Pith. sign in

REVIEW 5 major objections 6 minor 28 references

One cut-point phase-type distributions in Reliability. An application to Resistive Random Access Memories

T0 review · 5 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read The paper claims that a phase-type distribution with one unknown cut-point fits RRAM switching data with far fewer phases and passes goodness-of-fit tests that defeat standard phase-type models.

desk verdict A coherent piecewise-PH construction whose empirical claims are overstated: the fit comparison is fair but the parameter-count claim conflates phases with parameters. read the letter →

arxiv 2501.07949 v1 pith:UEHNGOWJ submitted 2025-01-14 stat.ME

classification stat.ME MSC 62N0560J2762F10
keywords onecut-pointphase-typedistributionnon-homogeneousreliabilitymaximumlikelihoodresistiverandomaccessmemoryRRAMvariabilityAnderson-DarlingtestErlang
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 introduces a new lifetime distribution, the one cut-point phase-type (PH) distribution, as a first step toward non-homogeneous phase-type models. The idea is to let the underlying Markov process run with one transition-intensity matrix T1 before an unknown time a and a second matrix T2 after a, so the hazard can change regime at a single point. The authors derive matrix-algebraic formulas for the density, reliability, hazard, cumulative hazard, characteristic function, and moments, and construct a maximum-likelihood estimator. In tests on 1000 resistive-switching cycles from HfO2-based RRAM devices, the cut-point model fits reset voltage with 14 phases where a standard 200-phase Erlang PH is rejected by the Anderson-Darling test, and reset current with 12 phases where a 353-phase PH fails. The paper's claim is that this construction solves the classic PH trade-off between a large number of phases and poor tail fit.

What carries the argument

The defining object is the one cut-point PH distribution with representation (α, T1, T2, a): before time a, the Markov process has intensity matrix T1 and initial distribution α; after a, if absorption has not happened, the transient distribution is α $e^{{T1 a}}$ and the process continues with intensity matrix T2. This gives a piecewise-matrix-exponential reliability function, R(x)=α $e^{{T1 x}}$ e for x ≤ a and R(x)=α $e^{{T1 a}}$ $e^{{T2 (x-a)}}$ e for x > a, which is what lets two Erlang blocks with different rates λ1 and λ2 capture the regime change with very few phases.

What would settle it

Simulate a two-phase abrupt-switch PH with known parameters, run the paper's MLE, and check that the estimated a, λ1, and λ2 recover the truth; then fit the same model to data simulated from a smoothly varying hazard. If the Anderson-Darling p-value stays above 0.05 for the smooth case, the cut-point is merely a flexible approximation, not evidence of a regime change. On the real data, a split-sample test—choosing the phase count on one half of the 1000 cycles and testing fit on the other—would show whether the reported p-values survive honest model selection.

Watch

Extended reading notes

Core claim

The central claim is that allowing one abrupt regime switch in the intensities of a phase-type distribution produces a dramatically more parsimonious and better-fitting model for RRAM switching parameters. For reset voltage, the one cut-point PH with an Erlang structure passes the Anderson-Darling test (p = 0.023) using 14 phases, while a homogeneous PH with 200 phases is rejected (p < 0.0001). For reset current, 12 phases pass (p = 0.141) where 353 phases fail (p = 0.003); set voltage passes with 11 phases (p = 0.0571) versus 89 phases (p = 0.0147); set current passes with 2 phases (p = 0.0819). The paper derives the full set of reliability quantities and the likelihood for the representation (α, T1, T2, a), and implements the estimation in R.

Load-bearing premise

The central premise is that the device's internal Markov intensities are exactly constant up to a single unknown time a and then instantly switch to new constants; if real switching is gradual, multi-regime, or the phase count is picked on the same data, the cut-point model's apparent advantage is partly an artifact of fitting a flexible piecewise-constant shape.

Editorial extensions

If this is right

  • If the model is right, RRAM switching statistics can be summarized by a handful of parameters, making variability simulation and circuit-level analysis practical instead of requiring hundreds of phases.
  • The closed-form hazard and cumulative hazard allow direct computation of failure rates, quantiles, and device reliability over operating ranges.
  • The maximum-likelihood framework means existing PH fitting practice (including EM-style routines) can be adapted to the two-block structure with little extra machinery.
  • Because the same construction works for any non-negative lifetime, other reliability datasets with a visible hazard regime change could be modeled more parsimoniously than with standard PH.

Reading between the lines

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

  • The same matrix-algebraic pattern extends mechanically to multiple cut-points, giving a piecewise-constant PH approximation to any smooth hazard with few phases per segment—an implication the paper does not pursue.
  • The reported Anderson-Darling p-values do not account for phase-count selection on the same data; a hold-out or bootstrap validation would show whether the parsimony advantage survives honest model selection.
  • The estimated cut-point could be given physical meaning as a change in conductive-filament kinetics, and could be tested experimentally by comparing estimates across temperatures or device stacks.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 6 minor

Summary. The paper introduces a "one cut-point phase-type (PH) distribution" as a first step toward non-homogeneous PH distributions. The construction partitions the non-negative time axis at an unknown point a: before a the underlying Markov process has intensity matrix T1, and after a it has intensity matrix T2, with the state distribution at time a carried over via exp(T1 a). The authors derive the density, reliability, hazard, cumulative hazard, characteristic function, and first two moments in matrix-exponential form, and they give a maximum likelihood procedure. The proposed model is then applied to four resistive switching parameters (reset voltage, reset current, set voltage, set current) from RRAM devices, with Anderson-Darling tests used to compare the cut-point PH fit against a standard Erlang PH fit. The central claim is that the cut-point model provides a better fit with fewer phases and fewer parameters.

Significance. If the theoretical and empirical claims were fully established, this would be a useful addition to the reliability toolbox: a tractable, interpretable way to model lifetime data with a regime switch, with explicit formulas for standard reliability quantities and an MLE implementation. The matrix-exponential derivations in Section 3 are coherent and provide a solid starting point for further work on piecewise phase-type models. However, the empirical evidence as presented does not support the paper's main "fewer parameters" claim, and the goodness-of-fit evidence is not calibrated. The theoretical core is worth salvaging, but the paper in its current form overstates what the data show.

major comments (5)
  1. [Abstract and Section 4, Tables 1-4] The assertion that the one cut-point model "decreases the number of parameters" is contradicted by the paper's own tables. Each comparison is between an Erlang PH with a single rate lambda and a cut-point Erlang model with at least the three continuous parameters a, lambda1, and lambda2. Since setting T1 = T2 makes the cut-point family contain the Erlang baseline, the improved in-sample fit is expected from added flexibility alone and is not evidence of a genuine regime change. The paper reports phase counts (14 vs 200, 12 vs 353, etc.) as if they were parameter counts; these are different quantities.
  2. [Section 4, Tables 1-4] The Anderson-Darling p-values are computed after the number of phases and the cut-point a are selected on the same data, so they are not valid goodness-of-fit tests. In particular, for reset voltage the cut-point model has A-D p = 0.023, which is a rejection at the usual 5% level, yet the text describes the fit as satisfactory. The authors should either use a holdout sample, provide selection-corrected p-values via simulation, or explicitly reframe the comparison as exploratory rather than confirmatory.
  3. [Section 3] The MLE section does not discuss identifiability or regularity conditions. The parameters are not identifiable when T1 = T2 or when the cut-point lies outside the support of the data, and no conditions are given under which (a, T1, T2) are uniquely estimable. The reported confidence interval for a is presented without stating the distributional assumptions behind it; standard errors or a bootstrap for all estimated parameters should be provided.
  4. [Section 4, reset voltage subsection] The reset-voltage example is the motivating case (Figure 2), but the cut-point fit's A-D p = 0.023 is below 0.05. The conclusion that the cut-point model "improves significantly" the fit is therefore not supported by the reported test. This failure should be reported honestly and discussed, especially since the same example is used to motivate the entire construction.
  5. [Section 2 and Section 4] The model assumes exact piecewise-constant transition intensities with an instantaneous switch at the single cut-point a, but no sensitivity analysis is provided for this structural assumption. If the true rates vary continuously or change gradually, the estimated a, lambda1, and lambda2 are misspecified, and the improved fit may be an artifact of the piecewise-constant approximation. A diagnostic comparison with a two-cut-point model or a continuously varying rate model would help assess the robustness of the regime-switch interpretation.
minor comments (6)
  1. [Section 4, reset voltage subsection] The text says "In total 4 parameters were estimated" for reset voltage, but only a, lambda1, and lambda2 are listed as the estimated parameters; please clarify what the fourth parameter is or correct the count.
  2. [Section 3] The typeset matrix formulas for the characteristic function and moments appear corrupted in the submitted PDF, making them difficult to verify; please ensure a clean rendering of all matrix-exponential expressions.
  3. [References] Reference [20] is cited for the muhaz package, but the package name, version, and URL are not provided; also, "R-cran" should be written as "CRAN" throughout.
  4. [Throughout] There are several typos and minor wording issues: "bandwitdth" in reference 28, "one cut-points" in Table 2's caption, "HfAIO" versus "HfAlO" in the Abbreviations list, and "John Hopkins University Press" should be "Johns Hopkins University Press."
  5. [Data Availability] The data availability statement says data are available from the corresponding author on reasonable request; given that the methodology is claimed to be implemented in R, a public repository with code and data would strengthen reproducibility.
  6. [Section 1] The paper cites the inhomogeneous PH distributions of Albrecher and Bladt [15] as related work but does not compare the proposed one cut-point PH class with that existing framework; a brief comparison would help position the contribution.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the one cut-point PH distribution is defined constructively, and its properties and likelihood are derived directly from that definition.

full rationale

The derivation chain is self-contained. The one cut-point PH density is defined constructively in Section 3 from two PH generators T1, T2 and a cut-point a; reliability, hazard, cumulative hazard, characteristic function, moments and the likelihood are then obtained by direct matrix-exponential calculus from this definition (Sections 2 and 3), with no fitted quantity reintroduced as an output. The cited earlier PH-RRAM papers [11,12,13,25] are used only to motivate the model and to supply comparative context; they are not used to define the distribution, to derive its properties, or to justify a uniqueness claim, so the self-citations are not load-bearing. The empirical comparison in Section 4 fits parameters by maximum likelihood and reports in-sample Anderson-Darling p-values after data-dependent phase and cut-point selection; while this raises statistical calibration concerns, it is not a circular reduction of a prediction to its inputs. No equation in the paper is equivalent to another by construction, and no fitted parameter is renamed as a prediction. The claimed parameter-count advantage is questionable because the cut-point model has more free parameters than the one-parameter Erlang baseline, but that is a correctness issue, not a circularity. Overall, the paper's derivation is independent of its conclusions.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The model rests on standard PH assumptions plus a piecewise-constant regime assumption and an imposed Erlang structure. All estimated quantities, cut-point and phase counts are fitted or selected from the same data, and no independent physical entity is introduced.

free parameters (3)
  • cut-point a = Reset voltage 0.595 [0.571,0.619]; reset current 0.0072 [0.0068,0.0076]; set voltage 0.315 [0.296,0.334]; set current…
    Estimated by maximum likelihood in each application. It defines the regime switch and appears in every formula for the distribution, moments and likelihood.
  • per-period Erlang rates lambda1 and lambda2 = Reset voltage 16.74531 and 261.61844; reset current 1003.27 and 9652.37; set voltage 11.5570 and 73.7963; set current…
    Estimated by MLE for T1 and T2. These two intensities replace the single rate of the standard Erlang PH fit and are the core extra flexibility of the model.
  • number of phases in each Erlang block = Reset voltage 14; reset current 12; set voltage 11; set current 2
    The number of phases is selected by an 'exhaustive analysis' on the same data that are later tested with Anderson-Darling. This is a model-selection degree of freedom, not a parameter with an independent criterion.
assumptions (5)
  • domain assumption The device's internal behavior is a Markov process over n transient states plus an absorbing state, so the time to event is phase-type distributed.
    Section 2 assumes this to define T and T^0, inheriting the standard PH modeling framework from Neuts. It is a physical modeling assumption about RRAM switching, not a proven fact.
  • ad hoc to paper Intensities are constant on [0,a) and [a,infinity), switching instantaneously at the single cut-point a.
    This is the defining construction of the paper. No physical or empirical evidence is given for an abrupt switch at one point, as opposed to gradual or multiple changes.
  • ad hoc to paper In the applications, T1 and T2 have Erlang structure and the initial vector alpha is (1,0,...,0).
    Section 4 says this structure is assumed to simplify the model. It reduces the parameter count but is not derived from RRAM physics or data.
  • domain assumption The 1000 switching cycles are independent and identically distributed observations.
    The likelihood in Section 3 requires i.i.d. data. Cycle-to-cycle variability may include serial dependence over the 1000 cycles, which is not tested.
  • ad hoc to paper Standard MLE regularity and Anderson-Darling test validity hold after model selection.
    The cut-point likelihood is discontinuous in a because observations switch groups, and phase counts are chosen on the same data. No regularity proof or adjustment for model selection is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of One cut-point phase-type distributions in Reliability. An application to Resistive Random Access Memories." pith.science (2026). https://pith.science/paper/UEHNGOWJ

@misc{pith2026250107949,
  author       = {Pith},
  title        = {Pith review of: One cut-point phase-type distributions in Reliability. An application to Resistive Random Access Memories},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UEHNGOWJ}},
  note         = {Machine review of arXiv:2501.07949}
}
read the original abstract

A new probability distribution to study lifetime data in reliability is introduced in this paper. This one is a first approach to a non-homogeneous phase-type distribution. It is built by considering one cut-point in the non-negative semi-line of a phase-type distribution. The density function is defined and the main measures associated, such as the reliability function, hazard rate, cumulative hazard rate and the characteristic function are also worked out. This new class of distributions enables to decrease the number of parameter in the estimate when inference is considered. Besides, the likelihood distribution is built to estimate the model parameters by maximum likelihood. Several applications by considering Resistive Random Access Memories compare the adjustment when phase type distributions and one cut-point phase-type distributions are considered. The developed methodology has been computationally implemented in R-cran.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 20 canonical work pages

  1. [1]

    Lawless, J. F. Statistical Models and Methods for Lifetime Data ; Hoboken, NJ, USA: John Wiley & Sons, Inc. (Wiley Series in Probability and Statistics), 2002. doi: 10.1002/9781118033005

  2. [2]

    Coolen, F. P. A. Parametric Probability Distributions in Relia bility. In Encyclopedia of Quantitative Risk Analysis and Assessment; Chichester, UK: John Wiley & Sons, Ltd, 2008. doi: 10.1002/9780470061596.risk0466

  3. [3]

    D.; Prentice, R

    Kalbfleisch, J. D.; Prentice, R. L. The Statistical Analysis of Failure Time Data; Hoboken, NJ, USA: John Wiley & Sons, Inc. (Wiley Series in Probability and Statistics), 2002. doi: 10.1002/9781118032985

  4. [4]

    Analysis of Multivariate Survival Data; New York, NY: Springer New York (Statistics for Biology and Health), 2000

    Hougaard, P. Analysis of Multivariate Survival Data; New York, NY: Springer New York (Statistics for Biology and Health), 2000. doi: 10.1007/978-1-4612-1304-8

  5. [5]

    Meeker, W. Q. ; Escobar, L. A. ; Pascual, F. G. Statistical Methods for Reliability Data (2nd Ed.); Wiley Series in Probability and Statistics, 2021

  6. [6]

    Elmahdy, E. E. A new approach for Weibull modeling for reliability life data analysis. App. Math. Comput. 2015, 250, 708–720. doi: 10.1016/j.amc.2014.10.036

  7. [7]

    K.; Lemonte, A

    Shakhatreh, M. K.; Lemonte, A. J.; Moreno–Arenas, G. The log-normal modified Weibull distribution a nd its reliability impli- cations. Reliab. Eng. Syst. Safe. 2019, 188, 6–22. doi: 10.1016/j.ress.2019.03.014

  8. [8]

    Matrix geometric solutions in stochastic models

    Neuts, M.F. Matrix geometric solutions in stochastic models. An algorithmic approach ; Baltimore: John Hopkins University Press, 1981

Show all 28 references
  1. [9]

    Ruin probabilities; World Scientific, Chinese, 2000

    Asmussen, S. Ruin probabilities; World Scientific, Chinese, 2000

  2. [10]

    Fitting phase-type distributions via EM-algorithm

    Asmussen, S.; Nerman, O. Fitting phase-type distributions via EM-algorithm. Scand. J. Stat. 1996, 23 (4), 419–441. Available at: https://www.jstor.org/stable/4616418

  3. [11]

    Phase-type distributions for studying variability in resistive memories

    Acal, C.; Ruiz-Castro, J.E.; Aguilera, A.M.; Jiménez-Molinos, F.; Roldán, J.B. Phase-type distributions for studying variability in resistive memories. J. Comput. Appl. Math. 2019, 345, 23-32. doi: 10.1016/j.cam.2018.06.010

  4. [12]

    E.; Alonso, F

    Pérez, E.; Maldonado, D.; Acal, C.; Ruiz-Castro, J. E.; Alonso, F. J.; Aguilera, A. M.; Jiménez-Molinos, F.; Wenger, C.; Roldán, J. B. Analysis of the statistics of device-to-device and cycle-to-cycle variability in TiN/Ti/Al:HfO2/TiN RRAMs. Microelectron. Eng. 2019, 214, 104–...

  5. [13]

    E.; Aguilera, A

    Pérez, E.; Maldonado, D.; Acal, C.; Ruiz-Castro, J. E.; Aguilera, A. M.; Jiménez-Molinos, F.; Roldán, J. B.; Wenger, C. Advanced temperature dependent statistical analysis of forming voltage distributions for three differen t HfO2-based RRAM technolo- gies. Solid State Electro...

  6. [14]

    F.; Samorodnitsky, G

    Bladt, M.; Nielsen, B. F.; Samorodnitsky, G. Calculation of ruin probabilities for a dense class of heavy tailed distribution s. Scand. Actuar. J. 2015, 2015 (7), 573-591. doi: 10.1080/03461238.2013.865257

  7. [15]

    Inhomogeneous phase -type distributions and heavy tails

    Albrecher, H.; Bladt, M. Inhomogeneous phase -type distributions and heavy tails. J. Appl. Probab. 2019, 56 (4), 1044-1064. doi: 10.1017/jpr.2019.60

  8. [16]

    H.; Zenga, M

    Marshall, A. H.; Zenga, M. Experimenting with the Coxian phase -type distribution to uncover suitable fits. Methodol. Comput. Appl. 2012, 14, 71-86. doi: 10.1007/s11009-010-9174-y

  9. [17]

    C.; Wiper, M

    Ausin, M. C.; Wiper, M. P.; Lillo, R. E. Bayesian estimation for the M/G/1 queue using a phase-type approximation. J. Stat. Plan. Infer. 2004, 118, 83-101. doi: 10.1016/s0378-3758(02)00398-1

  10. [18]

    He, Q. M. Fundamentals of matrix-analytic methods (Vol. 365); New York, Springer, 2014

  11. [19]

    Cluster-based fitting of phase-type distributions to empirical data

    Reinecke, P.; Kraub, T.; Wolter, K. Cluster-based fitting of phase-type distributions to empirical data. Comput. Math. Appl. 2012, 64 (12), 3840-3851. doi: 10.1016/j.camwa.2012.03.016. Mathematics 2021, 9, x FOR PEER REVIEW 12 of 12

  12. [20]

    G.; Wang, J

    Mueller, H. G.; Wang, J. L. Hazard rate estimation under random censoring with varying kernels and bandwidths. Biometrics 1994, 50, 61-76

  13. [21]

    Resistive switching: from fundamentals of nanoionic redox processes to memristive device applications ; Wiley-VCH, 2015

    Ielmini, D.; Waser, R. Resistive switching: from fundamentals of nanoionic redox processes to memristive device applications ; Wiley-VCH, 2015

  14. [22]

    Utilizing the variability of resistive random access memory to implement reconfigurable physical unclonable func- tions

    Chen, A. Utilizing the variability of resistive random access memory to implement reconfigurable physical unclonable func- tions. IEEE Electr. Device L. 2015, 36 (2), 138-140. doi: 10.1109/LED.2014.2385870

  15. [23]

    Stochastic memory devices for security and computing

    Carboni, R.; Ielmini, D. Stochastic memory devices for security and computing. Adv. Electron. Mater. 2019, 5, 1900198. doi: 10.1002/aelm.201900198

  16. [24]

    B.; García, H.; Campabadal, F.; Dueñas, S.; Castán, H.; Jiménez-Molinos, F.; Roldán, J

    González-Cordero, G.; González, M. B.; García, H.; Campabadal, F.; Dueñas, S.; Castán, H.; Jiménez-Molinos, F.; Roldán, J. B. A physically based model for resistive memories including a detailed temperature and variability description. Microelectron. Eng. 2017, 178, 26-29. doi...

  17. [25]

    E.; Acal, C.; Aguilera, A.M.; Roldán, J

    Ruiz-Castro, J. E.; Acal, C.; Aguilera, A.M.; Roldán, J. B. A complex model via phase -type distributions to study random tele- graph noise in resistive memories. Math. 2021, 9, 390. doi: 10.3390/math9040390

  18. [26]

    H.; Lu, P.; Nocedal, J.; Zhu, C

    Byrd, R. H.; Lu, P.; Nocedal, J.; Zhu, C. A limited memory algorithm for bound constrained optimization. SIAM J. Sci. Comput. 1995, 16, 1190-1208. doi: 10.1137/0916069

  19. [27]

    Silverman, B. W. Density estimation; London: Chapman and Hall, 1986

  20. [28]

    J.; Jones, M

    Sheater, S. J.; Jones, M. C. A reliable data -based bandwitdth selection method for kernel density estimation. J. Roy. Stat. Soc. B 1991, 53, 683-690

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.