Pith. sign in

REVIEW 2 major objections 5 minor 30 references

Beyond Marginal Validity: Finite-Sample Guarantees for Localized Conformal Prediction

T0 review · 2 major / 5 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 Conditional on the realized localization center, RLCP bounds the conditional-coverage gap and the oracle-relative length error uniformly over the localization ball with probability at least 1−δ over calibration, at the classical…

desk verdict Finite-sample uniform-local guarantees for RLCP with an honest fixed-score analysis; the length bound is vacuous when the localized score density vanishes, and the abstract should say so. read the letter →

arxiv 2608.06206 v1 pith:Q3D7DPP5 submitted 2026-08-06 stat.ML cs.LG

Anton Conrad , Rustam Isaev , Denis Belomestny , Eric Moulines , Sergey Samsonov This is my paper · ORCID
classification stat.MLcs.LG MSC 62G0862G1562G20
keywords conformalpredictionconditionalcoveragedistribution-freeinferencefinite-sampleguaranteeslocalizationquantileregressionrandomlylocalized
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

Conformal prediction guarantees marginal coverage on average over covariates, but that average can hide severe under-coverage in specific regions, and exact distribution-free conditional coverage is known to be unattainable in finite samples. Randomly localized conformal prediction (RLCP) attacks this gap by drawing an auxiliary localization center and reweighting calibration observations near it, while preserving the marginal guarantee. This paper proves finite-sample, high-probability control of the realized RLCP set: conditional on the drawn center, simultaneously for every covariate in the localization ball, the conditional-coverage gap and the interval-length error relative to the infeasible score oracle are bounded by explicit rates. Those rates decompose into a localization bias of order $h^\beta$ and a calibration term driven by the local effective sample size $n p_{\min} h^d$, which makes the bandwidth bias–variance trade-off explicit and, when balanced, yields the classical nonparametric rate $n^{-\beta/(2\beta+d)}$. For scores learned on an independent fold that target a pivotal score, the localization bias disappears and the learning error enters linearly.

What carries the argument

Three devices carry the argument. The reverse-law representation: after the auxiliary center $\tilde{x}$ is drawn from the kernel-smoothed law, the test covariate, conditional on $\tilde{x}$, has exactly the localized distribution $\Lambda_{\tilde{x},h}(du) = (w_{\tilde{x},h}(u)/Z_{\tilde{x},h})P_X(du)$, so RLCP is weighted conformal prediction under covariate shift and inherits finite-sample marginal validity. The calibration event $\Omega^{\mathrm{cal}}_{n,\tilde{x},h,\delta}$: a single event of probability at least $1-\delta$ on which the kernel-weighted empirical score CDF concentrates uniformly and the test-point weight is uniformly small, so every bound holds simultaneously for all $x \in B(\tilde{x}, h) \cap \mathcal{X}$. The inversion devices: the localized score-density floor $\kappa^{(S)}_{\alpha,\tilde{x},h}$ (the essential infimum of the localized score density on the quantile window) converts level error into quantile error, the length map $q \mapsto |C^{(S)}(x,q)|$ converts quantile error into length error, and Hölder continuity of the conditional score CDF converts the mismatch between the local mixture and the target covariate into the $h^\beta$ bias.

What would settle it

Run RLCP at the balanced bandwidth $h_\star \asymp (\log(4n)/(n p_{\min}))^{1/(2\beta+d)}$ on a score that satisfies the paper's A1–A3 but whose conditional score density vanishes on a band around the $(1-\alpha)$-quantile (for instance, a residual score with a gap in the noise density): if the realized oracle-relative length error does not shrink at the predicted $n^{-\beta/(2\beta+d)}$ rate, or stays bounded while the coverage gap shrinks, then the length claim fails precisely in the regime where $\kappa^{(S)}_{\alpha,\tilde{x},h} = 0$ makes the theorem's bound vacuous.

Watch

Extended reading notes

Core claim

With probability at least $1-\delta$ over the calibration sample, conditional on the realized localization center $\tilde{x}$, and for calibration sizes and bandwidths above explicit thresholds, the RLCP threshold deviates from the conditional score-oracle quantile by $O\big((A^{\mathrm{cal}}_{n,h}(\delta) + h^\beta)/\kappa^{(S)}_{\alpha,\tilde{x},h}\big)$, uniformly over $B(\tilde{x}, h) \cap \mathcal{X}$; from this, the conditional-coverage gap is bounded by $O(A^{\mathrm{cal}}_{n,h}(\delta) + h^\beta)$ (Theorem 3) and the oracle-relative length error by $O\big(L_{\mathrm{len}}(x)\,(A^{\mathrm{cal}}_{n,h}(\delta) + h^\beta)/\kappa^{(S)}_{\alpha,\tilde{x},h}\big)$ (Theorem 2), where $A^{\mathrm{cal}}_{n,h}(\delta) = \sqrt{\log(4/\delta)/(n p_{\min} h^d)} + \log(4/\delta)/(n p_{\min} h^d)$ captures the local effective sample size and $\kappa^{(S)}_{\alpha,\tilde{x},h}$ is the localized score-density floor. The first term in the rates is the localization bias coming from Hölder regularity of the conditional score law; the others are calibration error, and balancing them at $h_\star \asymp (\log(4n)/(n p_{\min}))^{1/(2\beta+d)}$ gives the classical pointwise nonparametric rate $n^{-\beta/(2\beta+d)}$ up to logarithmic factors. When the score is learned on an independent fold and targets a pivotal population score — one whose $(1-\alpha)$-quantile is the same at every covariate, as with conformalized quantile regression and the probability-integral-transform score — the localization bias vanishes, and Theorems 6 and 7 deliver the same uniform local control with the error split into calibration error plus a linearly entering uniform score-estimation error.

Load-bearing premise

The load-bearing premise is that the localized score distribution keeps strictly positive density across the quantile window; the paper's own passage before Theorem 2 concedes that this is not guaranteed by the main assumptions, and when that density floor is zero the stated length bound is formally true but vacuous, so the oracle-length guarantee rests on the separate shape-function condition of Lemma 1.

Editorial extensions

If this is right

  • Balancing the two error sources pins the optimal bandwidth at $h_\star \asymp (\log(4n)/(n p_{\min}))^{1/(2\beta+d)}$, where both the coverage gap and the length error attain the classical rate $n^{-\beta/(2\beta+d)}$ up to logarithmic factors (equations (19)–(20)).
  • Because the calibration event is sample-specific and common to the whole ball, the guarantee is simultaneous: no union bound over test covariates and no averaging over the auxiliary randomization is required for uniformity over $B(\tilde{x}, h) \cap \mathcal{X}$.
  • Under a pivotal target score the localization bias disappears and the bandwidth enters only through the effective calibration size $n p_{\min} h^d$, so the score-estimation error enters linearly: improving the learned score sharpens the localized oracle comparison at exactly the training rate.
  • The assumptions, including the density-floor condition via Lemma 1, are verified for residual, conformalized-quantile-regression, and distributional scores, so the bounds apply to the standard score constructions in practice.

Reading between the lines

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

  • A design lesson the paper leaves implicit: when a strong, near-pivotal score is available, the bandwidth should be pushed as large as the density-floor constraint allows rather than tuned to the fixed-score optimum, since localization then mainly buys effective sample size while the score itself carries covariate adaptivity.
  • The vacuous-kappa caveat points to a concrete stress test: scores whose conditional density is tiny at the target quantile (for example, extreme-quantile regions in heavy-tailed responses) should show realized length errors far above the predicted rate, and mapping where the degradation begins would delineate the exact class of scores for which the oracle-tracking claim holds.
  • The calibration-versus-localization decomposition is the same structure that governs locally weighted regression, suggesting the $n^{-\beta/(2\beta+d)}$ rate is the natural minimax benchmark for any locally weighted conformal procedure; the pivotal-score result further predicts that with a well-learned score, wide-bandwidth (near-marginal) calibration is nearly optimal, since localization adds no
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

2 major / 5 minor

Summary. The paper develops finite-sample guarantees for randomly localized conformal prediction (RLCP) targeted at the realized prediction set rather than at marginal or averaged quantities. For a fixed score, under Hölder regularity of the conditional score CDF, kernel and design conditions, and a length-regularity assumption, Theorems 2 and 3 give, with probability at least 1−δ over the calibration sample and conditional on the realized localization center, uniform bounds over the localization ball for the conditional-coverage gap and for the oracle-relative length error. The bounds combine a calibration term A_cal_{n,h}(δ) with a localization bias h^β; the length bound also carries the inverse of the localized score-density floor κ. For a data-split learned score that targets a pivotal population score, Theorems 6 and 7 give analogous bounds in which calibration error and score-estimation error enter separately and linearly. The residual, CQR, and distributional scores are shown to satisfy the structural assumptions, and numerical experiments illustrate the predicted rates and the coverage-length trade-off in a disclosed real-data hybrid implementation.

Significance. If the results are taken with the required positivity of the localized score-density floor, the paper is a substantial contribution: it gives high-probability, realized-center oracle comparisons for RLCP, explicitly separates calibration error from localization bias, and clarifies when learned pivotal scores remove the localization bias. The proofs are detailed and appear internally consistent, with explicit lemmas for concentration, quantile inversion, and score perturbation, and the experimental sections are thorough and transparent, including exact seeds, bandwidth grids, and disclosure that the real-data procedure is a hybrid rather than formal RLCP. The main caveat is that the advertised length guarantee is non-vacuous only under an additional density-minorization condition that is not part of A1–A3; the paper itself notes this after Lemma 1, but the abstract and introduction do not.

major comments (2)
  1. [Abstract and Section 3.2, Theorem 2] The abstract's claim that, for any fixed score, the paper proves finite-sample bounds for the length error relative to the oracle is not supported as stated. The right-hand side of Theorem 2 contains the factor 1/κ(S)_{α,x-tilde,h}, and under the convention 1/0=∞ the bound is formally true but vacuous whenever κ=0. Assumptions A1–A3 do not imply κ>0: A1 is an upper bound on the score density, A2 is a Hölder condition on the score CDF, and A3 concerns length regularity. The only sufficient condition supplied, Lemma 1's inequality (17), requires a covariate-local lower bound on p_{S|X} through a shape function, and it is verified for the residual, CQR, and distributional scores but not for arbitrary fixed scores. The text after Lemma 1 acknowledges the vacuity, so the theorem itself is formally correct, but the abstract and the introduction should qualify the length guarantee by κ>0 or by an equivalent density-floor assumption.
  2. [Section 4.2, Theorems 6 and 7 and Proposition 18] The same κ-positivity issue propagates to the learned-score results. The bounds in (28) and (29) are multiplied by 1/κ(S⋆)_{α,x-tilde,h}, yet the hypothesis list of Theorem 6 does not state κ(S⋆)>0. Assumptions A1(S⋆), A4(α), and A5 do not imply this positivity; Lemma 15's identification of the localized quantile already assumes κ(S⋆)>0, and Proposition 18 includes it as a hypothesis in the text. As written, Theorem 6 is non-vacuous only for target scores whose localized density floor is positive. The theorem statements should include this condition explicitly, or the theorems should be qualified as informative only when κ(S⋆)>0.
minor comments (5)
  1. [Abstract] The phrase 'for any fixed score' should be replaced by a formulation that explicitly conditions on the localized density minorization condition (13) for the length result.
  2. [Section 1, Eq. (1) and following text] The description of the length bound as a 'constant multiple' is imprecise because the factor 1/κ(S)_{α,x-tilde,h} is not a universal structural constant; it depends on the realized center, the bandwidth, and the score, and it can be infinite under the stated convention.
  3. [Section 3.3, Proposition 4 and Lemma 1] For the fixed-score examples, positivity of κ is verified only for the symmetric level choice (15), whereas Theorems 2 and 3 are stated for arbitrary α−<α<α+. The paper should clarify whether the main fixed-score length theorem is intended for those symmetric levels or whether the user must verify (13) separately for other level choices.
  4. [Section 5 and Appendix M] The real-data procedure is carefully disclosed as a hybrid, but referring to it simply as 'RLCP' in the decile figures and tables may mislead readers; a label such as 'RLCP-hybrid' would make the distinction from formal RLCP clearer.
  5. [Figure 5 caption] In the displayed expression for the decomposition proxy, the placeholder 'slow' appears where κ(S⋆) is intended; this should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: bounds follow from stated structural conditions and independent concentration results; the κ>0 caveat is a vacuity limitation, not a circular step.

full rationale

The paper's fixed-score analysis derives Theorem 3's coverage gap from Bernstein/Bousquet/Talagrand concentration (Proposition 11), A2 Hölder bias, and direct CDF comparisons; Theorem 2 additionally converts threshold error to set-length error via A3 and the quantile-inversion cost 1/κ. None of these steps restates its conclusion: κ(S)_{α,x̃,h} is defined as an essential infimum of the localized score density (13), not as a fitted parameter, and the theorem's RHS is not used to define any input. Lemma 1 gives a sufficient, unverified-for-arbitrary-scores condition for κ>0, and the paper explicitly records that the length bound is vacuous when κ=0 ('While Theorem 2 below remains valid for κ=0, the resulting upper bound is vacuous'), which is a limitation rather than circularity. The learned-score results (Theorems 6, 7) take A5's uniform score-estimation rate as an assumption and propagate it linearly through Lemmas 16-17; the rate is not fitted to the same outcomes being bounded. No load-bearing self-citations appear: references to Hore-Barber, Min et al., Romano et al., etc. are external prior results, and the reverse-law representation is used as a starting point, not as the paper's own uniqueness claim. The simulations are explicitly rate diagnostics and do not enter the theorem chain. Thus there is no exhibited reduction of any claimed prediction to its own inputs.

Assumptions & free parameters 0 free parameters · 8 assumptions · 0 invented entities

No fitted parameters are introduced into the theorems; all constants are structural and appear as assumptions. The only chosen schedule is the theoretical bandwidth h* in the discussion, used to balance the bound, not fitted to data. The auxiliary localization center is the existing RLCP construction of Hore and Barber, not a newly invented entity.

assumptions (8)
  • domain assumption P1: marginal density p_X bounded below by p_min and X is (c_reg, r0)-regular.
    Used via Lemma 10 to prove Z_x~,h >= c0 p_min h^d, which drives every effective sample-size threshold and the calibration event.
  • domain assumption K1: kernel K is bounded, symmetric, integrates to one, supported on B(0,1) and lower-bounded by kappa on B(0,eta).
    The kernel lower bound is what produces the positive local mass c0 p_min h^d; the symmetry is needed for the reverse-law identity (8).
  • domain assumption A1(S): conditional score density bounded above by sigma_up.
    Converts threshold deviations into coverage deviations in Lemmas 16 and 25, and is used throughout the learned-score analysis.
  • domain assumption A2(S): conditional score CDF is beta-Holder in the covariate, uniformly in t.
    Controls the localization bias h^beta in Theorems 2 and 3 and in Proposition 14.
  • domain assumption A3(S): the map q -> |C(S)(x,q)| is L_len(x)-Lipschitz.
    Turns the quantile-oracle deviation into the oracle-relative length error in Theorem 2 and Theorem 6.
  • domain assumption A4(alpha): target score S* is (1-alpha)-pivotal.
    Used in Lemma 15 to show the localized oracle quantile equals the pivot tau*, removing the h^beta term in the learned-score theorems.
  • domain assumption A5(S): uniform high-probability bound on the essential supremum of the score estimation error.
    Encodes the training-fold error rate for Theorems 6 and 7; the examples verify it for local-polynomial CQR and NW conditional-CDF estimators.
  • domain assumption Lemma 1 condition (17): conditional score density is lower-bounded by sigma_psi(u,h) psi(F_{S|X}(s|z)).
    Sufficient condition for the localized density minorization kappa > 0, which is needed for non-vacuous length bounds; without it Theorem 2 is vacuous.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Beyond Marginal Validity: Finite-Sample Guarantees for Localized Conformal Prediction." pith.science (2026). https://pith.science/paper/Q3D7DPP5

@misc{pith2026260806206,
  author       = {Pith},
  title        = {Pith review of: Beyond Marginal Validity: Finite-Sample Guarantees for Localized Conformal Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q3D7DPP5}},
  note         = {Machine review of arXiv:2608.06206}
}
abstract

Conformal prediction endows arbitrary black-box predictors with finite-sample, distribution-free marginal coverage, yet marginal validity can hide severe covariate-specific miscalibration, while exact distribution-free conditional coverage is finite-sample unattainable. Randomly localized conformal prediction (RLCP) mitigates this gap by calibrating near the test point while preserving marginal coverage. Existing theory, however, lacks finite-sample guarantees for the realized localized set that jointly control conditional validity and oracle efficiency. We provide such guarantees. For any fixed score, under H\"older regularity of the conditional score CDF and standard density and kernel assumptions, we prove high-probability bounds, uniform over a realized localization neighbourhood, for the conditional-coverage gap and the length error relative to the oracle. The bounds decompose into an $O(h^\beta)$ localization bias and a calibration term decreasing with calibration size, clarifying the bandwidth bias-variance tradeoff and when RLCP tracks the oracle. We also analyze data-split learned scores: when the score targets a pivotal score, as in conformalized quantile regression, uniform local guarantees decompose into fixed-score calibration and uniform score-estimation errors, showing that improved learning sharpens localized guarantees.

Figures

Figures reproduced from arXiv: 2608.06206 by the authors.

Figure 1
Figure 1. Oracle tracking at the balanced bandwidth [PITH_FULL_IMAGE:figures/full_fig_p017_1.png] view at source ↗
Figure 2
Figure 2. Bias–variance trade-off of the formal RLCP [PITH_FULL_IMAGE:figures/full_fig_p062_2.png] view at source ↗
Figure 3
Figure 3. Companion conditional-coverage diagnostic for the bandwidth study of [PITH_FULL_IMAGE:figures/full_fig_p063_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Uniform score-estimation error metrics (148) against the training size m, for the local-linear CQR pair (Proposition 8) and the Nadaraya–Watson conditional-CDF estimator (Proposition 9), at β = 0.5, 1. Each metric is a numerical proxy for the essential supremum in (146…
Figure 5
Figure 5. Figure 5: Measured RLCP oracle errors versus the empirical decomposition proxy [PITH_FULL_IMAGE:figures/full_fig_p065_5.png]
Figure 6
Figure 6. Figure 6: Real-data decile diagnostics. Top: conditional coverage by decile of KDE-estimated [PITH_FULL_IMAGE:figures/full_fig_p066_6.png]
Figure 7
Figure 7. Figure 7: Real-data bandwidth sensitivity. Each quantity is recomputed at bandwidth multipliers [PITH_FULL_IMAGE:figures/full_fig_p067_7.png]
Figure 8
Figure 8. Figure 8: Worst-slice coverage gain versus mean-length ratio, paired against [PITH_FULL_IMAGE:figures/full_fig_p068_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 12 canonical work pages

  1. [1]

    Tsybakov

    Jean-Yves Audibert and Alexandre B. Tsybakov. Fast learning rates for plug-in classifiers.The Annals of Statistics, 35(2):608–633, 2007. doi: 10.1214/009053606000001217

    doi:10.1214/009053606000001217 2007

  2. [2]

    Tibshirani

    Rina Foygel Barber and Ryan J. Tibshirani. Unifying different theories of conformal prediction. Electronic Journal of Statistics, 20(1):1428–1474, 2026. doi: 10.1214/26-EJS2513. URL https://doi.org/10.1214/26-EJS2513

    work page doi:10.1214/26-ejs2513 2026

  3. [3]

    Candès, Aaditya Ramdas, and Ryan J

    Rina Foygel Barber, Emmanuel J. Candès, Aaditya Ramdas, and Ryan J. Tibshirani. The limits of distribution-free conditional predictive inference.Information and Inference: A Journal of the IMA, 10(2):455–482, 2021. doi: 10.1093/imaiai/iaaa017. URLhttps://doi.org/10.1093/ imaiai/iaaa017

    doi:10.1093/imaiai/iaaa017 2021

  4. [4]

    Bobkov and Michel Ledoux.One-Dimensional Empirical Measures, Order Statistics, and Kantorovich Transport Distances, volume 261 ofMemoirs of the American Mathematical Society

    Sergey G. Bobkov and Michel Ledoux.One-Dimensional Empirical Measures, Order Statistics, and Kantorovich Transport Distances, volume 261 ofMemoirs of the American Mathematical Society. American Mathematical Society, 2019. doi: 10.1090/memo/1259

    doi:10.1090/memo/1259 2019

  5. [5]

    Oxford University Press, Oxford, 2013

    Stéphane Boucheron, Gábor Lugosi, and Pascal Massart.Concentration Inequalities: A Nonasymptotic Theory of Independence. Oxford University Press, Oxford, 2013. ISBN 9780199535255. doi: 10.1093/acprof:oso/9780199535255.001.0001. URL https://doi.org/10. 1093/acprof:oso/9780199535255.001.0001. 59 Beyond Marginal ValidityConrad et al

    arXiv 2013

  6. [6]

    Pope, and Michael A

    Thomas Brooks, D. Pope, and Michael A. Marcolini. Airfoil self-noise. UCI Machine Learning Repository, 1989. URLhttps://doi.org/10.24432/C5VW2C

    doi:10.24432/c5vw2c 1989

  7. [7]

    Nonparametric estimates of regression quantiles and their local Bahadur representation.The Annals of Statistics, 19(2):760–777, 1991

    Probal Chaudhuri. Nonparametric estimates of regression quantiles and their local Bahadur representation.The Annals of Statistics, 19(2):760–777, 1991. doi: 10.1214/aos/1176348119

    arXiv 1991

  8. [8]

    Generic inference onquantileandquantileeffectfunctionsfordiscreteoutcomes.Journal of the American Statistical Association, 115(529):123–137, 2020

    Victor Chernozhukov, Iván Fernández-Val, Blaise Melly, and Kaspar Wüthrich. Generic inference onquantileandquantileeffectfunctionsfordiscreteoutcomes.Journal of the American Statistical Association, 115(529):123–137, 2020. doi: 10.1080/01621459.2019.1611581

    arXiv 2020

Show all 30 references
  1. [9]

    Distributional conformal prediction

    Victor Chernozhukov, Kaspar Wüthrich, and Yinchu Zhu. Distributional conformal prediction. Proceedings of the National Academy of Sciences, 118(48):e2107794118, 2021. doi: 10.1073/pnas. 2107794118. URLhttps://arxiv.org/abs/1909.07889

    2021 arXiv

  2. [10]

    Dhillon, George Deligiannidis, and Tom Rainforth

    Guneet S. Dhillon, George Deligiannidis, and Tom Rainforth. On the expected size of conformal prediction sets. InProceedings of The 27th International Conference on Artificial Intelligence and Statistics, volume 238 ofProceedings of Machine Learning Research, pages 1549–1557. ...

    2024

  3. [11]

    Uwe Einmahl and David M. Mason. Uniform in bandwidth consistency of kernel-type function estimators.The Annals of Statistics, 33(3):1380–1403, 2005. doi: 10.1214/009053605000000129. URLhttps://doi.org/10.1214/009053605000000129

    2005 doi

  4. [12]

    A note on generalized inverses.Mathematical Methods of Operations Research, 77(3):423–432, 2013

    Paul Embrechts and Marius Hofert. A note on generalized inverses.Mathematical Methods of Operations Research, 77(3):423–432, 2013. doi: 10.1007/s00186-013-0436-7

    2013 doi

  5. [13]

    Bike sharing

    Hadi Fanaee-T. Bike sharing. UCI Machine Learning Repository, 2013. URLhttps://doi. org/10.24432/C5W894

    2013 doi

  6. [14]

    Conformal prediction with localization.arXiv preprint arXiv:1908.08558, 2019

    Leying Guan. Conformal prediction with localization.arXiv preprint arXiv:1908.08558, 2019. doi: 10.48550/arXiv.1908.08558. URLhttps://arxiv.org/abs/1908.08558

    1908 doi arXiv

  7. [15]

    Localized conformal prediction: A generalized inference framework for conformal prediction.Biometrika, 110(1):33–50, 2023

    Leying Guan. Localized conformal prediction: A generalized inference framework for conformal prediction.Biometrika, 110(1):33–50, 2023. doi: 10.1093/biomet/asac040. URL https: //doi.org/10.1093/biomet/asac040

    2023 doi

  8. [16]

    Uniform bias study and Bahadur representation for local polynomial estimators of the conditional quantile function.Econometric Theory, 28(1):87–129,

    Emmanuel Guerre and Camille Sabbah. Uniform bias study and Bahadur representation for local polynomial estimators of the conditional quantile function.Econometric Theory, 28(1):87–129,

  9. [17]

    Conformal prediction with local weights: Ran- domization enables robust guarantees.Journal of the Royal Statistical Society Series B: Statistical Methodology, 87(2):549–578, 2025

    Rohan Hore and Rina Foygel Barber. Conformal prediction with local weights: Ran- domization enables robust guarantees.Journal of the Royal Statistical Society Series B: Statistical Methodology, 87(2):549–578, 2025. doi: 10.1093/jrsssb/qkae103. URL https: //doi.org/10.1093/jrss...

    2025 doi

  10. [18]

    Rafael Izbicki, Gilson Shimizu, and Rafael B. Stern. Cd-split and hpd-split: Efficient conformal regions in high dimensions.Journal of Machine Learning Research, 23(87):1–32, 2022. URL https://jmlr.org/papers/v23/20-797.html

    2022

  11. [19]

    Pappas, and Hamed Hassani

    Shayan Kiyani, George J. Pappas, and Hamed Hassani. Length optimization in conformal prediction. InAdvances in Neural Information Processing Systems, volume 37, 2024. doi: 10.52202/079017-3158. URLhttps://doi.org/10.52202/079017-3158. 60 Beyond Marginal ValidityConrad et al

    2024 doi

  12. [20]

    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(1):71–96, 2014. doi: 10.1111/rssb.12021

    2014 doi

  13. [21]

    Tibshirani, and Larry Wasserman

    Jing Lei, Max G’Sell, Alessandro Rinaldo, Ryan J. Tibshirani, and Larry Wasserman. Distribution-free predictive inference for regression.Journal of the American Statistical Associ- ation, 113(523):1094–1111, 2018. doi: 10.1080/01621459.2017.1307116

    2018

  14. [22]

    A unified theory of conditional coverage in conformal prediction with applications.arXiv preprint arXiv:2605.11602, 2026

    Yinjie Min, Liuhua Peng, and Changliang Zou. A unified theory of conditional coverage in conformal prediction with applications.arXiv preprint arXiv:2605.11602, 2026. doi: 10.48550/ arXiv.2605.11602. URLhttps://arxiv.org/abs/2605.11602

    2026 doi arXiv

  15. [23]

    Physicochemical properties of protein tertiary structure

    Prashant Rana. Physicochemical properties of protein tertiary structure. UCI Machine Learning Repository, 2013. URLhttps://doi.org/10.24432/C5QW3H

    2013 doi

  16. [24]

    Yaniv Romano, Evan Patterson, and Emmanuel J. Candès. Conformalized quantile regression. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alché-Buc, E. Fox, and R. Garnett, editors, Advances in Neural Information Processing Systems, volume 32, pages 3543–3553. Curran Associa...

    2019

  17. [25]

    Uniform confidence bands for local polynomial quantile estimators.ESAIM: Probability and Statistics, 18:265–276, 2014

    Camille Sabbah. Uniform confidence bands for local polynomial quantile estimators.ESAIM: Probability and Statistics, 18:265–276, 2014. doi: 10.1051/ps/2013035

    2014

  18. [26]

    A tutorial on conformal prediction.Journal of Machine Learning Research, 9:371–421, 2008

    Glenn Shafer and Vladimir Vovk. A tutorial on conformal prediction.Journal of Machine Learning Research, 9:371–421, 2008. URLhttps://jmlr.org/papers/v9/shafer08a.html

    2008

  19. [27]

    Springer, 2005

    Vladimir Vovk, Alexander Gammerman, and Glenn Shafer.Algorithmic Learning in a Random World. Springer, 2005. doi: 10.1007/b106715. URLhttps://doi.org/10.1007/b106715

    2005 doi

  20. [28]

    Non-asymptotic analysis of efficiency in con- formalized regression

    Yunzhen Yao, Lie He, and Michael Gastpar. Non-asymptotic analysis of efficiency in con- formalized regression. In C. Vondrick, B. Hariharan, C. Raffel, L. Pinto, D. Yang, and A. Faust, editors,International Conference on Learning Representations, volume 2026, pages 42546–42583...

    2026

  21. [29]

    Concrete compressive strength

    I-Cheng Yeh. Concrete compressive strength. UCI Machine Learning Repository, 1998. URL https://doi.org/10.24432/C5PK67. 61 Beyond Marginal ValidityConrad et al. 10¡1 100 set-length error j jC Rj ¡ jC orcj j (a) ¯ = 0:5 h ⋆ slope ¡0:50 [¡0:53; ¡ 0:49] balance ¡0:500 10¡1 100 (c...

    1998 doi

  22. [2012]

    URLhttps://doi.org/10.1017/S0266466611000132

    doi: 10.1017/S0266466611000132. URLhttps://doi.org/10.1017/S0266466611000132

    doi

Pith tools

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