Pith. sign in

REVIEW 2 major objections 4 minor 33 references

A Cubic Regularization Method for Multiobjective Optimization

T0 review · 2 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper claims a cubic regularization method using approximate gradients and Hessians and a nonmonotone acceptance rule reaches an $\epsilon$-approximate Pareto critical point in at most $O(C\epsilon^{-3/2})$ iterations, with explicit…

desk verdict A solid multiobjective cubic regularization paper with a genuine algorithmic novelty and correct main complexity bound, but the abstract's derivative-free dimension exponents don't match the proofs and need correcting. read the letter →

arxiv 2506.08181 v1 pith:TTVNI27L submitted 2025-06-09 math.OC

Douglas S. Gonçalves , Max L. N. Gonçalves , Jefferson G. Melo This is my paper
classification math.OC MSC 49M1565K0590C2690C29
keywords cubicregularizationmultiobjectiveoptimizationParetocriticalpointiterationcomplexityfinitedifferencesderivative-freenonconvexnonmonotonelinesearch
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 introduces M-CRM, a cubic regularization method for unconstrained multiobjective optimization that builds a model from the maximum of cubic-regularized component models and minimizes it at every iteration. It tries to establish that, when each objective has a globally Lipschitz Hessian and is bounded below, the method produces an $\epsilon$-approximate Pareto critical point in at most $O(C\epsilon^{-3/2})$ iterations, where $C$ collects the Lipschitz constant, initial data, and inexactness parameters. The method deliberately allows gradient and Hessian approximations, and it tunes the regularization parameter and the required derivative accuracy together through a nonmonotone line search, so no user-supplied decrease tolerance is needed. For finite-difference implementations the paper gives iteration and function-evaluation bounds that depend on the dimension $n$, the number of objectives $m$, and a stepsize exponent $\beta$; under local strong convexity it also shows superlinear convergence with exact gradients and quadratic convergence with exact gradients and Hessians. The contribution matters because it transfers the worst-case guarantees of scalar cubic regularization to the multiobjective setting while tolerating inexact derivative information.

What carries the argument

The carrying object is the regularized model $M_{x,\sigma}(y)=\max_{j\in\mathcal{J}} M^j_{x,\sigma}(y)$, with $M^j_{x,\sigma}(y)=\langle \bar\nabla F_j(x),y-x\rangle+\tfrac12\langle \bar\nabla^2 F_j(x)(y-x),y-x\rangle+\tfrac{\sigma}{6}\|y-x\|^3$, together with the approximate KKT conditions (3.2) on subproblem (1.3). The cubic term makes the model coercive, so a solution exists, and it yields the key decrease inequality (2.9), which converts Taylor-model error and finite-difference derivative errors into a step-size-dependent decrease. The nonmonotone acceptance test (3.3) then raises $\sigma$ (and tightens the derivative-error tolerances (3.1)) until that inequality holds, producing the telescoping sum that gives the $\epsilon^{-3/2}$ iteration bound.

What would settle it

On a smooth nonconvex multiobjective instance with known Lipschitz constants, run the method for a sequence of tolerances $\epsilon$ and record the first iteration where $\|\sum_j \lambda_t^j \nabla F_j(x_t)\le \epsilon$; the central claim predicts the count grows at most like a constant times $\epsilon^{-3/2}$, so observing a steeper growth rate, or a failure of the inner solver to meet (3.2), would refute the theorem.

Watch

Extended reading notes

Core claim

The central claim is that the scalar cubic-regularization paradigm survives the passage to several objectives: a point $(x_{t+1},\lambda_{t+1})$ generated from an approximate KKT point of subproblem (1.3) decreases every component of $F$ by an amount controlled by the cubic step, and this decrease is strong enough to make the weighted gradient sums $\sum_j \lambda_t^j \nabla F_j(x_t)$ tend to zero. Consequently all limit points of the sequence are Pareto critical, and an $\epsilon$-approximate Pareto critical point is certified in at most $O(C\epsilon^{-3/2})$ outer iterations under assumptions (A1)-(A2). When derivatives are unavailable, the paper shows that central finite-difference formulas with step proportional to $\|x_t-x_{t-1}\|/n^\beta$ satisfy the required error conditions, leading to $O(n^{1-\beta}\epsilon^{-3/2})$ iterations and $O(mn^{3-\beta}\epsilon^{-3/2})$ function evaluations for the derivative-free variant. Under the local convexity assumption (A3), the same sequence is shown to converge to a local Pareto solution, superlinearly if gradients are exact and quadratically if both gradients and Hessians are exact.

Load-bearing premise

The analysis assumes that every iteration can call a subroutine that returns an approximate solution of the cubic subproblem satisfying the fixed tolerance conditions (3.2), and that all objective Hessians are globally Lipschitz; if either fails, the stated iteration bounds are not guaranteed.

Editorial extensions

If this is right

  • If the central claim is right, smooth nonconvex multiobjective problems can be solved with provable worst-case guarantees using only approximate gradients and Hessians, matching the scalar cubic-regularization rate of $O(\epsilon^{-3/2})$.
  • The derivative-free variant makes the cost of missing analytic derivatives explicit: with $\beta\in[0,1]$ controlling the finite-difference step, the iteration count is $O(n^{1-\beta}\epsilon^{-3/2})$ and the function-evaluation count is $O(mn^{3-\beta}\epsilon^{-3/2})$, allowing a trade-off between larger steps (fewer, cheaper iterations) and smaller steps (more reliable derivative approximations
  • Under local convexity the method does not sacrifice speed: exact gradients give superlinear convergence, and exact gradients plus Hessians give quadratic convergence, so the inexactness controls mainly affect the earlier, global phase of the run.
  • Because the acceptance rule is nonmonotone and parameter-free, the algorithm can be used as a black-box stop when the weighted gradient norm, or its finite-difference proxy, falls below a tolerance, without pre-tuning a sufficient-decrease parameter.

Reading between the lines

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

  • The practical cost of the derivative-free variant is dominated by the shrinking finite-difference step: as the iterates approach a Pareto critical point, $\|x_t-x_{t-1}\|\to 0$, so $h\to 0$ and function evaluations eventually suffer cancellation; choosing $\beta$ close to $1$ keeps $h$ larger but weakens the iteration bound, so the experimental choice $\beta=1/2$ is a plausible middle point.
  • Subproblem (1.3) has a strictly feasible point and satisfies a constraint qualification, so any generic nonlinear programming solver that returns a point meeting (3.2) can be substituted for the augmented-Lagrangian solver used in the experiments; the method's guarantees are solver-agnostic.
  • The error conditions (3.1) only require that gradient and Hessian errors decay with the previous step, so subsampled or randomized derivative estimators could plausibly be plugged in place of finite differences, at the price of probabilistic versions of Lemmas 3.9-3.14.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. The paper introduces a multiobjective cubic regularization method (M-CRM) for nonconvex unconstrained multiobjective optimization. At each outer iteration, the algorithm constructs cubic-regularized models from approximate gradients and Hessians of the objective components, solves the resulting subproblem approximately in the KKT sense, and adjusts both the regularization parameter and the derivative-approximation accuracy through a nonmonotone line search. The main theoretical results are an O(C ε^{-3/2}) iteration bound for reaching an ε-approximate Pareto critical point under globally Lipschitz Hessians and bounded-below objectives, finite-difference variants of the algorithm, global convergence to weak Pareto critical points, and local superlinear or quadratic convergence under strong convexity of the objectives on a level set. Numerical experiments compare exact and inexact versions of the method against a safeguarded multiobjective Newton method, and a derivative-free variant is applied to a regularized logistic regression problem.

Significance. If the stated complexity claims are corrected, this is a valuable contribution. The exact-derivative version delivers the expected O(ε^{-3/2}) complexity for multiobjective cubic regularization, the joint adaptive control of the regularization parameter and derivative accuracy is a useful algorithmic idea, and the numerical study is reasonably extensive for a methods paper. The paper also contains careful proofs of global convergence and local rates, and the derivative-free implementation is accompanied by detailed finite-difference error bounds. The main weakness is that the abstract and introduction advertise finite-difference complexity exponents that are stronger than what the proofs establish; the correct exponents already appear in Remarks 3.13 and 3.15, so the discrepancy is internally visible and, in principle, fixable by rewriting the advertised claims.

major comments (2)
  1. [Abstract, Section 1, and Remark 3.15] The derivative-free complexity claims in the Abstract and Section 1 are not supported by the proof. Lemma 3.9 gives κ̄_G = L n^{(1-2β)/2} and Lemma 3.14 gives κ̄_H = L n^{1-β} for the finite-difference gradient and Hessian. Substituting these into Δ_0 in Eq. (3.6), the dominant term for β<1 is α(2κ̄_H)/2 ~ L n^{1-β}, so Δ_0 ~ n^{3(1-β)/2}. Theorem 3.3 therefore yields O(n^{3(1-β)/2} ε^{-3/2}) iterations, and multiplying by the O(m n^2) function evaluations per inner iteration (Lemma 3.2) gives O(m n^{(7-3β)/2} ε^{-3/2}) function evaluations. These are exactly the bounds stated in Remark 3.15, but they contradict the Abstract's O(n^{1-β} ε^{-3/2}) iterations and O(m n^{3-β} ε^{-3/2}) function evaluations. The advertised exponents are off by a factor of (3/2) in the iteration exponent and by n^{1/2} in the evaluation exponent; the Abstract and Section 1 should be corrected to match Remark 3.15, unless a different finite-difference step choice can justify the stronger bound.
  2. [Section 1 and Remark 3.13] The same dimension-exponent error appears in the claim for exact gradients with Hessians computed by finite differences based on gradient values. Section 1 states an iteration bound of O(n^{(1-2β)/2} ε^{-3/2}) and O(m n^{(3-2β)/2}) function and gradient evaluations for β∈[0,1/2]. However, Lemma 3.12 gives κ̄_H = L n^{(1-2β)/2}; with κ̄_G = 0, Eq. (3.6) gives Δ_0 ~ n^{3(1-2β)/4}, so Theorem 3.3 yields O(n^{3(1-2β)/4} ε^{-3/2}) iterations and, with O(m n) evaluations per inner iteration, O(m n^{(7-6β)/4} ε^{-3/2}) total evaluations. These are precisely the bounds stated in Remark 3.13. The Section 1 claims again omit the 3/2 power on the dimension-dependent factor and should be corrected.
minor comments (4)
  1. [Lemma 3.12, Eq. (3.38)] The displayed formula for the central-difference Hessian appears to have a typo in the last column: it reads ∇F_j(x+he_n) - ∇F_j(x-he_1), which should presumably be ∇F_j(x+he_n) - ∇F_j(x-he_n).
  2. [Lemma 3.9 and Algorithm 2] The expression for the step size h in Lemma 3.9 is typeset ambiguously: h := √6‖x_t-x_{t-1}‖/(n^β α^{(i-1)})^{1/2} should be written as h = sqrt(6) ‖x_t-x_{t-1}‖ / (n^β α^{i-1})^{1/2}, and the same ambiguity affects Algorithm 2's definition of h. The derivation in Appendix A confirms the intended exponent, but the notation should be clarified.
  3. [Lemma 3.2, Eq. (3.5)] The formula for δ_T has an unbalanced parenthesis: it should read δ_T ≤ δ[2T + log_α(2(L + 3α(2κ̄_G + κ̄_H)) + σ_1) - log_α(σ_1)].
  4. [Figure 2 and Section 4.2] The caption of Figure 2 refers to 'M-CRM-I-DF2' while the text consistently calls the method 'M-CRM-I-DF'; the caption should be made consistent.

Circularity Check

0 steps flagged · score 1.0 of 10

No load-bearing circularity; the finite-difference exponent mismatch in the abstract is an internal arithmetic error, not a circular derivation.

full rationale

Walking the derivation chain, the central complexity bound is not an input in disguise. Theorem 3.3 obtains the sum-of-cubes bound (3.7) by telescoping the nonmonotone sufficient-decrease condition (3.3); that condition is proved in Corollary 2.3 solely from assumption (A1), the error tolerances (3.1), and the cubic subproblem conditions (3.2). The constants Delta_0 and Delta_1 in (3.6) are explicit combinations of L, theta, kappa_G, kappa_H, sigma_1, alpha, and the initial gap F_j(x_1) - F_j^*, with no parameter fitted to the target epsilon-approximate Pareto-critical output. The derivative-free variant rests on Lemmas 3.9 and 3.14, which give kappa_G = L n^{(1-2beta)/2} and kappa_H = L n^{1-beta} from Taylor bounds; substituting into Delta_0 yields the O(n^{3(1-beta)/2} epsilon^{-3/2}) iteration bound and O(m n^{(7-3beta)/2} epsilon^{-3/2}) evaluation bound stated in Remark 3.15. The self-citations ([20], [21], [22], [24]) are related-work pointers, an experimental baseline, and a consistency comparison; the statement in Remark 3.13 that the scalar case recovers the complexity results of the CRM from [24] is a check against an independent derivation, not a load-bearing premise. No uniqueness theorem or ansatz is imported from prior work by the same authors. I therefore find no circular reduction. Separately, the abstract and Section 1 advertise O(n^{1-beta} epsilon^{-3/2}) iterations and O(m n^{3-beta} epsilon^{-3/2}) evaluations, which are not supported by Remark 3.15's proof; this is an arithmetic/expository inconsistency in the advertised exponents, not circularity.

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

The only freely chosen quantities are algorithmic parameters such as alpha, theta, sigma_1, beta, and the error thresholds. None are fitted to data to force a desired conclusion. The essential assumptions are the Lipschitz Hessian condition, boundedness below, and the availability of an approximate subproblem solver. No new physical entities are introduced.

free parameters (5)
  • beta = 0.5 in Algorithm 2 experiments
    User-specified exponent in the finite-difference step size h. It affects the dimension dependence of the iteration complexity. The paper's abstract and Remark 3.15 give conflicting complexity exponents as a function of beta.
  • alpha = 2 in experiments
    Line-search expansion factor in the inner loop of Algorithm 1. It appears in the complexity constants but does not change the order in epsilon.
  • theta = 0.9 for M-CRM-I, 10^-8/||...||^2 for M-CRM-E
    Inexactness tolerance in the approximate KKT condition (3.2). It enters the complexity constant Delta_0.
  • sigma_1 = 2e-2 in experiments
    Initial regularization parameter. It appears in the complexity constants and the line search acceptance condition.
  • kappaG_bar and kappaH_bar = derived from L and n for finite differences; user inputs in general
    Thresholds in the derivative error condition (3.1). They are chosen by hand in the abstract algorithm and expressed via L and n in the finite-difference implementations.
assumptions (6)
  • domain assumption A1: Hessians of all objective components are globally L_j-Lipschitz continuous
    Used throughout the paper to derive the cubic upper bound (2.3) and the gradient difference bound (2.4), which are essential for the line search termination and complexity analysis.
  • domain assumption A2: Each objective F_j is bounded below
    Needed for the telescoping sum in Theorem 3.3 to yield a finite complexity bound. Without boundedness, the function value could go to -infinity.
  • domain assumption A3: Each F_j is strongly convex on a relevant level set (for local rates)
    Used only in Section 3.2 to establish superlinear and quadratic local convergence rates, not for the global complexity result.
  • domain assumption A4: Exact gradients and Hessian error bounded by a multiple of min(||x_t - x_{t-1}||, zeta||g_t||) (for local rates)
    Imposed in the local convergence analysis to control the effect of Hessian approximation and get the quadratic contraction.
  • domain assumption At every iteration, an approximate KKT point of the subproblem can be computed satisfying (3.2)
    The algorithm and its complexity analysis assume this oracle. For the nonconvex subproblem (1.3), no practical solver is guaranteed to produce such a point, which is the most fragile premise.
  • standard math Standard Taylor expansion bounds for functions with Lipschitz Hessians
    Equations (2.3) and (2.4) are classical results used to bound function decrease and gradient residual; they are invoked without proof beyond a brief citation to the Lipschitz property.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Cubic Regularization Method for Multiobjective Optimization." pith.science (2026). https://pith.science/paper/TTVNI27L

@misc{pith2026250608181,
  author       = {Pith},
  title        = {Pith review of: A Cubic Regularization Method for Multiobjective Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TTVNI27L}},
  note         = {Machine review of arXiv:2506.08181}
}
abstract

This work introduces a new cubic regularization method for nonconvex unconstrained multiobjective optimization problems. At each iteration of the method, a model associated with the cubic regularization of each component of the objective function is minimized. This model allows approximations for the first- and second-order derivatives, which must satisfy suitable error conditions. One interesting feature of the proposed algorithm is that the regularization parameter of the model and the accuracy of the derivative approximations are jointly adjusted using a nonmonotone line search criterion. Implementations of the method, where derivative information is computed using finite difference strategies, are discussed. It is shown that, under the assumption that the Hessians of the objectives are globally Lipschitz continuous, the method requires at most $\mathcal{O}(C\epsilon^{-3/2})$ iterations to generate an $\epsilon$-approximate Pareto critical. In particular, if the first- and second-order derivative information is computed using finite differences based solely on function values, the method requires at most $\mathcal{O}(n^{1-\beta}\epsilon^{-3/2})$ iterations, corresponding to $\mathcal{O}\left(mn^{3-\beta}\varepsilon^{-\frac{3}{2}}\right)$ function evaluations, where \(n\) is the dimension of the domain of the objective function, $m$ is the number of objectives, and $\beta \in [0,1]$ is a constant associated with the stepsize used in the finite-difference approximation. We further discuss the global convergence and local convergence rate of the method. Specifically, under the local convexity assumption, we show that the method achieves superlinear convergence when the first derivative is computed exactly, and quadratic convergence when both first- and second-order derivatives are exact.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

33 extracted references · 31 canonical work pages

  1. [1]

    Andreani, E

    R. Andreani, E. G. Birgin, J. M. Mart\' i nez, and M. L. Schuverdt , On A ugmented L agrangian methods with general lower-level constraints , SIAM J. Optim., 18 (2008), pp. 1286--1309

    work page 2008

  2. [2]

    P. B. Assun c \ a o, O. P. Ferreira, and L. F. Prudente , Conditional gradient method for multiobjective optimization , Comput. Optim. Appl., 78 (2021), pp. 741--768, https://doi.org/10.1007/s10589-020-00260-5

    work page doi:10.1007/s10589-020-00260-5 2021

  3. [3]

    Bellavia, G

    S. Bellavia, G. Gurioli, and B. Morini , Adaptive cubic regularization methods with dynamic inexact Hessian information and applications to finite-sum minimization , IMA J. Numer. Anal., 41 (2020), pp. 764--799

    work page 2020

  4. [4]

    Bellavia, G

    S. Bellavia, G. Gurioli, B. Morini, and P. L. Toint , Quadratic and cubic regularisation methods with inexact function and random derivatives for finite-sum minimisation , in In: Proceedings of the 21st International Conference on Computational Science and Its Applications, 2021, pp. 258--267

    work page 2021

  5. [5]

    Bello-Cruz, L

    Y. Bello-Cruz, L. R. Lucambio P\'erez, and J. G. Melo , Convergence of the projected gradient method for quasiconvex multiobjective optimization , Nonlinear Anal. Theory Methods Appl., 74 (2011), pp. 5268 -- 5273

    work page 2011

  6. [6]

    Bello-Cruz, J

    Y. Bello-Cruz, J. G. Melo, and R. V. G. Serra , A proximal gradient splitting method for solving convex vector optimization problems , Optimization, 71 (2022), pp. 33--53

    work page 2022

  7. [7]

    G. C. Bento, J. X. Cruz-Neto, and A. Soubeyran , A proximal point-type method for multicriteria optimization , Set-Valued Var Anal., 22 (2014), pp. 557--573

    work page 2014

  8. [8]

    Birgin and J

    E. Birgin and J. Martinez , Practical Augmented Lagrangian Methods for Constrained Optimization , SIAM, Philadelphia, 2014

    work page 2014

Show all 33 references
  1. [9]

    E. G. Birgin, J. L. Gardenghi, J. M. Mart \' nez, S. A. Santos, and P. L. Toint , Worst-case evaluation complexity for unconstrained nonlinear optimization using high-order regularized models , Math. Program., 163 (2017), pp. 359--368

    2017

  2. [10]

    Bonnel, A

    H. Bonnel, A. N. Iusem, and B. F. Svaiter , Proximal methods in vector optimization , SIAM J. Optim., 15 (2005), pp. 953--970

    2005

  3. [11]

    Calder \'o n, M

    L. Calder \'o n, M. A. Diniz-Ehrhardt, and J. M. Mart \' nez , On high-order model regularization for multiobjective optimization , Optim. Methods Softw., 37 (2022), pp. 175--191

    2022

  4. [12]

    Cartis, N

    C. Cartis, N. I. M. Gould, and P. L. Toint , On the complexity of steepest descent, N ewton's and regularized newton's methods for nonconvex unconstrained optimization problems , SIAM J. Optim., 20 (2010), pp. 2833--2852

    2010

  5. [13]

    Doikov and G

    N. Doikov and G. N. Grapiglia , First and zeroth-order implementations of the regularized N ewton method with lazy approximated hessians , 2023, https://arxiv.org/abs/2309.02412

    2023 arXiv

  6. [14]

    Fliege, L

    J. Fliege, L. M. Gra \ n a Drummond, and B. F. Svaiter , N ewton's method for multiobjective optimization , SIAM J. Optim., 20 (2009), pp. 602--626

    2009

  7. [15]

    Fliege and B

    J. Fliege and B. F. Svaiter , Steepest descent methods for multicriteria optimization , Math. Method. Oper. Res., 51 (2000), pp. 479--494

    2000

  8. [16]

    E. H. Fukuda and L. M. G. Drummond , Inexact projected gradient method for vector optimization , Comput. Optim. Appl., 54 (2013), pp. 473--493

    2013

  9. [17]

    E. H. Fukuda and L. M. Gra \ n a Drummond , On the convergence of the projected gradient method for vector optimization , Optimization, 60 (2011), pp. 1009--1021

    2011

  10. [18]

    E. H. Fukuda, H. Tanabe, and N. Yamashita , Proximal gradient methods for multiobjective optimization and their applications , Comput. Optim. Appl., 72 (2019), pp. 339--361

    2019

  11. [19]

    E. H. Fukuda, H. Tanabe, and N. Yamashita , An accelerated proximal gradient method for multiobjective optimization , Comput. Optim. Appl., 86 (2023), pp. 421--455

    2023

  12. [20]

    D. S. Gon c alves, M. L. N. Gon c alves, and J. G. Melo , An away-step frank-wolfe algorithm for constrained multiobjective optimization , Comput. Optim. Appl., 88 (2024), pp. 759--781

    2024

  13. [21]

    M. L. N. Gon c alves , Subsampled cubic regularization method for finite-sum minimization , Optimization, 0 (2024), pp. 1--24

    2024

  14. [22]

    M. L. N. Gon c alves, F. S. Lima, and L. F. Prudente , Globally convergent N ewton-type methods for multiobjective optimization , Comput. Optim. Appl., 83 (2022), pp. 403--434

    2022

  15. [23]

    L. M. Gra \ n a Drummond and B. F. Svaiter , A steepest descent method for vector optimization , J. Comput. Appl. Math., 175 (2005), pp. 395 -- 414

    2005

  16. [24]

    G. N. Grapiglia, M. L. N. Gon c alves, and G. N. Silva , A cubic regularization of N ewton's method with finite difference hessian approximations , Numer. Algorithms, 90 (2022), pp. 607--630

    2022

  17. [25]

    Griewank , The modification of N ewton method for unconstrained optimization by bounding cubic terms

    A. Griewank , The modification of N ewton method for unconstrained optimization by bounding cubic terms. , Technical Report NA/12, Department of Applied Mathematics and Theoretical Physics. University of Cambridge, Cambridge (1981)

    1981

  18. [26]

    Kohler and A

    J. Kohler and A. Lucchi , Sub-sampled cubic regularization for non-convex optimization , in In Proceedings of the 34th International Conference on Machine Learning (ICML 2017), vol. 70, 2017, pp. 1895 -- 1904

    2017

  19. [27]

    M. S. Louzeiro, G. N. Silva, J. Yuan, and D. Zhang , An adaptive cubic regularization quasi- N ewton method on R iemannian manifolds , 2024, https://arxiv.org/abs/2402.12464

    2024 arXiv

  20. [28]

    L. R. Lucambio P\'erez and L. F. Prudente , Nonlinear conjugate gradient methods for vector optimization , SIAM J. Optim., 28 (2018), pp. 2690--2720

    2018

  21. [29]

    Nesterov and B

    Y. Nesterov and B. T. Polyak , Cubic regularization of N ewton method and its global performance , Math. Program., 184 (2006), pp. 177--205

    2006

  22. [30]

    Nocedal and S

    J. Nocedal and S. J. Wright , Numerical Optimization , Springer-Verlag, 2 ed., 2006

    2006

  23. [31]

    Street, W

    W. Street, W. Wolberg, and O. Mangasarian , Nuclear feature extraction for breast tumor diagnosis , in IS&T/SPIE 1993 International Symposium on Electronic Imaging: Science and Technology, vol. 1905, San Jose, CA, 1993

    1993

  24. [32]

    Z. Wang, Y. Zhou, Y. Liang, and G. Lan , A note on inexact gradient and hessian conditions for cubic regularized N ewton's method , Oper. Res. Lett., 47 (2019), pp. 146--149

    2019

  25. [33]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION not #0 #1 if FUNCTION and 'skip pop #0 if FUNCTION or pop #1 'skip if FUNCTION new.block.check...

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.