ARC: Augmented-Rank Conformalization for Changepoint Localization --- Finite-Sample Validity and Distribution-Robust Efficiency

Chenchen Peng; Jie Zhang; Mixia Wu; Qijing Yan; Zhiqi Shen

REVIEW 1 major objections 4 minor 43 references

ARC: Augmented-Rank Conformalization for Changepoint Localization --- Finite-Sample Validity and Distribution-Robust Efficiency

T0 review · 1 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Rank-only scores make changepoint confidence sets transform-proof

desk verdict Rank-invariant conformal changepoint localization is a genuinely useful idea; the coverage and invariance theorems are sound, and the only real blemish is an explicitly unfinished proof of the length-contraction proposition. read the letter →

arxiv 2608.08424 v1 pith:ZYEKM3GJ submitted 2026-08-09 stat.ML cs.CVcs.LG

Chenchen Peng , Mixia Wu , Qijing Yan , Zhiqi Shen , Jie Zhang This is my paper · ORCID

classification stat.MLcs.CVcs.LG

keywords changepointlocalizationconformalinferencerankstatisticsconfidencesetdistribution-freeinvarianceefficiencytransferfinite-samplecoverage

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 proposes ARC, a family of changepoint-localization scores that read the data only through within-segment ranks. The aim is to keep the finite-sample coverage guarantee of conformal localization while making efficiency, meaning the length of the reported confidence set, transferable across distributions. ARC claims two properties: every frozen weight configuration, including a randomly initialized or mistrained network, yields coverage at least $1-\alpha$; and the entire confidence set is almost surely unchanged under strictly increasing transforms of the data, so set-length distributions certified once hold across the whole monotone orbit of a distribution pair. If true, users can certify set lengths on one distribution and carry them verbatim to heavy-tailed or skewed re-expressions, whereas plug-in likelihood-ratio scores see their set lengths drift and inflate.

What carries the argument

The machinery is a bank of rank-only segment statistics: a rank-CUSUM location channel built from maximized Wilcoxon evidence, a Mood-type squared-rank scale channel, fixed maximum or convex combinations of the two channels, and a single-hidden-layer network fed by downsampled rank paths, trained on synthetic data and frozen. Each candidate split $t$ is scored by the maximum of the relevant channel applied to the two segments, and a p-value is computed by within-segment permutation; the confidence set collects all $t$ whose p-value exceeds $\alpha$. Rank invariance carries the argument, since only the rank vector enters the score, and the frozen-parameter requirement keeps the permutation p-value exactly super-uniform.

What would settle it

Run ARC on a fixed window and again on the same window transformed by $g(x)=e^x$, sharing the permutation seed: the paper predicts identical p-values at every candidate, so any observed difference would disprove Theorem 2.

Watch

Extended reading notes

Core claim

ARC's central discovery is an efficiency-transfer theorem: because every ARC statistic is a function of within-segment ranks, applying any strictly increasing $g$ to the data leaves every rank, every split statistic, and every permutation p-value unchanged, so $C_{1-\alpha}(g(X))=C_{1-\alpha}(X)$ almost surely under shared permutation randomness. Consequently the joint law of the confidence set and changepoint depends on the pre- and post-change distributions only through their rank structure, and lengths certified once hold verbatim for every monotone-equivalent pair, however heavy-tailed. This contrasts with plug-in likelihood-ratio scores, whose set length changes with every re-expression. The paper also states finite-sample coverage for any frozen score and shows that the price of invariance is bounded by classical rank-test efficiency results.

Load-bearing premise

The guarantee rests on the score and any learned parameters being fixed before the test window is seen, and on the data inside each segment being exchangeable; if the combination rule is chosen after viewing the window, or if serial dependence breaks within-segment exchangeability, the coverage argument no longer holds exactly.

Editorial extensions

If this is right

A set-length distribution certified once, for example on Gaussian simulations, transfers verbatim to every monotone-equivalent pair: lognormal, logistic-compressed, or arbitrarily heavy-tailed images of the same rank experiment.
Plug-in likelihood-ratio scores retain coverage but their set lengths inflate under re-expression, whereas ARC set lengths are unchanged to the last replication across the tested monotone transforms.
No deployment-side calibration is needed for validity: every frozen weight state, including random initialization and a network trained on permuted labels, keeps coverage at the nominal level, with only set length affected.
On the well-log benchmark, combined ARC scores localize consensus annotations within three to five candidates, and an empty set is interpreted as a model-misfit signal at level $\alpha$ rather than a failure.
Under AR(1) serial dependence, i.i.d. within-segment permutations can drop coverage to about 0.30, and circular block permutations only repair it approximately to 0.84–0.92, so exact coverage is confined to within-segment exchangeable data.

Reading between the lines

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

If the transfer theorem is correct, localization-set calibration could become a table lookup: certify lengths once per rank structure and reuse them across every monotone-equivalent data pair, potentially removing per-dataset tuning of set sizes.
The same rank-invariance argument may extend to online changepoint localization built on e-values or exchangeability martingales, giving anytime-valid inference that is also invariant under monotone rescalings; the paper does not claim this extension.
A natural testable extension is per-coordinate rank channels for multivariate windows, turning type-awareness into coordinate attribution without forfeiting invariance; the paper gestures at but does not develop this direction.
The well-log results suggest ARC's practical benefit is insurance rather than uniform improvement: it is slightly longer than a tuned plug-in at the distribution the plug-in was built for, but it protects against distribution shift.

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.

Desk Editor's Note

read the letter

This paper is worth your time. It supplies the first score family for conformal changepoint localization whose confidence-set length is invariant under strictly increasing monotone transforms of the data. That is a real gap in the literature: coverage has been universal since Dandapanthula-Ramdas and Hore-Ramdas, but efficiency has been tied to the oracle likelihood ratio. ARC makes efficiency transferable by scoring only within-segment ranks.

The central theory is sound. Theorem 1 is the standard exact permutation-p-value argument, with the quantifier that any frozen weight configuration, including random or mistrained networks, retains finite-sample coverage. That is correctly stated and correctly proved. Theorem 2 is a direct consequence of rank invariance, and the paper is honest about the coupling caveat: shared Monte Carlo permutations make the confidence sets exactly equal, and it is the law of set length that transfers across the monotone orbit. The stress-test note checked this carefully and found no load-bearing flaw; I agree. The efficiency comparisons, the sabotage experiments, the well-log analysis, and the honest reporting of serial dependence all back up the claims. The code and result files are a real asset; the paper is reproducible in a way many conformal papers are not.

The soft spots are real but localized. Proposition 1, the length-contraction result, is explicitly only a sketch: two concentration steps in Appendix A.3 are stated without proof. That is not a fatal flaw, because the proposition is not used in Theorem 1 or Theorem 2 and the paper labels it a sketch, but it is presented as a theoretical result and should be either completed or formally downgraded to a conjecture with empirical support before the paper is treated as having fully discharged its theoretical claims. There is also a minor overstatement in Section 5.1: Table 2 says every entry is within Monte Carlo error of 0.90, but the t3 entry of 0.963 is more than three standard errors above nominal; this is a wording slip, not a substantive issue. The frozen-score requirement in Section 3.3 is a genuine limitation, but the paper states it clearly, and the price of type-adaptivity is honestly quantified.

This paper deserves a serious referee. I would send it out and ask for the Proposition 1 proof to be completed or its status downgraded, plus the Table 2 sentence fixed. The main contribution is solid, the simulations support it, and the invariance theorem is a genuine step forward for practitioners who need uncertainty sets on sensor streams under unknown marginal transformations.

Referee Report

1 major / 4 minor

Summary. The paper proposes ARC (Augmented-Rank Conformalization), a family of scores for conformal changepoint localization that depend on the data only through within-segment ranks: rank-CUSUM location and scale channels, fixed combinations, and a lightweight neural network score trained on synthetic data and frozen before deployment. The authors prove finite-sample coverage for any frozen score (Theorem 1), prove that the entire confidence set is almost surely invariant under strictly increasing marginal transformations (Theorem 2), and thereby obtain transfer of the set-length distribution across the monotone orbit of a distribution pair. They also state an asymptotic length-contraction proposition for the location channel (Proposition 1) with a proof sketch, and support the claims with simulations, a well-log case study, and explicit limitations covering serial dependence, trend alternatives, and unconditional coverage.

Significance. If the results hold, ARC fills a genuine gap: conformal wrappers make coverage universal but leave set length unprotected under heavy tails and distribution shift. The rank-based construction is simple, and the transfer theorem is a clean, falsifiable statement of distribution-robust efficiency. The paper is unusually transparent: the missing steps in Proposition 1 are acknowledged, serial dependence is tested and shown to break exactness, and conditional-on-detection coverage is separated from marginal coverage. The code and result files accompany the paper, and the simulations appear reproducible. Theorems 1 and 2 are standard but correctly executed; the main contribution is conceptual and should be of interest to the conformal inference and changepoint communities.

major comments (1)

[Section 4, Proposition 1; Appendix A.3] Proposition 1 is stated as a formal result, but the proof sketch explicitly omits two substantive steps: a sub-Gaussian tail bound for the maximum of standardized rank CUSUMs under permutation, and a union bound over the B Monte Carlo draws. In addition, the proposition statement does not specify that B must grow with n, although the proof sketch requires B to grow at least logarithmically in n. As written, the proposition is not proved. Because Theorems 1 and 2 do not depend on it, this does not undermine the central claims, but the authors should either supply a complete proof with the B_n growth condition made explicit, or reclassify the statement as a conjecture or remark with a clear caveat.

minor comments (4)

[Section 5.1, Table 2] The text says every coverage entry lies within Monte Carlo error of the nominal 0.90, but the t3 learned-combination cell reports 0.963, which is about 3.6 standard errors above nominal under 300 replications; please either provide a multiple-testing justification for the set of cells or recheck this cell.
[Section 5.2, Figure 2] The claim of exactly coincident p-value profiles relies on sharing the Monte Carlo permutation draws between the original and transformed windows; the text is clear about this, but it may help to state explicitly that without shared randomness the equality is distributional rather than pathwise.
[Section 3.2] In the definition of the scale channel, the quantities q_i, \bar q, and v_q are used before v_q is defined; please state whether v_q is the sample variance with divisor m-1 or m, since the normalization constant affects the statistic.
[Section 6] The well-log protocol centers each window on a consensus annotation, so the annotated location is candidate 24 by construction; the paper reports this, but it should be emphasized that the localization exercise is conditional on a changepoint being known to lie in the window.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: main theorems follow from rank invariance and standard permutation validity; efficiency measurements are simulations, not fitted inputs to the theorems.

full rationale

The paper's two central results are self-contained and non-circular. Theorem 1 inherits finite-sample coverage from the standard within-segment permutation argument: under the split hypothesis the two segments are separately exchangeable, the p-value in equation (1) is super-uniform for any deterministic score, and the frozen-weight assumption is used only to keep the score independent of the test window. No fitted parameter enters the proof, so the trained, random, and mistrained network states are not used to force validity. Theorem 2 follows directly from the defining property of ARC scores: each score is a deterministic function of within-segment ranks, and strictly increasing transformations preserve ranks. Coupling the Monte Carlo permutations makes the p-values and hence the confidence sets exactly equal, so the efficiency-transfer statement is an analytic consequence of the construction rather than a fitted prediction. The simulation section reports measured set lengths, coverage, and transfer behavior, but those numbers are not used as assumptions in the proofs. The only acknowledged incompleteness is Proposition 1 in Appendix A.3, whose sketch explicitly states that two technical steps, a sub-Gaussian tail bound and a union bound over Monte Carlo draws, would need to be discharged; that proposition is asymptotic, concerns length contraction, and is not used in Theorems 1 or 2. The paper also explicitly flags the limitations of serial dependence and trend-type alternatives. The one self-citation in the discussion points to a multivariate extension and is not load-bearing. Thus no circular step is present, and the derivation chain is independent of its own conclusions.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The central validity and invariance theorems do not depend on fitted parameters: Theorem 1 holds for any frozen score and Theorem 2 follows from rank invariance. Free parameters matter for reported empirical lengths and for the proposed concrete score, and Proposition 1 rests on two unproved concentration steps.

free parameters (5)

Simulation window n=120, tau=60, m0=10, m1=5, B=79, alpha=0.10
Hand-chosen simulation settings; central theorems hold for any values, but all reported lengths and coverage numbers are specific to these.
Learned convex combination weights w=(w1,w2) = 0.51, 0.49
Fitted by projected gradient descent on synthetic Gaussian training data and frozen; efficiency numbers in Section 5 depend on it.
Neural network g_theta weights (16->64->1) = trained weights, not enumerated
Trained on 6,000 synthetic Gaussian segments, frozen; Theorem 1 holds for any weights, but reported lengths depend on this trained state.
Network training hyperparameters and downsampling to 8 points = lr 1e-3, batch 64, 30 epochs, 16-dim phi
Chosen by hand; define the frozen score whose empirical lengths are reported.
Block length L=10 = 10
Chosen for AR(1) with phi=0.5 to exceed dependence range; coverage repair is only approximate.

assumptions (6)

domain assumption Within-segment exchangeability (piecewise-i.i.d.) of X_1..X_tau ~ F0 and X_{tau+1}..X_n ~ F1
Entered in Section 3.1; it makes p_tau super-uniform and is the basis of Theorem 1. The paper states serial dependence and trends violate it (Appendix B, Remark 1).
standard math Monte Carlo permutation p-values with +1 correction are valid under exchangeability
Invoked in Theorem 1 proof (Appendix A.1), citing Hemerik and Goeman [13] and Lehmann and Romano [21].
domain assumption Score and combination rule are frozen independently of the test window
Required by Theorem 1; Section 3.3 defines type-aware combination frozen before test, and Section 4 states the condition of Theorem 1.
standard math Strictly increasing g preserves all within-segment rank vectors
Lemma 1, used for Theorem 2.
domain assumption Stochastically ordered alternative rho = P(X'>X) != 1/2
Assumed in Proposition 1; pure scale changes may have rho=1/2 and are excluded.
ad hoc to paper Sub-Gaussian tail bound for maximum of standardized rank CUSUMs under permutation and union bound over B draws
Appendix A.3 explicitly says these two steps are not fully proved and would need discharge; Proposition 1 rests on them.

how reviews work

0 comments

Cite this review

Pith. "Pith review of ARC: Augmented-Rank Conformalization for Changepoint Localization --- Finite-Sample Validity and Distribution-Robust Efficiency." pith.science (2026). https://pith.science/paper/ZYEKM3GJ

@misc{pith2026260808424,
  author       = {Pith},
  title        = {Pith review of: ARC: Augmented-Rank Conformalization for Changepoint Localization --- Finite-Sample Validity and Distribution-Robust Efficiency},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZYEKM3GJ}},
  note         = {Machine review of arXiv:2608.08424}
}

read the original abstract

Conformal changepoint localization turns any score into a confidence set for the changepoint with finite-sample coverage. Coverage is universal; efficiency is not. The oracle score is a likelihood ratio, so practical scores estimate density ratios, and set length deteriorates under heavy tails, skewness, and distribution shift, where no length guarantee applies. We propose ARC (Augmented-Rank Conformalization), a family of scores depending on the data only through within-segment ranks: rank-CUSUM location and scale channels, their fixed combinations, and a lightweight neural score frozen after synthetic training. Every ARC score inherits finite-sample coverage for every frozen weight configuration, including random initialization and mistraining. The main result is an efficiency transfer theorem: the entire ARC confidence set is almost surely invariant under strictly increasing marginal transforms, so the set length distribution depends on the data pair only through its rank structure, and lengths certified once hold verbatim across its monotone orbit, whereas a plug-in score's length changes with every re-expression. Across different rank structures lengths do change, and are reported as such. Classical rank-test theory positions ARC as targeting the optimal invariant score at bounded cost. Simulations confirm nominal coverage for all scores, including sabotaged networks, identical sets under monotone transforms where plug-in scores inflate, and smooth degradation where plug-in sets become vacuous; on the well-log benchmark ARC localizes annotated shifts to three to five candidates and flags misfit by an empty set. Two boundaries are stated rather than hidden: serial dependence destroys exactness, and trend-type alternatives lie outside the piecewise-exchangeable model.

Figures

Figures reproduced from arXiv: 2608.08424 by the authors.

**Figure 1.** The ARC mechanism. Each candidate split t is scored through within-segment ranks only; scores are combined by a rule frozen before the test window is seen and calibrated by exact within-segment permutation, yielding the localization confidence set together with the existence test and change-type label. 3.2 The ARC score family Let R = (R1, . . . , Rm) denote the ranks of a segment of length m (no ties, almost surely… view at source ↗

**Figure 2.** Conformal p-value profiles pt on one Gaussian mean-shift window (n = 120, τ = 60, vertical line) and on its exp, cube, and logistic images, with shared permutation randomness. Left: the ARC learned combination, for which the four curves coincide exactly (Theorem 2), so one profile is visible. Right: the plug-in two-sample t score changes with each transform. differ visibly, since each transform presents the plug-in … view at source ↗

**Figure 3.** Mean confidence-set length by marginal distribution and score (values from Table [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Well-log analysis. (a) The full series with the nine consensus annotations (dashed) [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 23 canonical work pages

[1]

Nonparametric Multi Change Point Detection for Markov Chains via Adaptive Clustering

Imon Banerjee et al. Nonparametric multi change point detection for Markov chains via adaptive clustering, 2026. arXiv:2607.12369

work page Pith review arXiv 2026
[2]

Rafal Baranowski, Yining Chen, and Piotr Fryzlewicz. Narrowest-over-threshold detection of multiple change points and change-point-like features.Journal of the Royal Statistical Society: Series B (Statistical Methodology), 81:649–672, 2019. doi: 10.1111/rssb.12322

doi:10.1111/rssb.12322 2019
[3]

Theoretical guarantees for change local- ization using conformal p-values, 2025

Swapnaneel Bhattacharyya and Aaditya Ramdas. Theoretical guarantees for change local- ization using conformal p-values, 2025. arXiv:2510.08749

arXiv 2025
[4]

Richard Savage

Herman Chernoff and I. Richard Savage. Asymptotic normality and efficiency of certain 16 nonparametric test statistics.The Annals of Mathematical Statistics, 29(4):972–994, 1958. doi: 10.1214/aoms/1177706436

arXiv 1958
[5]

ART: Distribution-free and model-agnostic changepoint detection with finite-sample guarantees

Xiaolong Cui, Haoyu Geng, Guanghui Wang, Zhaojun Wang, and Changliang Zou. ART: Distribution-free and model-agnostic changepoint detection with finite-sample guarantees. Journal of the Royal Statistical Society Series B: Statistical Methodology, 2026. doi: 10. 1093/jrsssb/qkag037

work page 2026
[6]

Offline changepoint localization using a matrix of conformal p-values.Transactions on Machine Learning Research, 2026

Sanjit Dandapanthula and Aaditya Ramdas. Offline changepoint localization using a matrix of conformal p-values.Transactions on Machine Learning Research, 2026

work page 2026
[7]

Segmentation and estimation of change-point models: False positive control and confidence regions.The Annals of Statistics, 48:1615– 1647, 2020

Xiao Fang, Jian Li, and David Siegmund. Segmentation and estimation of change-point models: False positive control and confidence regions.The Annals of Statistics, 48:1615– 1647, 2020. doi: 10.1214/19-AOS1861

work page doi:10.1214/19-aos1861 2020
[8]

Changepoint detection in the presence of outliers

Paul Fearnhead and Guillem Rigaill. Changepoint detection in the presence of outliers. Journal of the American Statistical Association, 114(525):169–183, 2019. doi: 10.1080/ 01621459.2017.1385466

arXiv 2019

Show all 43 references

[9]

Multiscale change point inference.Journal of the Royal Statistical Society: Series B (Statistical Methodology), 76:495–580, 2014

Klaus Frick, Axel Munk, and Hannes Sieling. Multiscale change point inference.Journal of the Royal Statistical Society: Series B (Statistical Methodology), 76:495–580, 2014. doi: 10.1111/rssb.12047

2014 doi
[10]

Wild binary segmentation for multiple change-point detection.The Annals of Statistics, 42:2243–2281, 2014

Piotr Fryzlewicz. Wild binary segmentation for multiple change-point detection.The Annals of Statistics, 42:2243–2281, 2014. doi: 10.1214/14-AOS1245

2014 doi
[11]

Narrowest significance pursuit: Inference for multiple change-points in linear models.Journal of the American Statistical Association, 119:1633–1646, 2024

Piotr Fryzlewicz. Narrowest significance pursuit: Inference for multiple change-points in linear models.Journal of the American Statistical Association, 119:1633–1646, 2024. doi: 10.1080/01621459.2023.2211733

2024
[12]

Academic Press, New York, 1967

Jaroslav H´ ajek and Zbynˇ ekˇSid´ ak.Theory of Rank Tests. Academic Press, New York, 1967

1967
[13]

Exact testing with random permutations.TEST, 27: 811–825, 2018

Jesse Hemerik and Jelle Goeman. Exact testing with random permutations.TEST, 27: 811–825, 2018. doi: 10.1007/s11749-017-0571-1

2018 doi
[14]

Conformal changepoint localization, 2026

Rohan Hore and Aaditya Ramdas. Conformal changepoint localization, 2026. arXiv:2602.06267

2026 arXiv
[15]

Testing for a change in mean after changepoint detection.Journal of the Royal Statistical Society: Series B (Statistical Method- ology), 84:1082–1104, 2022

Sean Jewell, Paul Fearnhead, and Daniela Witten. Testing for a change in mean after changepoint detection.Journal of the Royal Statistical Society: Series B (Statistical Method- ology), 84:1082–1104, 2022. doi: 10.1111/rssb.12501

2022 doi
[16]

TUNE: Algorithm-agnostic inference after changepoint detection, 2024

Yinxu Jia, Jingru Liu, Guanghui Wang, Zhaojun Wang, and Changliang Zou. TUNE: Algorithm-agnostic inference after changepoint detection, 2024. arXiv:2409.15676

2024 arXiv
[17]

Kingma and Jimmy Ba

Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In Proceedings of the 3rd International Conference on Learning Representations (ICLR), 2015

2015
[18]

K¨ unsch

Hans R. K¨ unsch. The jackknife and the bootstrap for general stationary observations.The Annals of Statistics, 17(3):1217–1241, 1989. doi: 10.1214/aos/1176347265. 17

1989
[19]

W A VE: Interpretable high-dimensional change point detection via adaptive weighted variable selection.Mathematics, 14(13):2422,

Hui Lan, Luyue Qi, Jianyuan Xue, and Qijing Yan. W A VE: Interpretable high-dimensional change point detection via adaptive weighted variable selection.Mathematics, 14(13):2422,
[20]

Sequential change-point detection with FDR control in reconfigurable sensor networks.IEEE Transactions on Information Theory, 72 (5):3312–3326, 2026

Seungwon Lee, Yunxiao Chen, and Xiaoou Li. Sequential change-point detection with FDR control in reconfigurable sensor networks.IEEE Transactions on Information Theory, 72 (5):3312–3326, 2026. doi: 10.1109/TIT.2026.3669453

2026
[21]

Lehmann and Joseph P

Erich L. Lehmann and Joseph P. Romano.Testing Statistical Hypotheses. Springer, New York, 3rd edition, 2005

2005
[22]

Change-point detection and its modern ap- plications.Annual Review of Statistics and Its Application, 13:421–438, 2026

Jialiang Li, Jingli Wang, and Yuetao Yu. Change-point detection and its modern ap- plications.Annual Review of Statistics and Its Application, 13:421–438, 2026. doi: 10.1146/annurev-statistics-041124-044143

2026 doi
[23]

Robust mean change point testing in high-dimensional data with heavy tails.IEEE Transactions on Information Theory, 72 (1):571–609, 2026

Mengchu Li, Yudong Chen, Tengyao Wang, and Yi Yu. Robust mean change point testing in high-dimensional data with heavy tails.IEEE Transactions on Information Theory, 72 (1):571–609, 2026. doi: 10.1109/TIT.2025.3634207

2026
[24]

Integral probability metric-guided CUSUM-Net for nonparametric changepoint detection.Biometrika, 2026

Yunchen Li, Guanghui Wang, Shuntuo Xu, and Zhou Yu. Integral probability metric-guided CUSUM-Net for nonparametric changepoint detection.Biometrika, 2026. doi: 10.1093/ biomet/asag046

2026
[25]

Optimal nonparametric change point analysis.Electronic Journal of Statistics, 15:1154–1201, 2021

Oscar Hernan Madrid Padilla, Yi Yu, Daren Wang, and Alessandro Rinaldo. Optimal nonparametric change point analysis.Electronic Journal of Statistics, 15:1154–1201, 2021. doi: 10.1214/21-EJS1809

2021 doi
[26]

Alexander M. Mood. On the asymptotic efficiency of certain nonparametric two-sample tests.The Annals of Mathematical Statistics, 25(3):514–522, 1954. doi: 10.1214/aoms/ 1177728719

1954 doi
[27]

Calibrated persistent-laplacian CUSUM for online change-point detection,

Shan Nie et al. Calibrated persistent-laplacian CUSUM for online change-point detection,
[28]

Joseph J. K. ´O Ruanaidh and William J. Fitzgerald.Numerical Bayesian Methods Applied to Signal Processing. Springer, New York, 1996

1996
[29]

Post hoc inference for component attribution in multivariate change-point detection, 2026

Dhia-Elhaq Ouerfelli, Sylvain Arlot, Kevin Bleakley, and Patrick Pamphile. Post hoc inference for component attribution in multivariate change-point detection, 2026. arXiv:2607.14814

2026 arXiv
[30]

Ewan S. Page. Continuous inspection schemes.Biometrika, 41(1–2):100–115, 1954. doi: 10.1093/biomet/41.1-2.100

1954 doi
[31]

Distribution-free changepoint localization after se- quential change detection, 2026

Aytijhya Saha and Aaditya Ramdas. Distribution-free changepoint localization after se- quential change detection, 2026. arXiv:2606.01256

2026 arXiv
[32]

Post-detection inference for sequential changepoint localization.Journal of the Royal Statistical Society Series B: Statistical Methodology, 2026

Aytijhya Saha and Aaditya Ramdas. Post-detection inference for sequential changepoint localization.Journal of the Royal Statistical Society Series B: Statistical Methodology, 2026. doi: 10.1093/jrsssb/qkag069. 18

2026 doi
[33]

Online detection of changes in moment-based projections: When to retrain deep learners or update portfolios?Journal of Machine Learning Research, 27(2):1–50, 2026

Ansgar Steland. Online detection of changes in moment-based projections: When to retrain deep learners or update portfolios?Journal of Machine Learning Research, 27(2):1–50, 2026

2026
[34]

Online change point detection under heavy-tailedness and contamination, 2026

Edwin Yiu Nam Tang, Yudong Chen, Mengchu Li, and Yi Yu. Online change point detection under heavy-tailedness and contamination, 2026. arXiv:2606.09737

2026 arXiv
[35]

Gerrit J. J. van den Burg and Christopher K. I. Williams. An evaluation of change point detection algorithms, 2020. arXiv:2003.06222

2020 arXiv
[36]

Testing exchangeability on-line

Vladimir Vovk, Ilia Nouretdinov, and Alexander Gammerman. Testing exchangeability on-line. InProceedings of the 20th International Conference on Machine Learning, pages 768–775, 2003

2003
[37]

Springer, New York, 2005

Vladimir Vovk, Alexander Gammerman, and Glenn Shafer.Algorithmic Learning in a Random World. Springer, New York, 2005

2005
[38]

Online kernel CUSUM for change-point detection.Journal of the Royal Statistical Society Series B: Statistical Methodology, 2026

Song Wei and Yao Xie. Online kernel CUSUM for change-point detection.Journal of the Royal Statistical Society Series B: Statistical Methodology, 2026. doi: 10.1093/jrsssb/ qkag020

2026 doi
[39]

Individual comparisons by ranking methods.Biometrics Bulletin, 1(6): 80–83, 1945

Frank Wilcoxon. Individual comparisons by ranking methods.Biometrics Bulletin, 1(6): 80–83, 1945. doi: 10.2307/3001968

1945 doi
[40]

Beyond Euclidean summaries: Online change point detection for distribution-valued data, 2026

Yingyan Zeng, Yujing Huang, and Xiaoyu Chen. Beyond Euclidean summaries: Online change point detection for distribution-valued data, 2026. arXiv:2602.07252

2026 arXiv
[41]

Change-aware self-adaptive AI-aided Kalman filters with neural change point detection, 2026

Wenyi Zhang, Xiaoyong Ni, Nir Shlezinger, and Zengfu Wang. Change-aware self-adaptive AI-aided Kalman filters with neural change point detection, 2026. arXiv:2607.13387

2026 arXiv
[42]

Change-point detection for object-valued time series.Journal of Business & Economic Statistics, 44(1):255–269, 2026

Yi Zhang, Changbo Zhu, and Xiaofeng Shao. Change-point detection for object-valued time series.Journal of Business & Economic Statistics, 44(1):255–269, 2026. doi: 10.1080/ 07350015.2025.2520862. 19

2026
[2026]

doi: 10.3390/math14132422

doi

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Nonparametric Multi Change Point Detection for Markov Chains via Adaptive Clustering

Imon Banerjee et al. Nonparametric multi change point detection for Markov chains via adaptive clustering, 2026. arXiv:2607.12369

work page Pith review arXiv 2026

[2] [2]

Rafal Baranowski, Yining Chen, and Piotr Fryzlewicz. Narrowest-over-threshold detection of multiple change points and change-point-like features.Journal of the Royal Statistical Society: Series B (Statistical Methodology), 81:649–672, 2019. doi: 10.1111/rssb.12322

doi:10.1111/rssb.12322 2019

[3] [3]

Theoretical guarantees for change local- ization using conformal p-values, 2025

Swapnaneel Bhattacharyya and Aaditya Ramdas. Theoretical guarantees for change local- ization using conformal p-values, 2025. arXiv:2510.08749

arXiv 2025

[4] [4]

Richard Savage

Herman Chernoff and I. Richard Savage. Asymptotic normality and efficiency of certain 16 nonparametric test statistics.The Annals of Mathematical Statistics, 29(4):972–994, 1958. doi: 10.1214/aoms/1177706436

arXiv 1958

[5] [5]

ART: Distribution-free and model-agnostic changepoint detection with finite-sample guarantees

Xiaolong Cui, Haoyu Geng, Guanghui Wang, Zhaojun Wang, and Changliang Zou. ART: Distribution-free and model-agnostic changepoint detection with finite-sample guarantees. Journal of the Royal Statistical Society Series B: Statistical Methodology, 2026. doi: 10. 1093/jrsssb/qkag037

work page 2026

[6] [6]

Offline changepoint localization using a matrix of conformal p-values.Transactions on Machine Learning Research, 2026

Sanjit Dandapanthula and Aaditya Ramdas. Offline changepoint localization using a matrix of conformal p-values.Transactions on Machine Learning Research, 2026

work page 2026

[7] [7]

Segmentation and estimation of change-point models: False positive control and confidence regions.The Annals of Statistics, 48:1615– 1647, 2020

Xiao Fang, Jian Li, and David Siegmund. Segmentation and estimation of change-point models: False positive control and confidence regions.The Annals of Statistics, 48:1615– 1647, 2020. doi: 10.1214/19-AOS1861

work page doi:10.1214/19-aos1861 2020

[8] [8]

Changepoint detection in the presence of outliers

Paul Fearnhead and Guillem Rigaill. Changepoint detection in the presence of outliers. Journal of the American Statistical Association, 114(525):169–183, 2019. doi: 10.1080/ 01621459.2017.1385466

arXiv 2019