators, and the claimed multiply robust and covariate-shift extensions would be valuable additions to the conformal inference toolbox. The paper deserves credit for identifying a real limitation of standard conformal methods and for proposing a semiparametrically grounded solution. That said, the significance can only be assessed after the manuscript is made readable and after the key theoretical claim about the conformal step is stated with precise assumptions.

major comments (3)
  1. [Full text (all sections)] The full text is not readable: mathematical expressions, equations, and proofs appear as mis-encoded placeholders, and no theorem statements or simulation tables are legible. The abstract alone is insufficient to establish the correctness of the proposed construction. A complete, properly rendered manuscript with numbered assumptions, theorem statements, and proofs is required before the central claim can be evaluated.
  2. [Abstract, conformal risk control step] The abstract states that prediction regions are 'distribution-free valid' but does not state whether the guarantee is exact finite-sample, asymptotic after cross-fitting, or first-order asymptotic. The EIF is a first-order correction for the target quantile; it does not imply that individual pseudo-scores are exchangeable with the test score, which is what exact conformal risk control requires. If the pseudo-scores only match the target quantile in expectation, the tail behavior controlling miscoverage may not be preserved. The authors should either prove a (weighted) exchangeability property for their construction or explicitly state the weaker guarantee they obtain.
  3. [Abstract, semiparametric model assumptions] The abstract says the EIF is derived 'under a given semiparametric model for the coarsened data' but does not list the identifying conditions: coarsening at random, positivity of the missingness propensity, correct specification of the coarsening model, and rate conditions on the nuisance estimators. If the coarsening model is misspecified or the propensity can be arbitrarily close to zero, the EIF-based adjustment is biased and the 'distribution-free' property with respect to the outcome may no longer hold. These assumptions and regularity conditions must be stated precisely.
minor comments (2)
  1. [Full text] The document contains numerous uninterpretable character sequences, making it impossible to identify equation numbers, section numbers, or even the bibliography. A clean, correctly encoded PDF is a prerequisite for any meaningful review.
  2. [Abstract] The phrase 'multiply robust' is used without a definition in the abstract; it should be clarified whether this means robustness to misspecification of one of several nuisance models, and if so, which ones.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity identifiable; no equation-level reduction demonstrated.

full rationale

The available manuscript text is heavily corrupted by extraction artifacts, so the only clear derivation-chain statement is the abstract: the paper derives the efficient influence function (EIF) of the target quantile under a semiparametric coarsening model and then combines this with a conformal risk control procedure. On its face, this is a two-stage construction: the EIF is derived from the model, and the conformal threshold is selected to control a risk. There is no visible equation or passage showing that the conformal threshold is defined as the EIF-based quantile estimate, that a fitted parameter is renamed as a prediction, or that a central premise is imported from a self-citation. The skeptical concern that EIF-adjusted pseudo-scores may not be exchangeable with the test score is a validity/correctness objection, not a circularity objection, and the instructions explicitly require exhibiting a specific reduction (Eq. X = Eq. Y by construction) before flagging circularity. Because no such reduction can be located in the readable portions of the paper, the appropriate finding is no significant circularity.

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

Only the abstract is readable. No fitted constants are visible; the unspecified nuisance functions and the semiparametric coarsening model are the main inputs. The listed axioms are the minimal assumptions the abstract itself names.

assumptions (3)
  • domain assumption The coarsening mechanism follows a correctly specified semiparametric model, plausibly coarsening at random.
    The abstract premises the EIF derivation on a given semiparametric model for the coarsened data. If this model is misspecified, the efficient influence function is not valid for the actual data-generating process.
  • domain assumption Standard conformal exchangeability and risk-control assumptions hold, with any covariate shift handled by the paper's stated conditions.
    Conformal risk control guarantees distribution-free validity only under exchangeability or an explicit shift model. The abstract mentions covariate shift but does not state the exact conditions.
  • standard math Regularity conditions for pathwise differentiability and for nuisance function estimation rates hold.
    Efficient influence function derivations in semiparametric theory require such technical conditions. The abstract does not list them, and the readable text does not allow checking them.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multiply Robust Conformal Risk Control with Coarsened Data." pith.science (2026). https://pith.science/paper/OUKM67AI

@misc{pith2026250815489,
  author       = {Pith},
  title        = {Pith review of: Multiply Robust Conformal Risk Control with Coarsened Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OUKM67AI}},
  note         = {Machine review of arXiv:2508.15489}
}
read the original abstract

Conformal Prediction (CP) has recently received a tremendous amount of interest, leading to a wide range of new theoretical and methodological results for predictive inference with formal theoretical guarantees. However, the vast majority of CP methods assume that all units in the training data have fully observed data on both the outcome and covariates of primary interest, an assumption that rarely holds in practice. In reality, training data are often missing the outcome, a subset of covariates, or both on some units. In addition, time-to-event outcomes in the training set may be censored due to dropout or administrative end-of-follow-up. Accurately accounting for such coarsened data in the training sample while fulfilling the primary objective of well-calibrated conformal predictive inference, requires robustness and efficiency considerations. In this paper, we consider the general problem of obtaining distribution-free valid prediction regions for an outcome given coarsened training data. Leveraging modern semiparametric theory, we achieve our goal by deriving the efficient influence function of the quantile of the outcome we aim to predict, under a given semiparametric model for the coarsened data, carefully combined with a novel conformal risk control procedure. Our principled use of semiparametric theory has the key advantage of facilitating flexible machine learning methods such as random forests to learn the underlying nuisance functions of the semiparametric model. A straightforward application of the proposed general framework produces prediction intervals with stronger coverage properties under covariate shift, as well as the construction of multiply robust prediction sets in monotone missingness scenarios. We further illustrate the performance of our methods through various simulation studies.

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. A Joint Finite-Sample Certificate for Adaptive Selective Conformal Risk Control

    cs.LG 2026-06 unverdicted novelty 6.0 of 10

    A joint finite-sample certificate for adaptive selective conformal risk control that treats selected risk as a ratio and couples empirical-Bernstein, Clopper-Pearson, and closeness bounds.

Reference graph

Works this paper leans on

50 extracted references · 33 canonical work pages · cited by 1 Pith paper

  1. [1]

    Conformal risk control

    Anastasios N Angelopoulos, Stephen Bates, Adam Fisch, Lihua Lei, and Tal Schuster. Conformal risk control. arXiv preprint arXiv:2208.02814, 2022

    arXiv 2022

  2. [2]

    Near-optimal rate of consistency for linear models with missing values

    Alexis Ayme, Claire Boyer, Aymeric Dieuleveut, and Erwan Scornet. Near-optimal rate of consistency for linear models with missing values. In International Conference on Machine Learning, pages 1211--1243. PMLR, 2022

    work page 2022

  3. [3]

    Predictive inference with the jackknife+

    Rina Foygel Barber, Emmanuel J Candes, Aaditya Ramdas, and Ryan J Tibshirani. Predictive inference with the jackknife+. The Annals of Statistics, 49 0 (1): 0 486--507, 2021

    work page 2021

  4. [4]

    Conformal prediction beyond exchangeability

    Rina Foygel Barber, Emmanuel J Candes, Aaditya Ramdas, and Ryan J Tibshirani. Conformal prediction beyond exchangeability. The Annals of Statistics, 51 0 (2): 0 816--845, 2023

    2023

  5. [5]

    Predictive inference with weak supervision

    Maxime Cauchois, Suyash Gupta, Alnur Ali, and John C Duchi. Predictive inference with weak supervision. Journal of Machine Learning Research, 25 0 (118): 0 1--45, 2024 a

    work page 2024

  6. [6]

    Robust validation: Confident predictions even when distributions shift

    Maxime Cauchois, Suyash Gupta, Alnur Ali, and John C Duchi. Robust validation: Confident predictions even when distributions shift. Journal of the American Statistical Association, 119 0 (548): 0 3033--3044, 2024 b

    work page 2024

  7. [7]

    Distributional conformal prediction

    Victor Chernozhukov, Kaspar W \"u thrich, and Yinchu Zhu. Distributional conformal prediction. Proceedings of the National Academy of Sciences, 118 0 (48): 0 e2107794118, 2021

    work page 2021

  8. [8]

    Methotrexate and mortality in patients with rheumatoid arthritis: a prospective study

    Hyon K Choi, Miguel A Hern \'a n, John D Seeger, James M Robins, and Frederick Wolfe. Methotrexate and mortality in patients with rheumatoid arthritis: a prospective study. The Lancet, 359 0 (9313): 0 1173--1177, 2002

    work page 2002

Show all 50 references
  1. [9]

    Relative expected instantaneous loss bounds

    J \"u rgen Forster and Manfred K Warmuth. Relative expected instantaneous loss bounds. Journal of Computer and System Sciences, 64 0 (1): 0 76--102, 2002

    2002

  2. [10]

    The limits of distribution-free conditional predictive inference

    Rina Foygel Barber, Emmanuel J Candes, Aaditya Ramdas, and Ryan J Tibshirani. The limits of distribution-free conditional predictive inference. Information and Inference: A Journal of the IMA, 10 0 (2): 0 455--482, 2021

    2021

  3. [11]

    Conformal prediction with conditional guarantees

    Isaac Gibbs, John J Cherian, and Emmanuel J Cand \`e s. Conformal prediction with conditional guarantees. arXiv preprint arXiv:2305.12616, 2023

    2023 arXiv

  4. [12]

    Nearest neighbor based conformal prediction

    Laszlo Gy \"o rfi and Harro Walk. Nearest neighbor based conformal prediction. In Annales de l'ISUP, volume 63, pages 173--190, 2019

    2019

  5. [13]

    Introduction to the special section on missing data

    Julie Josse and Jerome P Reiter. Introduction to the special section on missing data. 2018

    2018

  6. [14]

    On the consistency of supervised learning with missing values

    Julie Josse, Jacob M Chen, Nicolas Prost, Ga \"e l Varoquaux, and Erwan Scornet. On the consistency of supervised learning with missing values. Statistical Papers, 65 0 (9): 0 5447--5479, 2024

    2024

  7. [15]

    Batch multivalid conformal prediction

    Christopher Jung, Georgy Noarov, Ramya Ramalingam, and Aaron Roth. Batch multivalid conformal prediction. arXiv preprint arXiv:2209.15145, 2022

    2022 arXiv

  8. [16]

    Predictive inference is free with the jackknife+-after-bootstrap

    Byol Kim, Chen Xu, and Rina Barber. Predictive inference is free with the jackknife+-after-bootstrap. Advances in Neural Information Processing Systems, 33: 0 4138--4149, 2020

    2020

  9. [17]

    Exchangeability, conformal prediction, and rank tests

    Arun Kumar Kuchibhotla. Exchangeability, conformal prediction, and rank tests. arXiv preprint arXiv:2005.06095, 2020

    2005 arXiv

  10. [18]

    Unified methods for censored longitudinal data and causality

    Mark J Laan and James M Robins. Unified methods for censored longitudinal data and causality. Springer, 2003

    2003

  11. [19]

    Neumiss networks: differentiable programming for supervised learning with missing values

    Marine Le Morvan, Julie Josse, Thomas Moreau, Erwan Scornet, and Ga \"e l Varoquaux. Neumiss networks: differentiable programming for supervised learning with missing values. Advances in Neural Information Processing Systems, 33: 0 5980--5990, 2020 a

    2020

  12. [20]

    Linear predictor on linearly-generated data with missing values: non consistency and solutions

    Marine Le Morvan, Nicolas Prost, Julie Josse, Erwan Scornet, and Ga \"e l Varoquaux. Linear predictor on linearly-generated data with missing values: non consistency and solutions. In International Conference on Artificial Intelligence and Statistics, pages 3165--3174. PMLR, 2020 b

    2020

  13. [21]

    What’sa good imputation to predict with missing values? Advances in Neural Information Processing Systems, 34: 0 11530--11540, 2021

    Marine Le Morvan, Julie Josse, Erwan Scornet, and Ga \"e l Varoquaux. What’sa good imputation to predict with missing values? Advances in Neural Information Processing Systems, 34: 0 11530--11540, 2021

    2021

  14. [22]

    Distribution-free prediction bands for non-parametric regression

    Jing Lei and Larry Wasserman. Distribution-free prediction bands for non-parametric regression. Journal of the Royal Statistical Society Series B: Statistical Methodology, 76 0 (1): 0 71--96, 2014

    2014

  15. [23]

    Distribution-free prediction sets

    Jing Lei, James Robins, and Larry Wasserman. Distribution-free prediction sets. Journal of the American Statistical Association, 108 0 (501): 0 278--287, 2013

    2013

  16. [24]

    Distribution-free predictive inference for regression

    Jing Lei, Max G’Sell, Alessandro Rinaldo, Ryan J Tibshirani, and Larry Wasserman. Distribution-free predictive inference for regression. Journal of the American Statistical Association, 113 0 (523): 0 1094--1111, 2018

    2018

  17. [25]

    Conformal inference of counterfactuals and individual treatment effects

    Lihua Lei and Emmanuel J Cand \'e s. Conformal inference of counterfactuals and individual treatment effects. Journal of the Royal Statistical Society Series B: Statistical Methodology, 83 0 (5): 0 911--938, 2021

    2021

  18. [26]

    Statistical analysis with missing data

    Roderick JA Little and Donald B Rubin. Statistical analysis with missing data. John Wiley & Sons, 2019

    2019

  19. [27]

    Inductive confidence machines for regression

    Harris Papadopoulos, Kostas Proedrou, Volodya Vovk, and Alex Gammerman. Inductive confidence machines for regression. In Machine learning: ECML 2002: 13th European conference on machine learning Helsinki, Finland, August 19--23, 2002 proceedings 13, pages 345--356. Springer, 2002

    2002

  20. [28]

    Super learner in prediction

    Eric C Polley and Mark J Van der Laan. Super learner in prediction. 2010

    2010

  21. [29]

    Prediction sets adaptive to unknown covariate shift

    Hongxiang Qiu, Edgar Dobriban, and Eric Tchetgen Tchetgen. Prediction sets adaptive to unknown covariate shift. Journal of the Royal Statistical Society Series B: Statistical Methodology, 85 0 (5): 0 1680--1705, 2023

    2023

  22. [30]

    Non-response models for the analysis of non-monotone ignorable missing data

    James M Robins and Richard D Gill. Non-response models for the analysis of non-monotone ignorable missing data. Statistics in medicine, 16 0 (1): 0 39--56, 1997

    1997

  23. [31]

    Estimation of regression coefficients when some regressors are not always observed

    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

    1994

  24. [32]

    Conformalized quantile regression

    Yaniv Romano, Evan Patterson, and Emmanuel Candes. Conformalized quantile regression. Advances in neural information processing systems, 32, 2019

    2019

  25. [33]

    Classification with valid and adaptive coverage

    Yaniv Romano, Matteo Sesia, and Emmanuel Candes. Classification with valid and adaptive coverage. Advances in neural information processing systems, 33: 0 3581--3591, 2020

    2020

  26. [34]

    Characterization of parameters with a mixed bias property

    Andrea Rotnitzky, Ezequiel Smucler, and James M Robins. Characterization of parameters with a mixed bias property. Biometrika, 108 0 (1): 0 231--238, 2021

    2021

  27. [35]

    Conformal prediction using conditional histograms

    Matteo Sesia and Yaniv Romano. Conformal prediction using conditional histograms. Advances in Neural Information Processing Systems, 34: 0 6304--6315, 2021

    2021

  28. [36]

    Conformal time-series forecasting

    Kamile Stankeviciute, Ahmed M Alaa, and Mihaela van der Schaar. Conformal time-series forecasting. Advances in neural information processing systems, 34: 0 6216--6228, 2021

    2021

  29. [37]

    On inverse probability weighting for nonmonotone missing at random data

    BaoLuo Sun and Eric J Tchetgen Tchetgen. On inverse probability weighting for nonmonotone missing at random data. Journal of the American Statistical Association, 113 0 (521): 0 369--379, 2018

    2018

  30. [38]

    Inverse-probability-weighted estimation for monotone and nonmonotone missing data

    BaoLuo Sun, Neil J Perkins, Stephen R Cole, Ofer Harel, Emily M Mitchell, Enrique F Schisterman, and Eric J Tchetgen Tchetgen. Inverse-probability-weighted estimation for monotone and nonmonotone missing data. American journal of epidemiology, 187 0 (3): 0 585--591, 2018

    2018

  31. [39]

    Statistical methods for robust inference in causal and missing data models

    Eric Joel Tchetgen Tchetgen. Statistical methods for robust inference in causal and missing data models. Harvard University, 2006

    2006

  32. [40]

    Conformal prediction under covariate shift

    Ryan J Tibshirani, Rina Foygel Barber, Emmanuel Candes, and Aaditya Ramdas. Conformal prediction under covariate shift. Advances in neural information processing systems, 32, 2019

    2019

  33. [41]

    Semiparametric theory and missing data, volume 4

    Anastasios A Tsiatis. Semiparametric theory and missing data, volume 4. Springer, 2006

    2006

  34. [42]

    The missing indicator method: From low to high dimensions

    Mike Van Ness, Tomas M Bosschieter, Roberto Halpin-Gregorio, and Madeleine Udell. The missing indicator method: From low to high dimensions. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 5004--5015, 2023

    2023

  35. [43]

    Conditional validity of inductive conformal predictors

    Vladimir Vovk. Conditional validity of inductive conformal predictors. In Asian conference on machine learning, pages 475--490. PMLR, 2012

    2012

  36. [44]

    Mondrian confidence machine

    Vladimir Vovk, David Lindsay, Ilia Nouretdinov, and Alex Gammerman. Mondrian confidence machine. Technical Report, 2003

    2003

  37. [45]

    Algorithmic learning in a random world, volume 29

    Vladimir Vovk, Alexander Gammerman, and Glenn Shafer. Algorithmic learning in a random world, volume 29. Springer, 2005

    2005

  38. [46]

    Finite-sample efficient conformal prediction

    Yachong Yang and Arun Kumar Kuchibhotla. Finite-sample efficient conformal prediction. arXiv preprint arXiv:2104.13871, 5, 2021

    2021 arXiv

  39. [47]

    Forster-warmuth counterfactual regression: A unified learning approach

    Yachong Yang, Arun Kumar Kuchibhotla, and Eric Tchetgen Tchetgen. Forster-warmuth counterfactual regression: A unified learning approach. arXiv preprint arXiv:2307.16798, 2023

    2023 arXiv

  40. [48]

    Doubly robust calibration of prediction sets under covariate shift

    Yachong Yang, Arun Kumar Kuchibhotla, and Eric Tchetgen Tchetgen. Doubly robust calibration of prediction sets under covariate shift. Journal of the Royal Statistical Society Series B: Statistical Methodology, page qkae009, 2024

    2024

  41. [49]

    Conformal prediction with missing values

    Margaux Zaffran, Aymeric Dieuleveut, Julie Josse, and Yaniv Romano. Conformal prediction with missing values. In International Conference on Machine Learning, pages 40578--40604. PMLR, 2023

    2023

  42. [50]

    Predictive uncertainty quantification with missing covariates

    Margaux Zaffran, Julie Josse, Yaniv Romano, and Aymeric Dieuleveut. Predictive uncertainty quantification with missing covariates. arXiv preprint arXiv:2405.15641, 2024

    2024 arXiv

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.