Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Inexact Proximal Point Algorithms for Zeroth-Order Global Optimization

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Sample-based proximal estimates can provably drive iterates to the global minimum

desk verdict A genuinely new theoretical result on global convergence of inexact proximal point methods, but the implemented TT- and MC-IPP algorithms do not meet the conditions the theorems require – worth peer review, needs a revision aligning claims with hypotheses. read the letter →

arxiv 2412.11485 v4 pith:Y7NNZJDP submitted 2024-12-16 math.OC

Minxin Zhang , Fuqun Han , Yat Tin Chow , Stanley Osher , Hayden Schaeffer This is my paper · ORCID
classification math.OC MSC 49M3765K0590C2690C56
keywords globaloptimizationzeroth-orderinexactproximalpointmethodGibbsmeasuretensortrainapproximationMonteCarlointegrationnonconvexderivative-free
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 tries to establish that a proximal point iteration can find the global minimum of a continuous nonconvex function even when the proximal operator is only estimated approximately, provided the squared estimation errors are summable over iterations. Its central theorem (Theorem 1) states that under mild coercivity and non-oscillation assumptions, the values $f(x_k)$ converge to $f_{\min}$ for deterministic inexact estimates satisfying $\sum_{k=0}^\infty \|y_k-\hat{x}_k\|^2<\infty$; Theorem 2 gives the same almost-sure conclusion for stochastic estimates satisfying a Borel-Cantelli-type tail condition. The paper also proves that the Gibbs-measure expectation converges to the proximal point at rate $O(\delta)$ as $\delta\to 0^+$, which justifies approximating proximal operators by sampling. If the results hold, global zeroth-order optimization becomes achievable with only approximate, sample-based proximal operators, and the two proposed algorithms, TT-IPP and MC-IPP, instantiate this in practice.

What carries the argument

The mechanism is the inexact proximal iteration with adaptive regularization: when consecutive iterates barely move, the parameter $t_k$ is increased toward its ceiling $T$, and this ceiling is what lets the Moreau-envelope argument push cluster points below $f_{\min}+\mu$; when the iterates move more, $t_k$ is decreased to encourage local refinement. To evaluate the proximal operator without gradients, the paper uses the Gibbs-measure identity $\mathrm{prox}^{\delta}_{tf}(x)=\int z\exp(-\varphi(z)/\delta)\,dz\,/\,\int \exp(-\varphi(z)/\delta)\,dz$, where $\varphi(z)=f(z)+\|z-x\|^2/(2t)$; the quadratic term and small $\delta$ concentrate the measure at the proximal point. The tensor-train variant builds a low-rank TT cross approximation of the discretized function $\exp(-f/\delta)$ on a mesh, with cost linear in the dimension, and the Monte Carlo variant samples from a Gaussian centered at $x$ with variance $\delta t I_d$ and applies an exponentially weighted moving average for variance reduction.

What would settle it

Implement TT-IPP or MC-IPP on a coercive continuous nonconvex function with multiple local minima using the Section 6 parameter values ($h=0.1$, $\gamma=1.1$ for TT-IPP; $C=1.1$, $c=0.9$ for MC-IPP) and record $\sum_{k=0}^\infty \|y_k-\hat{x}_k\|^2$; if this sum diverges while $f(x_k)$ still approaches $f_{\min}$, the observed convergence is not explained by Theorem 1's hypotheses, and if the iterates stall at a local minimizer whenever the sum diverges, the practical algorithms fail exactly where the theory's condition is violated.

Watch

Extended reading notes

Core claim

The central claim is that global minimization does not require exact proximal operators: an inexact proximal point iteration $x_{k+1}=\alpha_k y_k+(1-\alpha_k)x_k$, with $y_k$ an estimate of the true proximal point $\hat{x}_k\in\mathrm{prox}_{t_k f}(x_k)$, converges to the global minimum whenever the errors obey $\sum_{k=0}^\infty \|y_k-\hat{x}_k\|^2<\infty$ and the damping and regularization parameters satisfy $\alpha_{\min}>1-\eta_-$ and a sufficiently large ceiling $T$. Under Assumptions 1-3 (continuity, $p$-coercivity, and a no-flat-oscillation condition near the minimizer), Theorem 1 shows $\lim_{k\to\infty} f(x_k)=f_{\min}$; the proof controls accumulated errors, forces $t_k$ to its ceiling $T$, and uses a Moreau-envelope argument with $T>M/\mu$ to show every cluster point is a global minimizer. Theorem 2 extends the conclusion to stochastic estimates under the tail condition $\sum_k p_k<\infty$, with convergence almost surely. The paper further derives the $O(\delta)$ convergence rate of the Gibbs-measure estimator and the tensor-train estimation error bounds, and claims this is the first IPP method proved to converge to a global minimizer rather than only a stationary point.

Load-bearing premise

The guarantee rests on the cumulative squared errors of the approximate proximal estimates being finite, and the implemented TT and MC parameter choices do not verify the mesh-size and TT-error conditions (31) that would ensure this in high dimensions.

Editorial extensions

If this is right

  • Any deterministic zeroth-order estimator whose proximal errors are square-summable gives global convergence for coercive continuous functions satisfying Assumptions 1-3, so local minima are escaped by the IPP dynamics alone.
  • Stochastic estimators satisfying (13) converge almost surely, so adaptive sample-size schedules that shrink errors fast enough inherit the guarantee.
  • TT-IPP reaches the proven regime when the mesh spacing and TT error obey (31), and the experiments show function-evaluation cost growing nearly linearly with dimension on low-rank-structured integrands.
  • MC-IPP inherits the guarantee when the sample size grows sufficiently, and the experiments show it beats the compared solvers on most benchmarks under a fixed evaluation budget.
  • The $O(\delta)$ Gibbs estimator rate makes decreasing $\delta$ while keeping errors square-summable a valid route to global minimization, and the warm start (32) costs essentially one iteration.

Reading between the lines

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

  • If the summability condition is the true dividing line, then any sampling scheme whose squared error decays fast enough—adaptive MCMC, quasi-Monte Carlo, or learned surrogates—could be dropped into the same IPP framework; TT-IPP and MC-IPP are only the first two instantiations.
  • The mesh condition $h=O(\delta^{(d+2)/(2s)+1})$ in Corollary 2 is impractical at high dimension, and the experiments' fixed $h=0.1$ with $\gamma=1.1$ lie outside the proven regime; the paper's own heuristic choices therefore implicitly assume the worst-case quadrature bounds are pessimistic.
  • For functions where $\exp(-f/\delta)$ is approximately low-rank across iterations, warm-starting the TT approximation through Hadamard products could make subsequent proximal evaluations nearly free, pointing toward higher-dimensional global optimization than the current experiments.
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

3 major / 6 minor

Summary. The manuscript develops an inexact proximal point (IPP) framework for zeroth-order global minimization of nonconvex functions with a unique global minimizer. Under Assumptions 1–3, Theorem 1 proves convergence of the deterministic IPP iterate values to the global minimum when the proximal-estimation errors satisfy the summability condition (4), and Theorem 2 gives an almost-sure analogue under the probabilistic condition (13). Section 3 relates the proximal operator to the mean of a Gibbs measure and proves convergence of that mean to the proximal point as δ→0+, with an O(δ) rate under a nondegeneracy condition. Section 4 introduces tensor-train (TT) approximations of the Gibbs integrand, using the randomized TT cross algorithm, and provides error bounds for the resulting proximal estimates in Proposition 2 and Corollary 2. Section 5 defines two practical algorithms, TT-IPP and MC-IPP, with adaptive refinement and sampling schedules, and claims their convergence as corollaries of Theorems 1 and 2. Section 6 reports experiments on benchmark functions and two applications, showing competitive accuracy and function-evaluation counts for the proposed methods.

Significance. If the theoretical claims are fully established, the paper would provide the first convergence guarantee of an inexact proximal point method to a global minimizer under only summability of estimation errors, and it offers a coherent Gibbs-measure justification for sampling-based proximal estimates. The TT-based estimator is a genuine algorithmic contribution: it exploits Sobolev smoothness and low-rank structure to reduce the cost of the discretized integrals, and the code is publicly available. The experimental section is broad, covering many benchmarks and dimensions up to 100, with TT-IPP matching or beating several standard global solvers. However, the central link between the implemented algorithms and the theorem hypotheses is the weakest point: the sufficient conditions (4) and (13) are not verified for the parameter schedules actually used in the experiments, and the paper explicitly admits that the mesh requirement behind Corollary 3 is unrealistic in high dimensions. The proof of the deterministic theorem is essentially correct but contains a circularity in the choice of T, and the stochastic theorem inherits the same issue in a sample-path-dependent form.

major comments (3)
  1. [Section 2, Theorem 1 and Appendix A.1] The choice of T via (12) and (37) is circular. The bound M in (9) is derived after T is fixed, yet (12) requires T > M/μ, so no argument shows that a single T can be chosen in advance. In Theorem 2, M is sample-path-dependent, making (37) a condition on a random constant rather than a verifiable parameter choice. The proof needs either a uniform-in-T bound on M, an explicit construction of T, or a reformulation as an existence statement. This is load-bearing because the entire convergence proof relies on the stationarity of the limit point via (11)-(12).
  2. [Section 5.1, Corollary 3 and Algorithm 3] The implemented TT-IPP schedule does not satisfy the sufficient conditions used to invoke Theorem 1. Corollary 3 requires γ > (d+2)/(2s)+1 and, via (31), h_k = O(δ_k^{(d+2)/(2s)+1}) and ε_TT ≤ exp(-φ_min/δ_k) at every iteration. Algorithm 3 uses h0=0.1, γ=1.1, C=10, and refines only when h_k > C δ_k^γ, so h_k tracks C δ_k^{1.1} rather than the required power; for d=100 and s=2 the required exponent is about 26.5. The paper itself states after (29) that such a mesh bound is unrealistic, and no numerical or analytic evidence is provided that (4) holds for the tested parameter values. Consequently Theorem 1 does not currently prove global convergence of TT-IPP as implemented.
  3. [Section 5.2, Corollary 4 and Algorithm 4] The proof of Corollary 4 asserts that choosing C > 1 sufficiently large makes condition (13) hold, but the implemented values C=1.1, c=0.9, α_min=0.2, p=0.8 are not accompanied by any verification of the implied constants. The MC variance analysis preceding (35) is derived for a fixed δ, fixed sample size N, and no rejection step; Algorithm 4 adaptively changes δ, N, and α, and Line 7 introduces probability-p rejection with resampling, which the standard MC covariance computation at the top of Section 5.2 does not cover. Thus condition (13) of Theorem 2 is not established for MC-IPP as implemented, and the almost-sure convergence claim for the practical algorithm remains unsupported.
minor comments (6)
  1. [Section 2, Theorem 1 proof] After Eq. (5), 'inequity' should be 'inequality', and later in the proof 'arbitray κ' should be 'arbitrary κ'.
  2. [Definition 1] The subdifferential ∂f(x) is stated as a subset of R^n but should be R^d to match the domain of f.
  3. [Corollary 1 and Appendix A.5] The summation limits in the definition of z̄* are written as 'j−1' instead of 'j=1' in both places; this typo makes the formula unintelligible.
  4. [Algorithm 2] The while-loop condition uses '<' for the error test; presumably the loop should continue while the error exceeds τ_stop, so the inequality direction should be '>' or the text should clarify the stopping convention.
  5. [Appendix A.1] In the proof of Theorem 2, the sentence 'Combining with (4)' after Eq. (36) should refer to the almost-sure consequence of (13) together with the Borel–Cantelli argument, since (4) is not an assumption of Theorem 2.
  6. [Tables 2, 3, and 6] There are several typographical errors in the table entries, including 'Brwon R10' (should be 'Brown R10') and 'Alphine 1' (should be 'Alpine 1').

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the convergence theorems are conditional on explicit error-summability hypotheses and are proved from stated assumptions; the practical parameter-regime gap flagged by the paper is a verification issue, not a circular reduction.

full rationale

The derivation chain is not circular. Theorem 1 is an implication from Assumptions 1-3, alpha_min > 1 - eta_minus, sufficiently large T, and the explicit summability condition (4); its proof derives boundedness, shows t_k = T eventually, and proves f(x_k) -> f_min without assuming the conclusion. Condition (4) is an input hypothesis, not an output, and the paper does not claim it holds by construction for every setting. Theorem 2 similarly proves almost-sure convergence from the probabilistic summability condition (13) via Borel-Cantelli. The Gibbs-measure approximation results (Theorems 3-4, Corollary 1) are proved by standard Laplace/Morse asymptotics, with external citations used only for context, not as load-bearing replacements for the proofs. The TT error bounds rest on the external TT-cross theorem [50, Thm. 2.2] and the external Sobolev quadrature bound [31, Thm. 4.5], together with the paper's own quadrature estimates; none of these presuppose the global-convergence conclusion. The self-citations [26,52,71] are used for comparison and historical context and are not load-bearing for Theorem 1. The paper's own limitation statements - that the mesh requirement in (29) is 'unrealistic' for high dimensions and that numerical control parameters are 'chosen heuristically' - show that the implemented TT-IPP and MC-IPP may not be verified to satisfy (4)/(13) in the tested regimes. That is a correctness and verification gap about the bridge from Corollaries 3-4 to the experiments, not a case where a prediction is equivalent to its inputs by construction. No equation is defined in terms of the target conclusion, no fitted parameter is renamed as a prediction, and no uniqueness claim is imported from the authors' prior work to force a choice.

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

The central claim rests on a stated gap assumption (Assumption 3), standard coercivity and regularity assumptions, single-valuedness of the proximal operator, and external results on Laplace asymptotics, TT-cross approximation, and Sobolev quadrature. The free parameters are algorithm hyperparameters and preprocessing shifts, all disclosed in Table 1; they do not fit the conclusions to data but do mean the certified and tested regimes differ. No new physical entities or mediators are introduced.

free parameters (5)
  • Initial temperature delta_0 = 0.1 (0.5 in one TT-IPP test)
    Hand-chosen concentration scale; controls the trade-off between exploration and Gibbs-peak resolution; the theory only requires delta_k -> 0.
  • TT-IPP mesh size h_0 and refinement exponent gamma = h_0 = 0.1, gamma = 1.1 (Table 1)
    The theory requires h = O(delta^((d+2)/(2s)+1)) and gamma > (d+2)/(2s)+1 for convergence certification; the implemented values do not satisfy these for d >= 20 and moderate s, so the certified and tested regimes differ.
  • MC-IPP schedule constants (c, C, alpha_min, alpha_max, p, N_0) = c = 0.9, C = 1.1, alpha_min = 0.2, alpha_max = 0.3, p = 0.8, N_0 = 40d
    Hand-chosen tuning for the adaptive delta, sample size, damping, and rejection; Corollary 4 requires C large enough for (13), which is asserted but not verified for these values.
  • IPP schedule constants (T, tau, t_0, eta_plus, eta_minus, theta_1, theta_2, eps_bar, eta, m) = T = 20, tau = 0.5, t_0 = 1, eta_minus = 0.5/0.9, eta_plus = 2, theta_1 = 0.25, theta_2 = 0.75, eps_bar = 0.2, eta =…
    Implementation constants from Table 1; the theory constrains them only through inequalities (alpha_min > 1 - eta_minus, T sufficiently large), so these values are heuristic.
  • Function shift to set f_min = 0 = per-test shift
    Benchmarks are shifted and scaled so the minimum value is 0 and the minimizer lies in [-1,1]^d; this changes the exp(-f/delta) scale and the sign of phi_min in the error bounds (29)-(31).
assumptions (6)
  • ad hoc to paper Assumption 3: there exists mu > 0 such that 0 in partial-f(x) and f(x) < f_min + mu imply f(x) = f_min (gap condition separating global minima from other stationary points)
    Introduced in Section 2 as the mechanism that upgrades stationarity of the limit point to global minimality in Theorem 1. Without it the subsequential limit is only a stationary point. It is a real restriction on the landscape, not as 'mild' as the text claims.
  • domain assumption Assumptions 1-2: f is continuous with a global minimizer, and f is p-coercive
    Standard regularity used for existence and boundedness of the iterates at (9)-(12); invoked in the proofs of Theorems 1 and 2.
  • domain assumption Single-valuedness of prox_t f (Proposition 1 conditions: prox-regularity, nondegenerate minimizer, or sharpness)
    Needed for Corollaries 3-4 so that the Gibbs/TT/MC estimates have a unique target; the abstract's 'unique global minimizer' phrasing is replaced by these conditions in the theorems.
  • standard math Laplace and Morse asymptotics for the Gibbs expectation (Theorems 3-4)
    The convergence of the Gibbs expectation to the proximal point with rate O(delta) is a standard Laplace expansion, cited to [3], [21], [45], [72]; the paper repackages it with p-coercivity plus local C^2 hypotheses.
  • standard math TT-cross error guarantee (Theorem 5 from Oseledets-Tyrtyshnikov [50]) and Sobolev quadrature bound (from Kazashi-Suzuki-Goda [31])
    External results on which the TT error analysis (Proposition 2, Corollary 2) rests; not re-proved in this paper.
  • standard math Borel-Cantelli lemma and Monte Carlo variance bounds
    Used in the proof of Theorem 2 and in the MC sample-complexity argument of Section 5.2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Inexact Proximal Point Algorithms for Zeroth-Order Global Optimization." pith.science (2026). https://pith.science/paper/Y7NNZJDP

@misc{pith2026241211485,
  author       = {Pith},
  title        = {Pith review of: Inexact Proximal Point Algorithms for Zeroth-Order Global Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Y7NNZJDP}},
  note         = {Machine review of arXiv:2412.11485}
}
abstract

This work concerns the zeroth-order global minimization of continuous nonconvex functions with a unique global minimizer and possibly multiple local minimizers. We formulate a theoretical framework for inexact proximal point (IPP) methods for global optimization, establishing convergence guarantees under mild assumptions when either deterministic or stochastic estimates of proximal operators are used. The quadratic regularization in the proximal operator and the scaling effect of a parameter $\delta>0$ create a concentrated landscape of an associated Gibbs measure that is practically effective for sampling. The convergence of the expectation under the Gibbs measure as $\delta\to 0^+$ is established, and the convergence rate of $\mathcal O(\delta)$ is derived under additional assumptions. These results provide a theoretical foundation for evaluating proximal operators inexactly using sampling-based methods such as Monte Carlo (MC) integration. In addition, we propose a new approach based on tensor train (TT) approximation. This approach employs a randomized TT cross algorithm to efficiently construct a low-rank TT approximation of a discretized function using a small number of function evaluations, and we provide an error analysis for the TT-based estimation. We then propose two practical IPP algorithms, TT-IPP and MC-IPP. The TT-IPP algorithm leverages TT estimates of the proximal operators, while the MC-IPP algorithm employs MC integration to estimate the proximal operators. Both algorithms are designed to adaptively balance efficiency and accuracy in inexact evaluations of proximal operators. The effectiveness of the two algorithms is demonstrated through experiments on diverse benchmark functions and various applications.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Operator Splitting with Hamilton-Jacobi-based Proximals

    math.OC 2026-01 unverdicted novelty 7.0 of 10

    The authors show how to swap exact proximal operators for Hamilton-Jacobi Monte Carlo estimates in five splitting algorithms and claim almost-sure convergence, with proof gaps that need repair.

Reference graph

Works this paper leans on

73 extracted references · 67 canonical work pages · cited by 1 Pith paper

  1. [1]

    Ali, M.M., Khompatraporn, C., Zabinsky, Z.B.: A numerica l evaluation of several stochastic algorithms on selected continuous global optim ization test problems. J. Global Optim. 31, 635–672 (2005)

    work page 2005

  2. [2]

    Asi, H., Duchi, J.C.: Stochastic (approximate) proximal point methods: Convergence, optimality, and adaptivity. SIAM J. Optim. 29(3), 2257–2290 (2019)

    work page 2019

  3. [3]

    Sankhy¯ a Ser

    Athreya, K.B., Hwang, C.R.: Gibbs measures asymptotics. Sankhy¯ a Ser. A.72, 191–207 (2010)

    work page 2010

  4. [4]

    Attouch, H., Bolte, J., Svaiter, B.F.: Convergence of des cent methods for semi-algebraic and tame problems: proximal algorithms, forward–backward splitting, and regularized Gauss–Seidel methods. Math. Program. 137(1), 91–129 (2013)

    work page 2013

  5. [5]

    Advances in Neural Information P rocessing Systems 36 (2024)

    Batsheva, A., Chertkov, A., Ryzhakov, G., Oseledets, I.: Protes: probabilistic optimiza- tion with tensor sampling. Advances in Neural Information P rocessing Systems 36 (2024)

    work page 2024

  6. [6]

    Beck, J., Dia, B.M., Espath, L.F., Long, Q., Tempone, R.: F ast Bayesian experimental design: Laplace-based importance sampling for the expected information gain. Comput. Methods Appl. Mech. Eng. 334, 523–553 (2018)

    work page 2018

  7. [7]

    Bertsekas, D.P.: Incremental proximal methods for large scale convex optimization. Math. Program. 129(2), 163–195 (2011)

    work page 2011

  8. [8]

    Bolte, J., Sabach, S., Teboulle, M.: Proximal alternatin g linearized minimization for nonconvex and nonsmooth problems. Math. Program. 146(1), 459–494 (2014)

    work page 2014

Show all 73 references
  1. [9]

    Bernoulli 28(4), 2431–2458 (2022)

    Bras, P.: Convergence rates of Gibbs measures with degene rate minimum. Bernoulli 28(4), 2431–2458 (2022)

    2022

  2. [10]

    https://www.mathworks.com/matlabcentral/fileexchange/18593-differential-evolution (Accessed: 2024)

    Buehren, M.: Differential evolution. https://www.mathworks.com/matlabcentral/fileexchange/18593-differential-evolution (Accessed: 2024)

    2024

  3. [11]

    IEEE Transactions on Information Theory 61(4), 1985–2007 (2015)

    Candes, E.J., Li, X., Soltanolkotabi, M.: Phase retriev al via wirtinger flow: Theory and algorithms. IEEE Transactions on Information Theory 61(4), 1985–2007 (2015)

    2015

  4. [12]

    Cassioli, A., Di Lorenzo, D., Locatelli, M., Schoen, F., Sciandrone, M.: Machine learning for global optimization. Comput. Optim. Appl. 51, 279–303 (2012)

    2012

  5. [13]

    arXiv preprint arXiv:2209.1480 8 (2022)

    Chertkov, A., Ryzhakov, G., Novikov, G., Oseledets, I.: Optimization of functions given in the tensor train format. arXiv preprint arXiv:2209.1480 8 (2022)

    2022

  6. [14]

    Davis, D., Drusvyatskiy, D.: Proximal methods avoid act ive strict saddles of weakly convex functions. Found. Comput. Math. 22(2), 561–606 (2022)

    2022

  7. [15]

    Davis, D., Drusvyatskiy, D., MacPhee, K.J., Paquette, C .: Subgradient methods for sharp weakly convex functions. J. Optim. Theory Appl. 179, 962–982 (2018)

    2018

  8. [16]

    In: International Conference on Machine Learning, pp

    Dinh, L., Pascanu, R., Bengio, S., Bengio, Y.: Sharp mini ma can generalize for deep nets. In: International Conference on Machine Learning, pp . 1019–1028. PMLR (2017)

    2017

  9. [17]

    Engquist, B., Ren, K., Yang, Y.: Adaptive state-depende nt diffusion for derivative-free optimization. Commun. Appl. Math. Comput. 6(2), 1241–1269 (2024)

    2024

  10. [18]

    Evans, L.C.: Partial differential equations, vol. 19. Am erican Mathematical Society (2022) Inexact Proximal Point Algorithms for Zeroth-Order Global Optimization 33

    2022

  11. [19]

    Fornasier, M., Klock, T., Riedl, K.: Consensus-based op timization methods converge globally. SIAM J. Optim. 34(3), 2973–3004 (2024)

    2024

  12. [20]

    Fukushima, M., Mine, H.: A generalized proximal point algorithm for certain non-convex minimization problems. Int. J. Syst. Sci. 12(8), 989–1000 (1981)

    1981

  13. [21]

    Gamkrelidze, R.V.: Analysis I: Integral Representatio ns and Asymptotic Methods, vol. 13. Springer Science & Business Media (2012)

    2012

  14. [22]

    https://infinity77.net/global_optimization/ (Accessed: 2024)

    Gavana, A.: Global optimization benchmarks and AMPGO. https://infinity77.net/global_optimization/ (Accessed: 2024)

    2024

  15. [23]

    arXiv preprint arXiv: 2308.09050 (2023)

    Gomes, A.A., Gomes, D.A.: Derivative-free global minim ization in one dimension: Re- laxation, Monte Carlo, and sampling. arXiv preprint arXiv: 2308.09050 (2023)

    2023 arXiv

  16. [24]

    : Global optimization methods for computational electromagnetics

    Gottvald, A., Preis, K., Magele, C., Biro, O., Savini, A. : Global optimization methods for computational electromagnetics. IEEE Trans. Magn. 28(2), 1537–1540 (1992)

    1992

  17. [25]

    Griebel, M., Harbrecht, H.: Analysis of tensor approxim ation schemes for continuous functions. Found. Comput. Math. pp. 1–22 (2023)

    2023

  18. [26]

    Heaton, H., Wu Fung, S., Osher, S.: Global solutions to nonconvex problems by evolution of Hamilton-Jacobi PDEs. Commun. Appl. Math. Comput. 6(2), 790–810 (2024)

    2024

  19. [27]

    In: Selected Papers Of Alan J Hoffman: With Commentary, pp

    Hoffman, A.J.: On approximate solutions of systems of lin ear inequalities. In: Selected Papers Of Alan J Hoffman: With Commentary, pp. 174–176. World Scientific (2003)

    2003

  20. [28]

    Holland, J.H.: Genetic algorithms. Sci. Am. 267(1), 66–73 (1992)

    1992

  21. [29]

    Jongeneel, W., Yue, M.C., Kuhn, D.: Small errors in rando m zeroth-order optimization are imaginary. SIAM J. Optim. 34(3), 2638–2670 (2024)

    2024

  22. [30]

    In: Machine Learning and Knowledge Discovery in Databases: European Conference, EC ML PKDD 2016, Riva del Garda, Italy, September 19-23, 2016, Proceedings, Part I 16, pp

    Karimi, H., Nutini, J., Schmidt, M.: Linear convergence of gradient and proximal- gradient methods under the polyak-/suppress lojasiewicz condition. In: Machine Learning and Knowledge Discovery in Databases: European Conference, EC ML PKDD 2016, Riva del Garda, Italy, Septembe...

    2016

  23. [31]

    Kazashi, Y., Suzuki, Y., Goda, T.: Suboptimality of Gaus s–Hermite quadrature and optimality of the trapezoidal rule for functions with finite smoothness. SIAM J. Numer. Anal. 61(3), 1426–1448 (2023)

    2023

  24. [32]

    In: Proceedings of ICNN’95- international conference on neural networks, vol

    Kennedy, J., Eberhart, R.: Particle swarm optimization . In: Proceedings of ICNN’95- international conference on neural networks, vol. 4, pp. 19 42–1948. ieee (1995)

    1995

  25. [33]

    arXiv preprint arXiv:2307.15596 (2023)

    Khanh, P.D., Mordukhovich, B., Phat, V.T., Tran, D.B.: I nexact proximal methods for weakly convex functions. arXiv preprint arXiv:2307.15596 (2023)

    2023 arXiv

  26. [34]

    arXiv preprint arXiv:1412.6980 (2014)

    Kingma, D.P.: Adam: A method for stochastic optimizatio n. arXiv preprint arXiv:1412.6980 (2014)

    2014 arXiv

  27. [35]

    Science 220(4598), 671–680 (1983)

    Kirkpatrick, S., Gelatt Jr, C.D., Vecchi, M.P.: Optimiz ation by simulated annealing. Science 220(4598), 671–680 (1983)

    1983

  28. [36]

    SIAM Journal on Optimization 29(4), 2566–2593 (2019)

    Kong, W., Melo, J.G., Monteiro, R.D.: Complexity of a qua dratic penalty accelerated inexact proximal point method for solving linearly constra ined nonconvex composite programs. SIAM Journal on Optimization 29(4), 2566–2593 (2019)

    2019

  29. [37]

    Acta Numerica 28, 287–404 (2019)

    Larson, J., Menickelly, M., Wild, S.M.: Derivative-fre e optimization methods. Acta Numerica 28, 287–404 (2019)

    2019

  30. [38]

    Nature 521(7553), 436–444 (2015)

    LeCun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521(7553), 436–444 (2015)

    2015

  31. [39]

    arXiv preprint arXiv:2405.04150 (2024)

    Lei, M., Pong, T.K., Sun, S., Yue, M.C.: Subdifferentiall y polynomially bounded functions and Gaussian smoothing-based zeroth-order opti mization. arXiv preprint arXiv:2405.04150 (2024)

    2024 arXiv

  32. [40]

    Lemieux, C.: Monte Carlo and quasi-Monte Carlo sampling , vol. 20. Springer (2009)

    2009

  33. [41]

    Lepage, G.P.: VEGAS - an adaptive multi-dimensional int egration program. Tech. rep., Cornell University (1980)

    1980

  34. [42]

    arXiv preprin t arXiv:1811.11378 (2018)

    Liang, J., Monteiro, R.D.: A doubly accelerated inexact proximal point method for nonconvex composite optimization problems. arXiv preprin t arXiv:1811.11378 (2018)

    2018 arXiv

  35. [43]

    SIAM (2013)

    Locatelli, M., Schoen, F.: Global optimization: theory , algorithms, and applications. SIAM (2013)

    2013

  36. [44]

    Locatelli, M., Schoen, F.: (Global) optimization: hist orical notes and recent develop- ments. EURON J. Comput. Optim. 9, 100012 (2021)

    2021

  37. [45]

    Miller, P.D.: Applied asymptotic analysis, vol. 75. Ame rican Mathematical Soc. (2006)

    2006

  38. [46]

    Moreau, J.J.: Proximit´ e et dualit´ e dans un espace hilbertien. Bull. Soc. Math. Fr. 93, 273–299 (1965) 34 Minxin Zhang et al

    1965

  39. [47]

    Wiley-Interscience (1983)

    Nemirovskij, A.S., Yudin, D.B.: Problem Complexity and Method Efficiency in Opti- mization. Wiley-Interscience (1983)

    1983

  40. [48]

    Nesterov, Y., Spokoiny, V.: Random gradient-free minim ization of convex functions. Found. Comput. Math. 17(2), 527–566 (2017)

    2017

  41. [49]

    https://github.com/oseledets/TT-Toolbox (Accessed: 2024-11-20)

    Oseledets, I.: TT-toolbox. https://github.com/oseledets/TT-Toolbox (Accessed: 2024-11-20)

    2024

  42. [50]

    Linear Algebra Appl

    Oseledets, I., Tyrtyshnikov, E.: TT-cross approximati on for multidimensional arrays. Linear Algebra Appl. 432(1), 70–88 (2010)

    2010

  43. [51]

    Oseledets, I.V.: Tensor-train decomposition. SIAM J. S ci. Comput. 33(5), 2295–2317 (2011)

    2011

  44. [52]

    Osher, S., Heaton, H., Wu Fung, S.: A hamilton–jacobi-ba sed proximal operator. Proc. Natl. Acad. Sci. 120(14), e2220469120 (2023)

    2023

  45. [53]

    Parikh, N., Boyd, S., et al.: Proximal algorithms. Found . Trends Optim. 1(3), 127–239 (2014)

    2014

  46. [54]

    Mathematical Models an d Methods in Applied Sciences 27(01), 183–204 (2017)

    Pinnau, R., Totzeck, C., Tse, O., Martin, S.: A consensus -based model for global op- timization and its mean-field limit. Mathematical Models an d Methods in Applied Sciences 27(01), 183–204 (2017)

    2017

  47. [55]

    Poliquin, R.A., Rockafellar, R.T.: A calculus of prox-r egularity. J. Convex Anal. 17(1), 203–210 (2010)

    2010

  48. [56]

    Advances in Neural Information Processing Systems 35, 14236–14249 (2022)

    Qin, Z., Lidiak, A., Gong, Z., Tang, G., Wakin, M.B., Zhu, Z.: Error analysis of tensor- train cross approximation. Advances in Neural Information Processing Systems 35, 14236–14249 (2022)

    2022

  49. [57]

    Quarteroni, A., Sacco, R., Saleri, F.: Numerical mathem atics, vol. 37. Springer Science & Business Media (2010)

    2010

  50. [58]

    Reali, F., Priami, C., Marchetti, L.: Optimization algo rithms for computational systems biology. Front. Appl. Math. Stat. 3, 6 (2017)

    2017

  51. [59]

    In: International conference on machine learning, pp

    Reddi, S.J., Hefny, A., Sra, S., Poczos, B., Smola, A.: St ochastic variance reduction for nonconvex optimization. In: International conference on machine learning, pp. 314–323. PMLR (2016)

    2016

  52. [60]

    Rockafellar, R.T.: Monotone operators and the proximal point algorithm. SIAM J. Control Optim. 14(5), 877–898 (1976)

    1976

  53. [61]

    Rockafellar, R.T.: Advances in convergence and scope of the proximal point algorithm. J. Nonlinear and Convex Analysis 22(11), 2347–2374 (2021)

    2021

  54. [62]

    CRC Press (2013)

    Roncalli, T.: Introduction to risk parity and budgeting . CRC Press (2013)

    2013

  55. [63]

    Elsevier, New Delhi 16, 32–33 (2009)

    Ross, S.: Probability and statistics for engineers and s cientists. Elsevier, New Delhi 16, 32–33 (2009)

    2009

  56. [64]

    Saab, R., Yılmaz, ¨O.: Sparse recovery by non-convex optimization–instance optimality. Appl. Comput. Harmon. Anal. 29(1), 30–48 (2010)

    2010

  57. [65]

    Linear Algebra Appl

    Savostyanov, D.V.: Quasioptimality of maximum-volume cross interpolation of tensors. Linear Algebra Appl. 458, 217–244 (2014)

    2014

  58. [66]

    Shetty, S., Lembono, T., Loew, T., Calinon, S.: Tensor tr ain for global optimization problems in robotics. Int. J. Rob. Res. 43(6), 811–839 (2024)

    2024

  59. [67]

    Soley, M.B., Bergold, P., Batista, V.S.: Iterative power algorithm for global optimization with quantics tensor trains. J. Chem. Theory Comput. 17(6), 3280–3291 (2021)

    2021

  60. [68]

    Solodov, M.V., Svaiter, B.F.: A unified framework for som e inexact proximal point algorithms. Numer. Funct. Anal. Optim. 22(7-8), 1013–1035 (2001)

    2001

  61. [69]

    Advances in Neural Infor mation Processing Systems 35, 26052–26065 (2022)

    Sozykin, K., Chertkov, A., Schutski, R., Phan, A.H., Cic hocki, A.S., Oseledets, I.: TTOpt: A maximum volume quantized tensor train-based optim ization and its appli- cation to reinforcement learning. Advances in Neural Infor mation Processing Systems 35, 26052–26065 (2022)

    2022

  62. [70]

    Storn, R., Price, K.: Differential evolution–a simple an d efficient heuristic for global optimization over continuous spaces. J. Global Optim. 11, 341–359 (1997)

    1997

  63. [71]

    arXiv preprint arXiv:2406.02003 (2024)

    Tibshirani, R.J., Fung, S.W., Heaton, H., Osher, S.: Lap lace meets Moreau: Smooth approximation to infimal convolutions using Laplace’s meth od. arXiv preprint arXiv:2406.02003 (2024)

    2024 arXiv

  64. [72]

    SIAM (2001) Inexact Proximal Point Algorithms for Zeroth-Order Global Optimization 35 A Additional proofs A.1 Proof of Theorem 2 First, we show that lim k→∞ tk =T almost surely

    Wong, R.: Asymptotic approximations of integrals. SIAM (2001) Inexact Proximal Point Algorithms for Zeroth-Order Global Optimization 35 A Additional proofs A.1 Proof of Theorem 2 First, we show that lim k→∞ tk =T almost surely. By (13) ∑ ∞ k=0 P (‖ ‖yk− ˆxk ‖ ‖ 2 >ǫ k ) <∞. Th...

    2001

  65. [73]

    By Bolzano-Weiserstrass theorem, with probability 1, there exists a subsequence { ˆxkj } that converges to some limit x∞

    Combining with (4), it follows that { xk } k≥0 is uniformly bounded with probability 1 as well. By Bolzano-Weiserstrass theorem, with probability 1, there exists a subsequence { ˆxkj } that converges to some limit x∞. Following the same arguments in the proof of Theorem 1, if ...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.