Pith. sign in

REVIEW 2 major objections 5 minor 50 references

Proportional asymptotics of piecewise exponential proportional hazards models

T0 review · 2 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read A scalar saddle point captures the high-dimensional survival model's prediction error.

desk verdict A real first CGMT treatment of a survival model with a fixable proof gap and a likely-typo'd fixed-point equation; deserves refereeing, not desk rejection. read the letter →

arxiv 2501.18995 v1 pith:HD4O3OQY submitted 2025-01-31 math.ST cond-mat.dis-nnstat.TH

Emanuele Massa This is my paper · ORCID
classification math.STcond-mat.dis-nnstat.TH MSC 62N0162J0762E20
keywords high-dimensionalstatisticssurvivalanalysispiecewiseexponentialmodelproportionalhazardsridgeregularizationConvexGaussianMin-Maxtheoremreplicamethodcensoreddata
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

This paper claims that in the proportional asymptotics regime, where $p/n \to \zeta > 0$, the ridge-regularized piecewise exponential proportional hazards model has a complete low-dimensional description. For uncorrelated Gaussian covariates, the paper proves that the optimal penalized log-likelihood converges in probability to the saddle point of a scalar surrogate function, and that the estimator's projections along and orthogonal to the true coefficient vector converge to solutions of self-consistent equations. A reader should care because this turns a high-dimensional survival-fitting problem into a small system that yields exact expected prediction metrics—Harrell's c-index and an integrated Brier score ratio—as functions of $p/n$ and of the ridge penalty. The paper presents this as a rigorous step toward previously heuristic replica-method results for the Cox model.

What carries the argument

The carrying object is the Convex Gaussian Min-Max theorem (CGMT), a comparison principle that transfers Gaussian min-max optimization problems into simpler surrogate problems with the same optimal value and optimizer. Applied after rewriting the objective as a saddle point with a Lagrange multiplier $\varphi$, the theorem leads to a scalar surrogate built from the Moreau envelope of $g(x,\omega,\Delta,T)=\Lambda(T|\omega)e^x-\Delta x$; its proximal operator, expressed through the Lambert W function, closes the self-consistent equations. The number of baseline-hazard parameters $\ell$ is fixed as $n\to\infty$, which is what keeps the limiting problem finite-dimensional.

What would settle it

Set $n=400$, $p=ζn$ for several $ζ$, generate log-logistic survival data with 40% censoring as in Section 5, and trace the empirical minimizer along a ridge path $η\in(0.1,6)$. If the squared norm of $β̂$ or the orthogonal projection $ṥ̂_n$ grows without bound as $ζ$ approaches 1 or $η$ decreases, while equations (17)-(21) still have a finite solution, Theorem 2's convergence claim would fail.

Watch

Extended reading notes

Core claim

The central claim is that for data generated as $Y|X \sim f_0(\cdot|X'\beta_0)$ with $X \sim N(0,I_p)$ and right censoring independent of covariates, the minimum of the penalized piecewise exponential log-likelihood $L_n(\omega,\beta)$ converges in probability to the saddle point of the scalar function $L(\omega,w,v,\varphi,\tau)$ built from a Moreau envelope, and the estimator projections $\beta_0'\hat\beta_n/\|\beta_0\|$, $\|P_{\beta_0^\perp}\hat\beta_n\|$, and $\hat\omega_n$ converge to the unique solution $(w^\star,v^\star,\omega^\star)$ of the self-consistent equations. Because the out-of-sample linear predictor converges in distribution to $w^\star Z_0 + v^\star Q$, quantities such as the c-index and the ideal integrated Brier score can be evaluated exactly from the scalar limit. The proof uses the Convex Gaussian Min-Max theorem to replace the original high-dimensional min-max problem with an asymptotically equivalent scalar process, then identifies that process's saddle point.

Load-bearing premise

The argument depends on the assumption that the true solution of the fitting problem stays inside a fixed bounded region that does not grow with the number of observations; the paper asserts this restriction rather than proving that unconstrained minimizers remain bounded in probability.

Editorial extensions

If this is right

  • For any $ζ=p/n>0$, the expected test c-index is nearly flat in ridge strength $η$, while the value of $η$ that minimizes the integrated Brier score ratio grows with $ζ$, so denser problems need more regularization just to beat the null model.
  • The estimator's behavior splits cleanly: one scalar measures alignment with the true coefficient direction and another measures shrinkage in the orthogonal space, and each has a deterministic limit.
  • Prediction metrics can be computed from the saddle point without fitting the high-dimensional model, making regularization-path comparisons cheap.
  • The scalar limit provides a rigorous benchmark that heuristic replica calculations for the Cox model must reproduce.

Reading between the lines

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

  • If the compactness assumption were proved rather than asserted, the same proof structure would likely extend to correlated Gaussian designs, with the population covariance spectrum entering the self-consistent equations.
  • Because the replica heuristic only requires asymptotic Gaussianity of the linear predictor, the saddle-point characterization may survive for sub-Gaussian covariates; the author leaves this as a conjecture.
  • If the number of intervals $ℓ$ grows with $n$ so that the baseline hazard becomes saturated, the scalar limit might approach a Cox-model limit, but the proof as written requires $ℓ$ fixed and would need an independent argument.
  • The convergence-in-probability result suggests that fixed-point iteration on the self-consistent equations is a reliable computational shortcut; one could test how its accuracy degrades for small $n$ or $ζ$ near zero.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 analyzes a ridge-penalized piecewise exponential proportional hazards model in a high-dimensional regime where p/n tends to a positive constant, with Gaussian uncorrelated covariates and independent censoring. The main results are Theorem 1, which asserts that the normalized minimum of the penalized log-likelihood converges in probability to the saddle point of the scalar surrogate objective L in Eq. (13), and Theorem 2, which asserts that the estimator's projections w_hat_n, v_hat_n and the baseline log-hazard parameters omega_hat_n converge to the solution of the self-consistent equations (17)-(21). From these, the paper derives asymptotic expressions for prediction metrics such as Harrell's c-index and an oracle integrated Brier score. The proof follows the standard CGMT roadmap: the problem is rewritten as a min-max problem, Gordon's comparison is applied, directions are optimized, Moreau envelopes are introduced, and convexity arguments pass to the deterministic limit. Numerical experiments in Section 5 compare the theory with simulations for several values of the aspect ratio and ridge strength. Appendices A-D contain technical lemmas on integrability, derivatives of the Moreau envelope, pointwise convergence of the auxiliary objective, and localization of the minimizer.

Significance. If the proof is completed and the displayed equations are corrected, this would be a useful rigorous contribution: it extends Convex Gaussian Min-Max 'exact asymptotics' from generalized linear models to a parametric survival model with censoring, and it yields explicit, falsifiable predictions for discrimination and calibration metrics. The surrogate objective and the self-consistent equations are derived from CGMT rather than fitted to simulations, and the numerical experiments are an independent check; the paper also provides reproducible code. The main caveats are the unproved compactness/coercivity step in the CGMT application and the mismatch between Eq. (21) and the derivative computation in Proposition 8. Both are localizable and appear fixable, but they currently prevent the theorems, as stated, from being considered established.

major comments (2)
  1. [Section 4 and Theorem 3; Eqs. (12), (26)-(29)] The proof restricts the minimization over beta, xi and the maximization over phi to compact convex sets, with the statement that 'Intuition suggests that if a saddle point exists and the set is sufficiently large, then there is not going to be any difference between the bounded and unbounded problem.' No argument is supplied that the unconstrained minimizers of L_n in Eq. (8), or the optimal dual variables, stay in a fixed compact set with probability tending to one. Since the CGMT in Theorem 3 applies only to compact, convex sets, the comparison in Eqs. (26)-(29) and the limit in Eq. (12) are established only for a compact-constrained problem, not for the actual objective in Theorem 1. The localization part of Theorem 2 inherits this issue: the use of implication (28) and of Eqs. (109)-(111) in Appendix D requires exactly this boundedness-in-probability. A coercivity argument based on lower bounds for min_x g(x, omega, T, Delta) together with the ridge penalties would close the gap, but such an argument is absent from the manuscript.
  2. [Section 3, Theorem 2, Eq. (21)] Equation (21) does not follow from the stationarity condition computed in Proposition 8. Differentiating L in Eq. (13) with respect to omega_k and setting the derivative to zero gives alpha * omega_k = E[Delta * psi_k(T)] - exp(omega_k) * E[Psi_k(T) * exp(xi_hat)], where xi_hat is the proximal point. Solving this equation yields omega_k = (1/alpha) * E[Delta * psi_k(T)] - W0( (1/alpha) * E[Psi_k(T) * exp(xi_hat)] * exp{ (1/alpha) * E[Delta * psi_k(T)] } ). The displayed equation instead uses E[exp(xi_hat) * psi_k(T)] as the coefficient, E[Delta * Psi_k(T)] inside the exponential, and an unexplained factor eta in the denominators. Because Eq. (21) is used in Section 5 to produce the theoretical curves, this discrepancy must be corrected and the derivation shown; if it is a typographical error, the numerical predictions should be re-checked against the corrected equation.
minor comments (5)
  1. [Section 3 and 4] The symbol tau is used both for the user-specified knots of the piecewise exponential model and for the variational parameter tau in the surrogate objective L. This is confusing; consider renaming one of them.
  2. [Appendix C, proof of Proposition 11] In the sentence after Eq. (91), the text reads 'L^{omega,w,v}_n(.) converges in probability to L^{omega,w,v}_n(.)'; the limit should be L^{omega,w,v}(.), not L^{omega,w,v}_n(.).
  3. [Section 3, Theorems 1 and 2] The theorems are stated without the regularity conditions that the appendices actually use, such as finite moments of T, positive mass in each interval, and bounded knots. These conditions should be stated in the main text so that the statements are self-contained.
  4. [Section 5] In the comparison of theory and simulations, the paper says 'because of the convergence in probability established in (2)'; this should refer to Eq. (12) or to Theorem 2, not to Eq. (2).
  5. [Appendix A, Proposition 1 and Proposition 4] The minimum min_x g(x, omega, Delta, T) is used in several places, but for Delta = 0 the infimum is not attained; replace 'min' by 'inf' or add a brief note on the Delta = 0 case.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the surrogate objective and replica-symmetric equations are derived from the CGMT and convex analysis, not fitted to simulations; author self-citations are motivational only.

full rationale

The paper's central claims derive the asymptotic optimal value of the penalized log-likelihood from the Convex Gaussian Min-Max Theorem, reducing the high-dimensional problem to the scalar saddle-point problem L in equation (13). The replica-symmetric equations (17)-(21) are then obtained by differentiating the expected Moreau-envelope surrogate, as in Propositions 7 and 8, and solving the stationarity conditions. No parameter appearing in the surrogate or in the RS equations is fitted to the simulations; the numerical experiments in Section 5 are an independent check against the theoretical curves. The author's prior replica-based papers [44] and [9] are cited only as motivation or as heuristic results to be rigorously justified, not as input to the proof of Theorems 1 and 2. The compactness restriction in Section 4 is a genuine rigor gap, since Theorem 3 requires compact convex sets and the paper asserts without proof that the unconstrained optimizers remain in a fixed compact set; however, this is a correctness or proof-completeness concern, not a circularity. Similarly, the apparent mismatch in equation (21) relative to Proposition 8 is a typographical or correctness issue, not a circular reduction. The derivation chain is therefore self-contained conditional on the cited external CGMT theorem, and the predictions are not equivalent to the inputs by construction.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on Gaussianity, independent censoring, a finite-dimensional piecewise constant baseline, and an unproved compactness restriction. No parameters are fitted to data; zeta, eta, alpha, knots and ell are user-chosen inputs. No new entities are postulated.

free parameters (4)
  • zeta
    Asymptotic ratio p/n, a regime parameter in the limit statement, not fitted to data.
  • eta
    Ridge strength on beta, user-specified in experiments and swept along a path.
  • alpha = 0.01 in simulations
    Ridge strength on baseline hazard parameters, fixed in numerics.
  • ell = 11 intervals in simulations
    Number of piecewise intervals, user-specified and assumed fixed as n grows.
assumptions (5)
  • domain assumption Covariates are Gaussian and uncorrelated, X ~ N(0, I_p).
    Equation (1), throughout; the CGMT proof requires Gaussianity, and the paper explicitly sets uncorrelated design for simplicity.
  • domain assumption Censoring is uninformative and independent of covariates.
    Section 2 assumes Y independent of C given X, and C independent of X, so censoring terms factor out of the loss.
  • domain assumption The true event time depends on covariates only through the linear predictor X' beta0.
    Equation (1), Y|X ~ f0(.|X' beta0); the whole reduction separates beta into parallel and orthogonal components to beta0.
  • domain assumption The baseline hazard is piecewise constant on a finite set of intervals, with ell fixed as n tends to infinity.
    Equations (4)-(7); the paper notes the proof does not extend to the Cox model because ell would grow with n.
  • ad hoc to paper The global minimizers and saddle point variables stay in fixed compact sets independent of n.
    Section 4 restricts the min-max to compact convex sets and asserts no difference from the unbounded problem, but no proof of boundedness in probability is given. This is required for the CGMT comparison.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Proportional asymptotics of piecewise exponential proportional hazards models." pith.science (2026). https://pith.science/paper/HD4O3OQY

@misc{pith2026250118995,
  author       = {Pith},
  title        = {Pith review of: Proportional asymptotics of piecewise exponential proportional hazards models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HD4O3OQY}},
  note         = {Machine review of arXiv:2501.18995}
}
abstract

We study the flexible piecewise exponential model in a high dimensional setting where the number of covariates $p$ grows proportionally to the number of observations $n$ and under the hypothesis of random uncorrelated Gaussian designs. We prove rigorously that the optimal ridge penalized log-likelihood of the model converges in probability to the saddle point of a surrogate objective function. The technique of proof is the Convex Gaussian Min-Max theorem of Thrampoulidis, Oymak and Hassibi. An important consequence of this result, is that we can study the impact of the ridge regularization on the estimates of the parameter of the model and the prediction error as a function of the ratio $p/n > 0$. Furthermore, these results represent a first step toward rigorously proving the (conjectured) correctness of several results obtained with the heuristic replica method for the Cox semi-parametric model.

Figures

Figures reproduced from arXiv: 2501.18995 by the authors.

Figure 1
Figure 1. Simulated data (markers and error bars) against the theory (solid lines). Figures (1a,1b) show the value of [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Simulated data (markers and error bars) against the theory (solid lines): (left) the Test c-index; (right) [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 50 canonical work pages

  1. [1]

    Survival Analysis: Techniques for Censored and Truncated Data

    JP Klein and ML Moeschberger. Survival Analysis: Techniques for Censored and Truncated Data. Statistics for Biology and Health. Springer New York, 2005

    work page 2005

  2. [2]

    The Statistical Analysis of Failure Time Data

    JD Kalbfleisch and RL Prentice. The Statistical Analysis of Failure Time Data. Wiley Series in Probability and Statistics. Wiley, 2011

    work page 2011

  3. [3]

    J Ramjith, C Andolina, T Bousema, and MA Jonker. Flexible time-to-event models for double-interval-censored infectious disease data with clearance of the infection as a competing risk.Frontiers in Applied Mathematics and Statistics, 8, 2022

    work page 2022

  4. [4]

    Survival analysis in infectious disease research: describing events in time

    Cole SR and Michael GH. Survival analysis in infectious disease research: describing events in time. AIDS (London, England), 24, 2010

    work page 2010

  5. [5]

    Regression models and life-tables

    DR Cox. Regression models and life-tables. Journal of the Royal Statistical Society. Series B (Methodological), 34(2):187–220, 1972

    work page 1972

  6. [6]

    Cox’s regression model for counting processes: A large sample study

    PK Andersen and RD Gill. Cox’s regression model for counting processes: A large sample study. The Annals of Statistics, 10(4):1100–1120, 1982

    work page 1982

  7. [7]

    A large sample study of cox’s regression model

    AA Tsiatis. A large sample study of cox’s regression model. The Annals of Statistics, 9(1):93 – 108, 1981

    work page 1981

  8. [8]

    Replica analysis of overfitting in regression models for time-to-event data

    ACC Coolen, JE Barrett, P Paga, and CJ Perez-Vicente. Replica analysis of overfitting in regression models for time-to-event data. Journal of Physics A: Mathematical and Theoretical, 50(37):375001, 2017

    work page 2017

Show all 50 references
  1. [9]

    Replica analysis of overfitting in regression models for time to event data: the impact of censoring

    E Massa, A Mozeika, and ACC Coolen. Replica analysis of overfitting in regression models for time to event data: the impact of censoring. Journal of Physics A: Mathematical and Theoretical, 57(12):125003, 2024

    2024

  2. [10]

    High-Dimensional Statistics: A Non-Asymptotic Viewpoint

    MJ Wainwright. High-Dimensional Statistics: A Non-Asymptotic Viewpoint. Cambridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press, 2019

    2019

  3. [11]

    On robust regression with high-dimensional predictors

    N El Karoui, D Bean, PJ Bickel, C Lim, and B Yu. On robust regression with high-dimensional predictors. Proceedings of the National Academy of Sciences, 110(36):14557–14562, 2013

    2013

  4. [12]

    High-dimensional generalized linear models and the lasso

    SA van de Geer. High-dimensional generalized linear models and the lasso. The Annals of Statistics, 36(2):614 – 645, 2008

    2008

  5. [13]

    Regularization for cox’s proportional hazards model with np-dimensionality

    J Bradic, J Fan, and J Jiang. Regularization for cox’s proportional hazards model with np-dimensionality. The Annals of Statistics, 39(6):3092–3120, 2011

    2011

  6. [14]

    Non-asymptotic oracle inequalities for the high-dimensional cox regression via lasso

    K Shengchun and B Nan. Non-asymptotic oracle inequalities for the high-dimensional cox regression via lasso. Statistica Sinica, 2014

    2014

  7. [15]

    Regression shrinkage and selection via the lasso

    R Tibshirani. Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society. Series B (Methodological), 58(1):267–288, 1996

    1996

  8. [16]

    Statistics for High-Dimensional Data: Methods, Theory and Applications

    P Bühlmann and S van de Geer. Statistics for High-Dimensional Data: Methods, Theory and Applications . Springer Series in Statistics. Springer Berlin Heidelberg, 2011

    2011

  9. [17]

    Weak Convergence and Empirical Processes: With Applications to Statistics

    A Van der Vaart and J Wellner. Weak Convergence and Empirical Processes: With Applications to Statistics . Springer Series in Statistics. Springer New York, 2013

    2013

  10. [18]

    Asymptotic Statistics

    AW Van Der Vaart. Asymptotic Statistics. Asymptotic Statistics. Cambridge University Press, 2000

    2000

  11. [19]

    Empirical Processes in M-Estimation

    S van de Geer. Empirical Processes in M-Estimation. Cambridge Series in Statistical and Probabilistic Mathe- matics. Cambridge University Press, 2000. 9 Proportional asymptotics of piecewise exponential proportional hazards models A PREPRINT

    2000

  12. [20]

    Fundamentals of High-Dimensional Statistics: With Exercises and R Labs

    J Lederer. Fundamentals of High-Dimensional Statistics: With Exercises and R Labs. Springer Texts in Statistics. Springer International Publishing, 2021

    2021

  13. [21]

    The Lasso with general Gaussian designs with applica- tions to hypothesis testing

    Michael Celentano, Andrea Montanari, and Yuting Wei. The Lasso with general Gaussian designs with applica- tions to hypothesis testing. The Annals of Statistics, 51(5):2194 – 2220, 2023

    2023

  14. [22]

    Surprises in high-dimensional ridgeless least squares interpolation

    TJ Hastie, A Montanari, S Rosset, and RJ Tibshirani. Surprises in high-dimensional ridgeless least squares interpolation. Annals of statistics, 50 2:949–986, 2019

    2019

  15. [23]

    High-dimensional asymptotics of prediction: Ridge regression and classifi- cation

    Edgar Dobriban and Stefan Wager. High-dimensional asymptotics of prediction: Ridge regression and classifi- cation. arXiv: Statistics Theory, 2015

    2015

  16. [24]

    The distribution of the Lasso: Uniform control over sparse balls and adaptive parameter tuning

    L Miolane and A Montanari. The distribution of the Lasso: Uniform control over sparse balls and adaptive parameter tuning. The Annals of Statistics, 49(4):2313 – 2335, 2021

    2021

  17. [25]

    Optimal storage properties of neural network models

    E Gardner and B Derrida. Optimal storage properties of neural network models. Journal of Physics A: Mathe- matical and General, 21(1):271, 1988

    1988

  18. [26]

    On the impact of predictor geometry on the performance on high-dimensional ridge-regularized generalized robust regression estimators

    N El Karoui. On the impact of predictor geometry on the performance on high-dimensional ridge-regularized generalized robust regression estimators. Probability Theory and Related Fields, 170:95–175, 2018

    2018

  19. [27]

    A modern maximum-likelihood theory for high-dimensional logistic regression

    P Sur and EJ Candès. A modern maximum-likelihood theory for high-dimensional logistic regression. Proceed- ings of the National Academy of Sciences, 116(29):14516–14525, 2019

    2019

  20. [28]

    High dimensional robust m-estimation: asymptotic variance via approximate message passing

    DL Donoho and A Montanari. High dimensional robust m-estimation: asymptotic variance via approximate message passing. Probability Theory and Related Fields, 166:935–969, 2013

    2013

  21. [29]

    The gaussian min-max theorem in the presence of convexity, 2015

    C Thrampoulidis, S Oymak, and B Hassibi. The gaussian min-max theorem in the presence of convexity, 2015

    2015

  22. [30]

    Precise error analysis of regularized m -estimators in high dimen- sions

    C Thrampoulidis, E Abbasi, and B Hassibi. Precise error analysis of regularized m -estimators in high dimen- sions. IEEE Transactions on Information Theory, 64(8):5592–5628, 2018

    2018

  23. [31]

    Learning curves of generic features maps for realistic datasets with a teacher-student model*

    B Loureiro, C Gerbelot, H Cui, S Goldt, F Krzakala, M Mézard, and L Zdeborová. Learning curves of generic features maps for realistic datasets with a teacher-student model*. Journal of Statistical Mechanics: Theory and Experiment, 2022(11):114001, 2022

    2022

  24. [32]

    Piecewise Exponential Models for Survival Data with Covariates

    M Friedman. Piecewise Exponential Models for Survival Data with Covariates. The Annals of Statistics , 10(1):101 – 113, 1982

    1982

  25. [33]

    Marginal likelihoods based on Cox’s regression and life model

    JD Kalbfleisch and RL Prentice. Marginal likelihoods based on Cox’s regression and life model. Biometrika, 60(2):267–278, 1973

    1973

  26. [34]

    Discussion on professor cox’s paper

    NE Breslow. Discussion on professor cox’s paper. Journal of the Royal Statistical Society: Series B (Method- ological), 34(2):202–220, 1972

    1972

  27. [35]

    Analysis of survival data under the proportional hazards model

    NE Breslow. Analysis of survival data under the proportional hazards model. International Statistical Review / Revue Internationale de Statistique, 43(1):45–57, 1975

    1975

  28. [36]

    Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis

    FE Harrell. Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis. Graduate Texts in Mathematics. Springer, 2001

    2001

  29. [37]

    Time-to-event prediction with neural networks and cox regression.J

    H Kvamme, Ø Borgan, and I Scheel. Time-to-event prediction with neural networks and cox regression.J. Mach. Learn. Res., 20:129:1–129:30, 2019

    2019

  30. [38]

    Continuous and discrete-time survival prediction with neural networks.Lifetime Data Analysis, 2022

    H Kvamme and Ø Borgan. Continuous and discrete-time survival prediction with neural networks.Lifetime Data Analysis, 2022

    2022

  31. [39]

    Convex Analysis

    RT Rockafellar. Convex Analysis. Princeton Landmarks in Mathematics and Physics. Princeton University Press, 1997

    1997

  32. [40]

    On the lambert w function

    RM Corless, GH Gonnet, DEG Hare, DJ Jeffrey, and DE Knuth. On the lambert w function. Advances in Computational mathematics, 5:329–359, 1996

    1996

  33. [41]

    Some inequalities for gaussian processes and applications

    Y Gordon. Some inequalities for gaussian processes and applications. Israel Journal of Mathematics, 50, 1985

    1985

  34. [42]

    Assessing the performance of prediction models: a framework for traditional and novel measures

    EW Steyerberg, AJ Vickers, NR Cook, T Gerds, M Gonen, N Obuchowski, MJ Pencina, and MW Kattan. Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology (Cambridge, Mass.), 2010

    2010

  35. [43]

    Evaluating the Yield of Medical Tests

    FE Harrell, RM Califf, DB Pryor, KL Lee, and RA Rosati. Evaluating the Yield of Medical Tests. JAMA, 247(18):2543–2546, 1982

    1982

  36. [44]

    Penalization-induced shrinking without rotation in high dimensional glm regression: a cavity analysis

    E Massa, MA Jonker, and ACC Coolen. Penalization-induced shrinking without rotation in high dimensional glm regression: a cavity analysis. Journal of Physics A: Mathematical and Theoretical, 55(48):485002, 2022. 10 Proportional asymptotics of piecewise exponential proportional...

    2022

  37. [45]

    Universality of regularized regression estimators in high dimensions.The Annals of Statistics, 51(4):1799 – 1823, 2023

    Q Han and Y Shen. Universality of regularized regression estimators in high dimensions.The Annals of Statistics, 51(4):1799 – 1823, 2023

    2023

  38. [46]

    Universality laws for high-dimensional learning with random features

    H Hu and YM Lu. Universality laws for high-dimensional learning with random features. IEEE Transactions on Information Theory, 69:1932–1964, 2020

    1932

  39. [47]

    Universality of empirical risk minimization

    A Montanari and B Saeed. Universality of empirical risk minimization. ArXiv, abs/2202.08832, 2022. A PROPERTIES OF THE MOREAU ENVELOPE FOR THE PIECEWISE EXPONENTIAL MODEL In the following we assume that ∥ω∥ ≤Cbomega, w, v≤ Cβ and τ >0, furthermore τ1 < τ2 < · · ·< τℓ+1 < ∞. Pr...

    2022 arXiv

  40. [48]

    E h wZ0 + vQ 4i = w4E h Z 4 0 i + v4E h Q4 i is finite,

  41. [49]

    E h g(0, ω, ∆, T) 2i ≤ Pℓ k=1 exp(ωk)(τk+1 − τk) 2 ≤ Pℓ k=1 exp(2ωk) Pℓ k=1(τk+1 − τk)2 which is finite,

  42. [50]

    Ψk(T ) exp(ωk) W0 τ Λ(T |ω)e∆τ +wZ0+vQ τ Λ(T |ω) # ≤ (τk+1 − τk)EZ0,Q

    the term min x g(x, ω, ∆, T) is bounded by a deterministic constant (49), and hence so it is its square. The cross terms can be similarly bounded via the Cauchy-Schwartz inequality. Proposition 3 (FINITE V ARIANCE OF THE DERIV ATIVE) . The random function (ω, w, v, τ) → ˙g(., ...

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.