Pith. sign in

REVIEW 3 major objections 4 minor 47 references

Subset simulation's known weaknesses on multimodal or discontinuous failure problems are blamed on its MCMC sampler; swapping in a cyclic component-wise Intrepid kernel restores accurate failure-probability estimates through 1003 dimensions

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 03:58 UTC pith:S5T27WB5

load-bearing objection Worth a careful look: the component-wise cyclic Intrepid sampler is a real contribution and the 1003-dimensional result is striking, but the central causal claim is undercut by the simultaneous switch to single-chain seeding, and the single-run CoV estimates are shakier than the tables suggest. the 3 major comments →

arxiv 2607.22989 v1 pith:S5T27WB5 submitted 2026-07-25 stat.CO stat.ME

SuSIE: Subset Simulation with Intrepid Exploration

Adwait Sharma , Promit Chakroborty , Michael D. Shields This is my paper
classification stat.CO stat.ME MSC 60J2265C05
keywords subset simulationstructural reliabilityMarkov chain Monte Carlomultimodal samplingrare event estimationhigh-dimensional reliabilityIntrepid MCMCoverlapping batch means
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Subset simulation is a standard method for estimating rare failure probabilities in structural engineering, but it has a documented weakness: when failure arises from several disconnected regions, or when the performance function changes sharply, it can miss failure modes and under-report risk. This paper argues that the weakness is not in the subset-simulation decomposition itself but in the random-walk Metropolis sampler normally used to draw from intermediate conditional distributions. To test that, the authors replace that sampler with a modified 'Intrepid' sampler that alternates local Cartesian moves with global moves along constant-probability hyperspheres of the standard-normal parent distribution, cycling these moves with one long chain per subset level. Across four benchmark examples, including a 1003-dimensional snap-through buckling truss, the proposed SuSIE estimator matches a 100-million-sample direct Monte Carlo reference with coefficients of variation far below those of three comparison samplers, all of which miss the dominant failure region in the highest-dimensional case. If the claim holds, engineers can keep the subset simulation framework intact and simply upgrade the sampler.

Core claim

The paper's central claim is that subset simulation is theoretically sound; the product decomposition of the failure probability into conditional probabilities is not the problem. The problem is that standard MCMC kernels cannot sample the multimodal, discontinuous conditional distributions that appear when failure regions are disconnected or performance functions are sharply varying. The proposed SuSIE method replaces the usual modified Metropolis sampler with a cyclic component-wise Intrepid sampler: each step first proposes a global move in hyperspherical coordinates, perturbing the radius and angles of the standard-normal parent, then proposes a local component-wise Cartesian move. Becau

What carries the argument

The central object is the cyclic component-wise Intrepid MCMC transition, p_CI(u' | u) = ∫ p_local(u' | ũ) p_global(ũ | u) dũ, built from two Metropolis-Hastings kernels: a local kernel that updates Cartesian components one at a time against their standard-normal marginals, and a global kernel that updates hyperspherical coordinates (radius and angles) against the marginals of the standard normal in polar form, proposing jumps along constant-probability contours of the parent distribution. Because each component-wise Metropolis-Hastings kernel satisfies detailed balance, the composition of these kernels preserves the target conditional distribution; the cyclic structure makes exploration sys

Load-bearing premise

The load-bearing premise is that, within the finite chain lengths used, the sampler discovers and mixes among every failure mode that contributes materially to the failure probability; the paper's ergodicity argument only guarantees discovery in the limit, and the paper explicitly acknowledges that there is no finite-sample guarantee.

What would settle it

Construct a known two-modal failure domain in, say, ten dimensions with two equal-weight, well-separated modes, run SuSIE at ns = 500 across many independent trials, and count the fraction of trials in which the second mode is never sampled; if that fraction is not small, the claim that the cyclic global move restores exploration in high dimensions is false. A second check targets the variance estimator: compare the single-run overlapping-batch-means coefficient of variation against the empirical coefficient of variation over repeated trials at ns = 500; a persistent understatement, as seen in

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • For structural reliability problems with multimodal or discontinuous limit states, SuSIE yields failure-probability estimates consistent with direct Monte Carlo while requiring orders of magnitude fewer performance-function evaluations.
  • The standard subset simulation framework can be left intact: only the conditional sampler needs to be replaced; no auxiliary tuning run, temperature schedule, or restructuring of the algorithm is required.
  • In the 1003-dimensional snap-through buckling example, modified Metropolis, conditional sampling, and replica exchange all miss the dominant failure region, while SuSIE achieves a coefficient of variation below one even at ns = 500, suggesting the method extends to high-dimensional reliability problems where prior advanced samplers fail.
  • The method costs about twice as many performance-function evaluations per generated sample as standard modified Metropolis, because each global move is followed by a local move; the paper's unit-coefficient-of-variation comparisons contend that this extra cost is repaid by lower variance.
  • Because the local transition can be any valid MCMC kernel, the framework can be combined with other samplers to address further geometric challenges such as extremely narrow conditional distributions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • A natural test of the paper's explanation is to feed SuSIE the extremely narrow conditional-distribution problem with an ensemble or affine-invariant local kernel; if the coefficient of variation drops, the claim that only the local kernel was limiting in that case is corroborated.
  • The hyperspherical global move exploits the spherical symmetry of the standard-normal parent; the same cyclic construction should transfer to other elliptically contoured or transformation-based parent distributions, though the proposal densities would need to be re-derived.
  • If the Markov-chain central limit theorem for the non-reversible kernel is later established, the single-run overlapping-batch-means coefficient of variation could make adaptive stopping feasible; until then, the parenthetical single-run values in small-ns runs should be read as optimistic, as the 1003-dimensional example shows a single-run value of 0.62 versus an empirical value of 0.92.
  • The single-long-chain design trades away the independence across many short chains that made the classical subset-simulation variance estimator consistent, so any variance assessment must use chain-based estimators like overlapping batch means, which places more weight on the unproven central limit theorem.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes SuSIE, a variant of subset simulation in which the MCMC sampler is replaced by a cyclic component-wise version of the recently proposed Intrepid MCMC algorithm. The cyclic sampler alternates a global hyperspherical move with a local Cartesian move, and the method also changes the seed/chain configuration so that each intermediate level uses a single long Markov chain instead of many short chains. The authors claim that the known failures of SuS on problems with multiple/disconnected failure regions or discontinuous performance functions stem from the MCMC sampler rather than the SuS framework, and they support this with numerical comparisons against modified Metropolis (S1), conditional sampling (S2), and replica exchange (S3) on four examples, including a 1003-dimensional snap-through buckling problem. A single-run overlapping batch means (OBM) variance estimator is also proposed. The paper includes a proof of stationarity in Appendix A, an empirical CLT check, and an honest discussion of limitations, including a Black Swan example where the proposed method still performs poorly.

Significance. If the central claim holds, this is a practically useful contribution: it offers a relatively simple, drop-in replacement for the MCMC sampler inside SuS that appears to handle multimodality and sharp performance-function changes without modifying the SuS framework or requiring expensive auxiliary tuning. The numerical study is substantial — four diverse examples, 1000 independent trials per configuration, and 10^8-sample direct Monte Carlo references — and the 1003-dimensional snap-through buckling example is an informative stress test. The paper also carefully reports a limitation (the Black Swan problem) rather than overclaiming universality. However, the strength of the causal claim and the reliability of the reported single-run CoV estimates are both currently undermined by the issues detailed below.

major comments (3)
  1. [§3.1; Tables 1–5] The comparison between SuSIE (S4) and baselines S1–S3 varies two factors at once: the transition kernel and the chain-initiation scheme. Section 3.1 states that the algorithm differs from original SuS in two ways: cyclic component-wise Intrepid MCMC, and a single Markov chain of length ns seeded uniformly from the nsp0 threshold samples, whereas S1–S3 use the original multi-chain configuration. Thus the observed improvement could be due to the seeding scheme, which the paper itself argues improves exploration independently of the Intrepid proposal. To support the central claim in §1.2 that SuS limitations stem from the MCMC sampler, an ablation is needed: e.g., run S1/S2/S3 with the single-chain scheme, or run S4 with the multi-chain scheme. Without this, the causal attribution is not established.
  2. [Appendix A; Eqs. (12), (13), (19)] The transition densities as written are not probability kernels. For example, ∫ p_local(u'|u) du' = ∫ I[G(u')≤b_l] ∏ p_Ui(u'_i|u_i) du' = P(K_0(u,·)∈F_l) < 1, because the component-wise MH candidate can fall outside F_l. The algorithm in Step 2d rejects such moves by staying at the current state, but no corresponding point mass is included in Eq. (12); the same issue affects Eqs. (13) and (19). Consequently, in the Appendix A proof, Eq. (58) takes ∫ p_local(·|u') over the state space to be 1, which is not true. The stationarity proof is therefore invalid as written, although the algorithm itself may be correct once the missing rejection mass is included. The authors should correct the kernels and redo the proof.
  3. [§3.2; Appendix A; Tables 5–6] The parenthetical single-run OBM CoV estimates rely on a Markov chain CLT that is explicitly not proved (Appendix A states the regularity conditions are 'technically involved and therefore not pursued'). The empirical check in Table 7 is not supportive: at ns=10000 the KS statistic is 0.05, above the 5% critical value for 1000 realizations (~0.043), and σ̂_ns oscillates (0.43–0.53) without visible stabilization. The consequence is material: Table 5 (ns=500) shows OBM CoV 0.62 versus empirical CoV 0.92, and Table 6 shows 0.80 versus 1.14. As presented, the parenthetical CoV understates estimator uncertainty. Please either prove the CLT for the proposed non-reversible kernel, provide mixing/CLT diagnostics for the dimensions and chain lengths used (including the 1003-dimensional case), or clearly label the OBM CoV as tentative.
minor comments (4)
  1. [Table 1 header] Typo: 'probaiblity' should be 'probability'.
  2. [§1.1] Duplicate article: 'identify the the highest probability regions'.
  3. [Figure 8 caption] The caption says 'setting U1=U2=2.5 and U4=U5=U6=U7=−2.5', but the text in §4.3 says U5=U6=U7=−2.5 and U3, U4 are free. Please make the slice specification consistent.
  4. [Eq. (16)] The radial proposal q_R is Gaussian on [0,∞), so negative proposals are possible; they are rejected with zero target density. This is acceptable but might be stated explicitly for clarity.

Circularity Check

0 steps flagged

No circularity: the failure-probability predictions are benchmarked against direct MC, while the main weaknesses (confounded sampler/single-chain comparison and unproved CLT) are validity gaps rather than definitional reductions.

full rationale

I found no step in which a claimed prediction reduces by construction to a fitted input or to a self-citation. The SuSIE estimates in Tables 1-6 are obtained by running Algorithm 2 and are compared with an independent 10^8-sample direct MC reference; the intermediate p0 values are standard SuS inputs, and the final-level indicator average is not fitted to the reference. The self-citations to Intrepid MCMC [11] and to the replica-exchange baseline [38] supply ingredients/comparators but are not used as forced uniqueness premises, and the paper's own examples provide independent evidence. The most serious concern is experimental rather than circular: Section 3.1 states SuSIE differs from the original SuS in two ways (the cyclic Intrepid kernel and a single-chain seed scheme), so the headline attribution of SuS failures to the MCMC sampler is confounded. Appendix A also explicitly declines to prove the Markov chain CLT regularity conditions for the non-reversible kernel ('examining these regularity conditions ... is technically involved and therefore not pursued in the present study'), and Table 7's KS statistics are borderline; this weakens the OBM CoV values in Tables 1-6 but is an unverified variance-estimation assumption, not a definitional reduction. Under the required standard, no circular step can be exhibited.

Axiom & Free-Parameter Ledger

4 free parameters · 5 axioms · 0 invented entities

The paper introduces no fitted parameters in its central demonstrations: sigma_R is derived from the chi distribution, gamma_0 and p0 and the local proposal scale are inherited standard choices, and the OBM batch size follows Flegal & Jones. The load-bearing assumptions are the standard modified-Metropolis construction (borrowed from Au & Beck), the unproven Markov chain CLT for the non-reversible cyclic kernel (flagged by the authors), and Harris ergodicity argued from positivity. The mechanically derived benchmark models are domain assumptions. No new particles, forces, or entities are postulated.

free parameters (4)
  • gamma_0 (radial perturbation bound) = 2
    Bounds the global radial proposal Uniform(1/gamma_0, gamma_0) in Eq. (10). Taken from Chakroborty & Shields [11] as a 'recommended value'; not fitted in this paper.
  • p0 (intermediate failure probability) = 0.1
    Standard SuS constant (p0 in [0.1, 0.3] per Zuev et al. [46]); drives the adaptive threshold schedule b_l and the chain length.
  • Local Cartesian proposal scale = 1.0 (unit standard deviation Gaussian per component)
    Eq. (8): component-wise Gaussian proposals with unit variance, following Au & Beck [3]. No adaptation or tuning.
  • OBM batch size s = sqrt(ns)
    Algorithm 3; recommended by Flegal & Jones [15]. Directly controls the single-run CoV estimates in Tables 1–6.
axioms (5)
  • standard math Hyperspherical-coordinate marginal structure of an nd-dimensional standard normal: R ~ chi(nd) with f_R(r) proportional to r^(nd-1) exp(-r^2/2), Theta_i proportional to sin^(nd-i-1)(theta_i), Theta_{nd-1} uniform (Eqs. 14–15).
    Used to define the marginal targets for the component-wise global proposals and to derive sigma_R in Eq. (16). Standard result for the chi distribution and spherical coordinates.
  • domain assumption Component-wise MH kernels satisfying detailed balance w.r.t. the unconditional parent marginals, followed by an indicator acceptance I[G <= b_l], yield a Markov kernel with the correct conditional stationary distribution phi(u|F_l).
    The Au & Beck modified-Metropolis construction, adopted here for both plocal (Eqs. 11–12) and pglobal (Eqs. 17–19). Valid when the pre-indicator kernel is reversible w.r.t. the product density; the paper asserts this by construction and proves it in Appendix A.
  • domain assumption The Markov chain CLT holds for the non-reversible cyclic kernel, justifying the OBM variance estimator (Eq. 27) and hence the single-run CoV values.
    Explicitly not proven: Appendix A states the regularity conditions are 'technically involved and therefore not pursued in the present study' and substitutes a 2D empirical check (Table 7). The OBM-vs-empirical CoV gap at ns=500 in Table 5 (0.62 vs 0.92) shows the assumption is not uniformly safe.
  • domain assumption Harris ergodicity of the cyclic kernel, asserted from positivity of the transition density over F_l.
    Appendix A: 'pIntrepid(u'|u) > 0 for all u',u in F_l which implies that the Markov chain is Harris ergodic.' Plausible via positivity of the component kernels, but positivity alone does not give the geometric ergodicity needed for rapid convergence in high dimension.
  • domain assumption The benchmark models faithfully represent the physical failure modes of interest.
    Examples 2–4: the 2-DOF vibration response, the rigid-bar buckling equilibrium, and the von-Mises truss with 500 series springs (Appendix B, Eqs. 64–78). Reliability results inherit the fidelity of these mechanical models.

pith-pipeline@v1.3.0-alltime-deepseek · 26641 in / 31625 out tokens · 320067 ms · 2026-08-01T03:58:28.510595+00:00 · methodology

0 comments
read the original abstract

This work explores the challenges associated with the subset simulation framework - a well-established algorithm for estimating structural reliability - in settings involving multiple, possibly disconnected regions of failure, or involving discontinuous or sharply changing performance functions. We demonstrate that these drawbacks of subset simulation stem from the limitations of the Markov chain Monte Carlo (MCMC) sampler employed within the method, not from the subset simulation framework itself. Traditional random-walk Metropolis algorithms, which are known to struggle severely with sampling from multimodal distributions, are conventionally applied within subset simulation, which leads to inaccurate failure probability estimates in such cases. In this work, we instead utilize a modified version of the Intrepid MCMC sampler, which has recently been shown to be more effective than vanilla random-walk Metropolis algorithms in sampling from multimodal probability distributions. The subset simulation method with the proposed Intrepid sampler is demonstrated to address these complicating features. Several illustrative examples are considered, ranging from 2 to 1003 dimensions and exhibiting multiple failure regions or highly nonlinear performance functions, including both analytical problems and structural engineering applications.

Figures

Figures reproduced from arXiv: 2607.22989 by Adwait Sharma, Michael D. Shields, Promit Chakroborty.

Figure 1
Figure 1. Figure 1: Example 1: Propagation of conditional samples towards the failure region for [PITH_FULL_IMAGE:figures/full_fig_p011_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Example 1: Histogram of log of the failure probability estimates [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Two-degree-of-freedom mass-spring-damper system under sinusoidal excitation considered in Example 2. [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Example 2: Propagation of samples towards the failure region for [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Example 2: Histogram of log of the failure probability estimates [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Example 3: Eccentrically loaded rigid bars connected by a linear spring with torsional springs at the base. [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Example 3: Load (P1) vs deflection (θ) for various values of eccentricity (ϵ1) for the structure [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Example 3: 2-dimensional slice (U3 vs U4) of the 7-dimensional performance function G(U) setting U1 = U2 = 2.5 and U4 = U5 = U6 = U7 = −2.5. appears that although the mixing of sampler S2 is better than sampler S1, the lack of exploratory steps results in inferior performance compared to the proposed sampler S4 for smaller sample sizes per subset. Of particular interest here is the result for sampler S3. B… view at source ↗
Figure 9
Figure 9. Figure 9: Example 3: Histogram of log of the failure probability estimates [PITH_FULL_IMAGE:figures/full_fig_p018_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Example 4: A von-Mises truss prone to snap-through buckling. [PITH_FULL_IMAGE:figures/full_fig_p019_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Example 4: Load-deflection diagram showing snap-through buckling for the von-Mises truss. [PITH_FULL_IMAGE:figures/full_fig_p020_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Example 4: Histogram of log of the failure probability estimates [PITH_FULL_IMAGE:figures/full_fig_p021_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: CoV ( ¯δ) vs number of performance function calls Ncalls for samplers S1, S2, S3, and S4 for: (a) Example 1: Piece-wise linear function, (b) Example 2: Vibration of a two-DOF system, (c) Example 3: Stable symmetric buckling of rigid, and (d) Example 4: Snap-through buckling of a von-Mises truss. simulation [36, 47, 31, 45, 40, 12] may be leveraged and used along with the proposed sampler. Hence, future wo… view at source ↗
Figure 14
Figure 14. Figure 14: Black Swan problem: Propagation of samples towards the failure region for [PITH_FULL_IMAGE:figures/full_fig_p024_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Comparison of empirical CDF of √ ns  Pˆ l − Pl  and N(0, σˆ 2 ns ) [PITH_FULL_IMAGE:figures/full_fig_p028_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Undeformed and deformed configurations of the von-Mises truss. [PITH_FULL_IMAGE:figures/full_fig_p029_16.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

47 extracted references · 1 linked inside Pith

  1. [1]

    Deep learning-based methods in structural reliability analysis: a review.Measurement Science and Technology, 34(7):072001, 2023

    Sajad Saraygord Afshari, Chuan Zhao, Xinchen Zhuang, and Xihui Liang. Deep learning-based methods in structural reliability analysis: a review.Measurement Science and Technology, 34(7):072001, 2023

    2023

  2. [2]

    Distributed and adaptive darting monte carlo through regenerations

    Sungjin Ahn, Yutian Chen, and Max Welling. Distributed and adaptive darting monte carlo through regenerations. InArtificial Intelligence and Statistics, pages 108–116. PMLR, 2013

    2013

  3. [3]

    Estimation of small failure probabilities in high dimensions by subset simulation

    Siu-Kui Au and James L Beck. Estimation of small failure probabilities in high dimensions by subset simulation. Probabilistic engineering mechanics, 16(4):263–277, 2001. 30 APREPRINT- JULY28, 2026

    2001

  4. [4]

    Application of subset simulation methods to reliability benchmark problems.Structural safety, 29(3):183–193, 2007

    Siu Kui Au, Jianye Ching, and James L Beck. Application of subset simulation methods to reliability benchmark problems.Structural safety, 29(3):183–193, 2007

    2007

  5. [5]

    John Wiley & Sons, 2014

    Siu-Kui Au and Yu Wang.Engineering risk assessment with subset simulation. John Wiley & Sons, 2014

    2014

  6. [6]

    Asymptotic approximations for multinormal integrals.Journal of engineering mechanics, 110(3):357–366, 1984

    Karl Breitung. Asymptotic approximations for multinormal integrals.Journal of engineering mechanics, 110(3):357–366, 1984

    1984

  7. [7]

    The geometry of limit state function graphs and subset simulation: Counterexamples.Reliability Engineering & System Safety, 182:98–106, 2019

    Karl Breitung. The geometry of limit state function graphs and subset simulation: Counterexamples.Reliability Engineering & System Safety, 182:98–106, 2019

    2019

  8. [8]

    Karl Breitung. Sorm, design points, subset simulation, and markov chain monte carlo.ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering, 7(4):04021052, 2021

    2021

  9. [9]

    CRC press, 2011

    Steve Brooks, Andrew Gelman, Galin Jones, and Xiao-Li Meng.Handbook of markov chain monte carlo. CRC press, 2011

    2011

  10. [10]

    The tail wags the distribution: Only sample the tails for efficient reliability analysis.arXiv preprint arXiv:2505.18510, 2025

    Promit Chakroborty and Michael D Shields. The tail wags the distribution: Only sample the tails for efficient reliability analysis.arXiv preprint arXiv:2505.18510, 2025

    Pith/arXiv arXiv 2025

  11. [11]

    Intrepid mcmc: Metropolis-hastings with exploration.Computer Methods in Applied Mechanics and Engineering, 448:118402, 2026

    Promit Chakroborty and Michael D Shields. Intrepid mcmc: Metropolis-hastings with exploration.Computer Methods in Applied Mechanics and Engineering, 448:118402, 2026

    2026

  12. [12]

    Riemannian manifold hamiltonian monte carlo based subset simulation for reliability analysis in non-gaussian space.Structural Safety, 94:102134, 2022

    Weiming Chen, Ziqi Wang, Marco Broccardo, and Junho Song. Riemannian manifold hamiltonian monte carlo based subset simulation for reliability analysis in non-gaussian space.Structural Safety, 94:102134, 2022

    2022

  13. [13]

    Jianye Ching, Siu-Kui Au, and James L Beck. Reliability estimation for dynamical systems subject to stochastic excitation using subset simulation with splitting.Computer methods in applied mechanics and engineering, 194(12-16):1557–1579, 2005

    2005

  14. [14]

    Implementation of machine learning techniques into the subset simulation method.Structural Safety, 79:12–25, 2019

    Fengkun Cui and Michel Ghosn. Implementation of machine learning techniques into the subset simulation method.Structural Safety, 79:12–25, 2019

    2019

  15. [15]

    Flegal and Galin L

    James M. Flegal and Galin L. Jones. Batch means and spectral variance estimators in markov chain monte carlo. The Annals of Statistics, 38(2), April 2010

    2010

  16. [16]

    Adaptive monte carlo augmented with normalizing flows.Proceedings of the National Academy of Sciences, 119(10):e2109420119, 2022

    Marylou Gabrié, Grant M Rotskoff, and Eric Vanden-Eijnden. Adaptive monte carlo augmented with normalizing flows.Proceedings of the National Academy of Sciences, 119(10):e2109420119, 2022

    2022

  17. [17]

    Markov chain monte carlo maximum likelihood

    Charles J Geyer. Markov chain monte carlo maximum likelihood. InComputing Science and Statistics: Proceed- ings of the 23rd Symposium on the Interface, page 156, New York, 1991. American Statistical Association

    1991

  18. [18]

    Practical markov chain monte carlo.Statistical science, pages 473–483, 1992

    Charles J Geyer. Practical markov chain monte carlo.Statistical science, pages 473–483, 1992

    1992

  19. [19]

    Estimating monte carlo variance from multiple markov chains.Scandinavian Journal of Statistics, 2025

    Kushagra Gupta and Dootika Vats. Estimating monte carlo variance from multiple markov chains.Scandinavian Journal of Statistics, 2025

    2025

  20. [20]

    Exact and invariant second-moment code format.Journal of the Engineering Mechanics division, 100(1):111–121, 1974

    Abraham M Hasofer and Niels C Lind. Exact and invariant second-moment code format.Journal of the Engineering Mechanics division, 100(1):111–121, 1974

    1974

  21. [21]

    Improvement of second-order reliability estimates by importance sampling.Journal of Engineering Mechanics, 114(12):2195–2199, 1988

    Michael Hohenbichler and Ruediger Rackwitz. Improvement of second-order reliability estimates by importance sampling.Journal of Engineering Mechanics, 114(12):2195–2199, 1988

    1988

  22. [22]

    Galin L. Jones. On the markov chain central limit theorem.Probability Surveys, 1(none), January 2004

    2004

  23. [23]

    Application of spherical subset simulation method and auxiliary domain method on a benchmark reliability study.Structural Safety, 29(3):194–207, 2007

    LS Katafygiotis and SH Cheung. Application of spherical subset simulation method and auxiliary domain method on a benchmark reliability study.Structural Safety, 29(3):194–207, 2007

    2007

  24. [24]

    Niching subset simulation.Probabilistic Engineering Mechanics, 79:103729, 2025

    Hugh J Kinnear and FA DiazDelaO. Niching subset simulation.Probabilistic Engineering Mechanics, 79:103729, 2025

    2025

  25. [25]

    Mcmc methods for multi-modal distributions

    Krzysztof Łatuszy ´nski, Matthew T Moores, and Timothée Stumpf-Fétizon. Mcmc methods for multi-modal distributions. InHandbook of Markov Chain Monte Carlo, pages 368–396. Chapman and Hall/CRC, 2026

    2026

  26. [26]

    Relaxed subset simulation for reliability estimation.Reliability Engineering & System Safety, 264:111302, 2025

    Binbin Li, Weili Xia, and Zihan Liao. Relaxed subset simulation for reliability estimation.Reliability Engineering & System Safety, 264:111302, 2025

    2025

  27. [27]

    John wiley & sons, 2018

    Robert E Melchers and André T Beck.Structural reliability analysis and prediction. John wiley & sons, 2018

    2018

  28. [28]

    Active learning for structural reliability: Survey, general framework and benchmark.Structural Safety, 96:102174, 2022

    Maliki Moustapha, Stefano Marelli, and Bruno Sudret. Active learning for structural reliability: Survey, general framework and benchmark.Structural Safety, 96:102174, 2022

    2022

  29. [29]

    Determination des distribution don’t les marges sont donnees.Comptes rendus de l’Académie des Sciences, 225:42–43, 1962

    Andre Nataf. Determination des distribution don’t les marges sont donnees.Comptes rendus de l’Académie des Sciences, 225:42–43, 1962. 31 APREPRINT- JULY28, 2026

    1962

  30. [30]

    Mcmc using hamiltonian dynamics.Handbook of markov chain monte carlo, pages 47–95, 2011

    Radford M Neal. Mcmc using hamiltonian dynamics.Handbook of markov chain monte carlo, pages 47–95, 2011

    2011

  31. [31]

    Mcmc algorithms for subset simulation

    Iason Papaioannou, Wolfgang Betz, Kilian Zwirglmaier, and Daniel Straub. Mcmc algorithms for subset simulation. Probabilistic Engineering Mechanics, 41:89–103, 2015

    2015

  32. [32]

    A framework for adaptive MCMC targeting multimodal distributions.The Annals of Statistics, 48(5):2930 – 2952, 2020

    Emilia Pompe, Chris Holmes, and Krzysztof Łatuszy´nski. A framework for adaptive MCMC targeting multimodal distributions.The Annals of Statistics, 48(5):2930 – 2952, 2020

    2020

  33. [33]

    Reliability analysis—a review and some perspectives.Structural safety, 23(4):365–395, 2001

    Rüdiger Rackwitz. Reliability analysis—a review and some perspectives.Structural safety, 23(4):365–395, 2001

    2001

  34. [34]

    Sesc: A new subset simulation method for rare-events estimation.Mechanical Systems and Signal Processing, 150:107139, 2021

    Mohsen Rashki. Sesc: A new subset simulation method for rare-events estimation.Mechanical Systems and Signal Processing, 150:107139, 2021

    2021

  35. [35]

    Remarks on a multivariate transformation.The annals of mathematical statistics, 23(3):470– 472, 1952

    Murray Rosenblatt. Remarks on a multivariate transformation.The annals of mathematical statistics, 23(3):470– 472, 1952

    1952

  36. [36]

    Modified metropolis–hastings algorithm with reduced chain correlation for efficient subset simulation.Probabilistic Engineering Mechanics, 26(2):331–341, 2011

    AM Santoso, KK Phoon, and ST Quek. Modified metropolis–hastings algorithm with reduced chain correlation for efficient subset simulation.Probabilistic Engineering Mechanics, 26(2):331–341, 2011

    2011

  37. [37]

    Benchmark study on reliability estimation in higher dimensions of structural systems–an overview.Structural Safety, 29(3):167–182, 2007

    Gerhart I Schuëller and Helmuth J Pradlwarter. Benchmark study on reliability estimation in higher dimensions of structural systems–an overview.Structural Safety, 29(3):167–182, 2007

    2007

  38. [38]

    Modified replica exchange-based mcmc algorithm for estimation of structural reliability based on particle splitting method.Probabilistic Engineering Mechanics, 72:103448, 2023

    Adwait Sharma and CS Manohar. Modified replica exchange-based mcmc algorithm for estimation of structural reliability based on particle splitting method.Probabilistic Engineering Mechanics, 72:103448, 2023

    2023

  39. [39]

    Sampling variance reduction in structural reliability estimation via sequential mcmc sampling

    Adwait Sharma and CS Manohar. Sampling variance reduction in structural reliability estimation via sequential mcmc sampling. InAdvances in Reliability and Safety Assessment for Critical Systems: Proceedings of the 5th National Conference on Reliability and Safety (NCRS 2022), page 243. Springer Nature, 2023

    2022

  40. [40]

    Subset simulation for problems with strongly non- gaussian, highly anisotropic, and degenerate distributions.Computers & Structures, 245:106431, 2021

    Michael D Shields, Dimitris G Giovanis, and VS Sundar. Subset simulation for problems with strongly non- gaussian, highly anisotropic, and degenerate distributions.Computers & Structures, 245:106431, 2021

    2021

  41. [41]

    Monte carlo and variance reduction methods for structural reliability analysis: A comprehensive review.Probabilistic Engineering Mechanics, 73:103479, 2023

    Chenxiao Song and Reiichiro Kawai. Monte carlo and variance reduction methods for structural reliability analysis: A comprehensive review.Probabilistic Engineering Mechanics, 73:103479, 2023

    2023

  42. [42]

    A review and assessment of importance sampling methods for reliability analysis.Structural Safety, 97:102216, 2022

    Armin Tabandeh, Gaofeng Jia, and Paolo Gardoni. A review and assessment of importance sampling methods for reliability analysis.Structural Safety, 97:102216, 2022

    2022

  43. [43]

    Reliability analysis of complex systems using subset simulations with hamiltonian neural networks.Structural Safety, 109:102475, 2024

    Denny Thaler, Somayajulu LN Dhulipala, Franz Bamer, Bernd Markert, and Michael D Shields. Reliability analysis of complex systems using subset simulations with hamiltonian neural networks.Structural Safety, 109:102475, 2024

    2024

  44. [44]

    Markov chains for exploring posterior distributions.the Annals of Statistics, pages 1701–1728, 1994

    Luke Tierney. Markov chains for exploring posterior distributions.the Annals of Statistics, pages 1701–1728, 1994

    1994

  45. [45]

    Hamiltonian monte carlo methods for subset simulation in reliability analysis.Structural Safety, 76:51–67, 2019

    Ziqi Wang, Marco Broccardo, and Junho Song. Hamiltonian monte carlo methods for subset simulation in reliability analysis.Structural Safety, 76:51–67, 2019

    2019

  46. [46]

    Bayesian post-processor and other enhancements of subset simulation for estimating failure probabilities in high dimensions.Computers & structures, 92:283–296, 2012

    Konstantin M Zuev, James L Beck, Siu-Kui Au, and Lambros S Katafygiotis. Bayesian post-processor and other enhancements of subset simulation for estimating failure probabilities in high dimensions.Computers & structures, 92:283–296, 2012

    2012

  47. [47]

    Modified metropolis–hastings algorithm with delayed rejection

    Konstantin M Zuev and Lambros S Katafygiotis. Modified metropolis–hastings algorithm with delayed rejection. Probabilistic Engineering Mechanics, 26(3):405–412, 2011. 32

    2011