Pith. sign in

REVIEW 3 major objections 4 minor 2 cited by

Convergence analysis of wide shallow neural operators within the framework of Neural Tangent Kernel

T0 review · 3 major / 4 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 Wide shallow neural operators are provably trainable to zero error under gradient descent.

desk verdict NTK convergence for shallow Chen-Chen operators fills a small gap, but a false positive-definiteness lemma and a scrambled appendix need a fix before this is citable. read the letter →

arxiv 2412.05545 v3 pith:47H4MLMS submitted 2024-12-07 cs.LG cs.DSmath.OC

Xianliang Xu , Ye Li , Zhongyi Huang This is my paper · ORCID
classification cs.LGcs.DSmath.OC MSC 68T07
keywords neuraltangentkerneloperatorgradientdescentconvergenceanalysisover-parameterizationshallownetworkphysics-informedlinear
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 a missing piece in the theory of operator learning: a training-error guarantee for a wide shallow neural operator trained by gradient descent. Working in the neural tangent kernel regime, where random initialization and over-parameterization keep every weight near its starting point, the authors prove that continuous-time gradient flow drives the squared training error to zero exponentially fast, and that discrete-time gradient descent converges at a linear rate to a global minimum with high probability. The same conclusion is extended to physics-informed shallow neural operators, where the loss includes PDE residuals at interior points and boundary conditions. If correct, these results say that interpolating the training data is not an obstacle for shallow operator networks once the width is large enough, and that the speed of training is controlled by the smallest eigenvalues of the limiting neural tangent kernel matrices.

What carries the argument

The load-bearing object is the neural tangent kernel Gram matrix of the model, split into two parts: $H^\infty=H^\infty_1\otimes H^\infty_2$ coming from the trunk weights and $\tilde H^\infty=\tilde H^\infty_1\otimes\tilde H^\infty_2$ from the branch weights. The argument needs the smallest eigenvalues $\lambda_0,\tilde\lambda_0$ to be strictly positive, which Lemma 1 obtains from the non-parallelism of the sampled input functions and output points. Around those fixed matrices, the proof controls two perturbations: the Gram matrices at initialization are close to their infinite-width limits, and they stay close to their initial values through training because the weights move at most $O(1/\sqrt m)$; a Taylor-residual recursion captures the discrete-time update. Together these facts keep the smallest eigenvalues bounded away from zero and turn the loss dynamics into a contraction.

What would settle it

Take a dataset with two identical input functions $u_i=u_j$ but different output labels, train the same shallow operator with a small step size, and track $\|z-G_t(u)\|_2^2$; Lemma 1 then fails because $H^\infty_1$ is singular, and the predicted uniform linear contraction should break down, or the width required to keep it should blow up as the eigenvalue gap goes to zero.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the neural tangent kernel analysis of finite-dimensional ReLU networks transfers to shallow operator networks. For the branch–trunk model $G(u)(y)=\frac{1}{\sqrt m}\sum_{r=1}^{m}[\frac{1}{\sqrt p}\sum_{k=1}^{p}\tilde a_{rk}\sigma(\tilde w_{rk}^{\top}u)]\sigma(w_r^{\top}y)$, the infinite-width Gram matrices $H^\infty$ and $\tilde H^\infty$ factor as Kronecker products, and under the condition that no two sampled input functions $u_i$ are parallel and no two sampled output points $y_j$ are parallel, both are strictly positive definite with smallest eigenvalues $\lambda_0$ and $\tilde\lambda_0$. Theorem 2 states that, with high probability, discrete-time gradient descent satisfies $\|z-G_t(u)\|_2^2 \le \bigl(1-\eta(\lambda_0+\tilde\lambda_0)/2\bigr)^{t}\|z-G_0(u)\|_2^2$ provided the width $m$ is of order $\Omega\bigl(n_1^4 n_2^4 \log(n_1 n_2/\delta)\log^3(m/\delta)/(\min(\lambda_0,\tilde\lambda_0)^2(\lambda_0+\tilde\lambda_0)^2)\bigr)$; Theorem 1 gives the continuous-time bound $\exp(-(\lambda_0+\tilde\lambda_0)t)$. The same mechanism proves the analogous linear-rate statement for physics-informed training, with width scaling polynomial in the sample count and PDE dimension $d$.

Load-bearing premise

The proof hinges on the two infinite-width kernel matrices having strictly positive smallest eigenvalues, which is guaranteed only when no two sampled input functions are parallel and no two sampled output points are parallel; if the data contain parallel or identical samples, the linear-rate conclusion collapses.

Editorial extensions

If this is right

  • With enough width, gradient descent finds an interpolating model, so the training loss can be driven to zero.
  • The convergence rate is linear, with contraction factor $1-\eta(\lambda_0+\tilde\lambda_0)/2$, so the speed is controlled by the smallest NTK eigenvalues, not by the data dimension.
  • The same guarantee holds in continuous time, with an exponential rate $\exp(-(\lambda_0+\tilde\lambda_0)t)$, and for physics-informed training of the same architecture on PDE residuals plus boundary data.
  • The strict positive-definiteness condition quantifies a data-dependent requirement: no two sampled input functions, and no two sampled output locations, may be parallel.

Reading between the lines

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

  • The paper leaves open the practical size of the width requirement; because it involves fourth powers of both sample counts and inverse eigenvalue gaps, the implied $m$ may be very large, so whether the NTK regime is reachable for realistic operator-learning datasets is not settled.
  • The same Kronecker-structure argument suggests a route for other operator architectures, such as encoders that randomly project input functions, as long as their limiting Gram matrices can be shown strictly positive definite.
  • A natural testable extension is to monitor $\|H(t)-H(0)\|_F$ during training; the proof predicts it stays $O(1/\sqrt m)$, so a numerical check would immediately confirm or contradict the NTK regime.
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

3 major / 4 minor

Summary. The paper studies the training dynamics of wide shallow neural operators of the form G(u)(y) = m^{-1/2} Σ_r [p^{-1/2} Σ_k a_{rk} σ(\tilde{w}_{rk}^T u)] σ(w_r^T y), where u is identified with its evaluations at fixed sensors. Under random Gaussian initialization and over-parameterization, the authors claim that gradient flow (Theorem 1) and gradient descent (Theorem 2) converge to a globally optimal training loss at a linear rate, with rates governed by the smallest eigenvalues λ0 and \tilde{λ}_0 of limiting NTK Gram matrices H^∞ and \tilde{H}^∞. A third theorem extends the analysis to physics-informed neural operators for a parabolic PDE with a ReLU^3 trunk activation, asserting linear convergence of the residual. The proofs follow the standard NTK template: show the random Gram matrices concentrate around their kernel limits, prove stability of the Gram matrices under small weight movements, and then close an induction on weight deviations and residual norms.

Significance. If the results are correct, this is a useful first training-error guarantee for shallow neural operators in the NTK regime, complementing existing approximation and generalization analyses. The paper's explicit treatment of the Kronecker structure of the limiting Gram matrices, its discrete-time residual recursion (Lemma 4), and its extension to a physics-informed setting are concrete strengths. The claimed linear rates are falsifiable and the width scalings are explicit, which is valuable for subsequent work. However, several load-bearing statements need repair: the positive-definiteness lemma is incomplete as stated, the PINN dynamics equation is internally inconsistent, and the appendix's duplicate lemma numbering makes the proof of Theorem 1 difficult to verify as submitted.

major comments (3)
  1. [Section 3, Lemma 1 (and Section 5, Lemma 6)] The stated condition for positive definiteness is insufficient because the zero vector satisfies 'no two samples are parallel' vacuously. For example, with n1 = 1 and u1 = 0, the matrix H1^∞ has entry E[σ(0)^2] = 0, so H^∞ = H1^∞ ⊗ H2^∞ = 0 and λ0 = 0; the rates and width bounds in Theorems 1–3 are then undefined. The same failure occurs for y1 = 0. The paper should explicitly assume ui ≠ 0 and yj ≠ 0 (or, as Remark 1 suggests for a related issue, augment both nets with bias terms so that zero inputs become nonzero). This is a load-bearing correction, not a cosmetic one, since all subsequent theorems depend on λ0, \tilde{λ}_0 > 0.
  2. [Section 5, paragraph after the definitions of s(u_i) and h(u_i)] The dynamics equation dGt(u)/dt = (H(t) + \tilde{H}(t))(z − Gt(u)) is inconsistent with the definition of Gt(u) as the residual vector ((s(u1), h(u1)), …, (s(un1), h(un1))). Since s and h are defined as scaled residuals (LG − f) and (G − g), the correct continuous-time residual dynamics should be dGt/dt = −(H(t) + \tilde{H}(t))Gt(u), with the target being zero, not z = ((f,g),…). As written, the equation has fixed point G∞ = z = (f,g), which is not the intended zero-residual solution. Theorem 3's claim of convergence to zero therefore does not follow from the stated dynamics; the notation must be fixed (either define Gt as the prediction and state the dynamics for the residual, or set the target to zero).
  3. [Appendix, Section 7.4, Proof of Theorem 1] The proof of Theorem 1 is not verifiable in its current numbering. The proof says 'Theorem 1 is a direct corollary of Lemma 6 and Lemma 9', but Lemma 6 already denotes the PINN positive-definiteness lemma in Section 5, and Lemma 9 denotes the PINN recursion in Section 5, neither of which is the stated result used here. In addition, the same number 'Lemma 8' appears twice (Sections 8.7 and 8.8), and the auxiliary references in the appendix (e.g., 'Theorem 3.1 in [17]', 'Theorem 2.1 in [19]') do not align with the main reference list entries. The mathematical arguments may be repairable, but the appendix needs a complete renumbering and cross-reference cleanup before the claims can be checked.
minor comments (4)
  1. [Throughout] There are numerous typos and dangling phrases, including 'shallw' (Section 2), 'demenstra ting' (Conclusion), 'Condtion' (Condition 2 in Section 5), and '∈ [n1]' used where n2 or n3 is meant in the PINN Gram-matrix definitions.
  2. [Section 3, Lemma 2 vs Theorem 1] Lemma 2 states a width requirement without the log^3(m/δ) factor that appears in Theorem 1; the proof of Theorem 1 later combines additional constraints, but the lemma as stated is weaker than the theorem's condition and should be aligned.
  3. [Section 3, Remark 2] The remark refers to 'from (11)' when the relevant inequality is equation (9) in the proof sketch of Theorem 1; equation (11) is a later discrete-time definition, so the cross-reference is wrong.
  4. [Section 4, Corollary 1 proof] In the displayed requirement for m, the argument of one logarithm is written as log(n δ) without the denominator, and the line 'where (65), (59) and the condition in Lemma 2' uses numbers that do not correspond to the sequentially numbered lemmas of the paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the convergence rates follow from external NTK kernel-positivity and concentration results, not from the conclusion being assumed or from fitted/renamed inputs.

full rationale

The paper's derivation is a standard NTK analysis applied to a two-layer (branch/trunk) parameterization. The central constants lambda_0 and lambda_tilde_0 are defined as the least eigenvalues of the limiting Gram matrices H_infinity and H_tilde_infinity (Lemma 1), whose strict positive definiteness is justified by external theorems (Theorem 3.1 in [17], Theorem 2.1 in [19], and Lemma 3.2 in [20]); these are independent results and are not assumed to equal anything about the training loss. The training-error contraction in Eq. (9) and Theorem 1 is obtained from lambda_min(H(t)) >= lambda_0/2 and lambda_min(H_tilde(t)) >= lambda_tilde_0/2, and these spectral lower bounds are themselves consequences of the perturbation lemmas (Lemmas 2 and 3), not of the target bound. The discrete-time recursion in Lemma 4 is an exact identity for gradient descent, and Lemma 5 bounds the second-order residual I(t); neither the residual bound nor the induction in Condition 1 presumes the conclusion. No parameter is fitted to a subset of the training data and then reported as a prediction; the only data-dependent quantities are the spectral gaps lambda_0 and lambda_tilde_0, which are inputs to the rate, not objects being predicted. There are no self-citations: the cited works [17], [19], and [20] are by different author groups and are external. The reviewer-identified gap concerning zero samples satisfying the non-parallelism condition vacuously (so that lambda_0 = 0) is a correctness/assumption issue in Lemma 1, not a circularity: even if Lemma 1's statement is incomplete, the theorems do not reduce to their own inputs by construction. Therefore the appropriate circularity score is 0.

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

The central claim rests on standard NTK machinery: bounded data, non-parallel sampling, standard Gaussian/Rademacher initialization, and external kernel-positivity results. No new physical entities or ad hoc fitted constants are introduced; the eigenvalue parameters λ0 and λ̃0 are data-dependent but not fitted.

assumptions (5)
  • domain assumption No two samples in {u_i} are parallel and no two samples in {y_j} are parallel (Lemma 1).
    Used to prove strict positive definiteness of H∞ and H̃∞, which is the basis for the linear convergence rate.
  • domain assumption Input/output sample norms are bounded: ‖u_i‖_2 = O(1), ‖y_j‖_2 = O(1).
    Used throughout the concentration and stability estimates, e.g., in deriving ψ2 norms of ReLU features.
  • domain assumption Initialization: w_r(0) ~ N(0,I), w̃_rk(0) ~ N(0,I), a_rk(0) ~ Unif{-1,1}, with a_rk fixed during training.
    This specific random initialization is assumed in all theorems and is needed for the NTK approximation results.
  • standard math Known positive definiteness of ReLU feature kernels (Theorem 3.1 in [17], Theorem 2.1 in [19], Lemma 3.2 in [20]).
    The proof of Lemma 1 invokes these external results to establish linear independence of features; the paper does not reprove them fully.
  • standard math Sub-Weibull concentration inequalities for sums of random variables (Theorem 3.1 in [22], Bernstein inequality).
    Used in Lemmas 2, 3, 7, 8 to control deviations of empirical Gram matrices from their expectations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Convergence analysis of wide shallow neural operators within the framework of Neural Tangent Kernel." pith.science (2026). https://pith.science/paper/47H4MLMS

@misc{pith2026241205545,
  author       = {Pith},
  title        = {Pith review of: Convergence analysis of wide shallow neural operators within the framework of Neural Tangent Kernel},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/47H4MLMS}},
  note         = {Machine review of arXiv:2412.05545}
}
read the original abstract

Neural operators are aiming at approximating operators mapping between Banach spaces of functions, achieving much success in the field of scientific computing. Compared to certain deep learning-based solvers, such as Physics-Informed Neural Networks (PINNs), Deep Ritz Method (DRM), neural operators can solve a class of Partial Differential Equations (PDEs). Although much work has been done to analyze the approximation and generalization error of neural operators, there is still a lack of analysis on their training error. In this work, we conduct the convergence analysis of gradient descent for the wide shallow neural operators and physics-informed shallow neural operators within the framework of Neural Tangent Kernel (NTK). The core idea lies on the fact that over-parameterization and random initialization together ensure that each weight vector remains near its initialization throughout all iterations, yielding the linear convergence of gradient descent. In this work, we demonstrate that under the setting of over-parametrization, gradient descent can find the global minimum regardless of whether it is in continuous time or discrete time.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Optimization and generalization analysis for two-layer physics-informed neural networks without over-parametrization

    cs.LG 2025-07 reject novelty 5.0 of 10

    A two-layer PINN can be trained by SGD to O(epsilon) loss with width independent of the number of samples, provided the target lies in a custom function class and the SGD trajectory does not explode.

  2. Optimal Convergence Rates for Neural Operators

    stat.ML 2024-12 conditional novelty 5.0 of 10

    Two-layer neural operators trained with early-stopped gradient descent achieve the same minimax convergence rates as kernel methods in the neural tangent kernel regime.

Reference graph

Works this paper leans on

26 extracted references · 22 canonical work pages · cited by 2 Pith papers

  1. [15]

    Optimizatio n for neural operator learn- ing: Wider networks are better

    B. Shrimali, A. Banerjee, and P. Cisneros-Velarde, “Optimizatio n for neural operator learn- ing: Wider networks are better.”

    work page

  2. [17]

    Gradient descent p rovably optimizes over- parameterized neural networks,

    S. S. Du, X. Zhai, B. Poczos, and A. Singh, “Gradient descent p rovably optimizes over- parameterized neural networks,” arXiv preprint arXiv:1810.02054 , 2018

    arXiv 2018

  3. [19]

    Relu deep neural networks a nd linear finite elements,

    J. He, L. Li, J. Xu, and C. Zheng, “Relu deep neural networks a nd linear finite elements,” arXiv preprint arXiv:1807.03973 , 2018

    arXiv 2018

  4. [1]

    Physics-informe d neural networks: A deep learning framework for solving forward and inverse problems involvin g nonlinear partial differential equations,

    M. Raissi, P. Perdikaris, and G. E. Karniadakis, “Physics-informe d neural networks: A deep learning framework for solving forward and inverse problems involvin g nonlinear partial differential equations,” Journal of Computational physics , vol. 378, pp. 686–707, 2019

    work page 2019

  5. [2]

    The deep ritz method: a deep learning-based numerical algorithm for solving variational problems,

    B. Yu et al. , “The deep ritz method: a deep learning-based numerical algorithm for solving variational problems,” Communications in Mathematics and Statistics , vol. 6, no. 1, pp. 1–12, 2018

    work page 2018

  6. [3]

    Optimal approximation rate of relu networks in terms of width and depth,

    Z. Shen, H. Yang, and S. Zhang, “Optimal approximation rate of relu networks in terms of width and depth,” Journal de Math´ ematiques Pures et Appliqu´ ees, vol. 157, pp. 101–135, 2022

    work page 2022

  7. [4]

    Deep network approxim ation for smooth functions,

    J. Lu, Z. Shen, H. Yang, and S. Zhang, “Deep network approxim ation for smooth functions,” SIAM Journal on Mathematical Analysis , vol. 53, no. 5, pp. 5465–5506, 2021

    work page 2021

  8. [5]

    Error bounds for approximations with deep relu n etworks,

    D. Yarotsky, “Error bounds for approximations with deep relu n etworks,” Neural networks, vol. 94, pp. 103–114, 2017

    work page 2017

Show all 26 references
  1. [6]

    Model reduction and neural networks for parametric pdes,

    K. Bhattacharya, B. Hosseini, N. B. Kovachki, and A. M. Stuart , “Model reduction and neural networks for parametric pdes,” The SMAI journal of computational mathematics , vol. 7, pp. 121–157, 2021

    2021

  2. [7]

    Universal approximation to nonlinear oper ators by neural net- works with arbitrary activation functions and its application to dyna mical systems,

    T. Chen and H. Chen, “Universal approximation to nonlinear oper ators by neural net- works with arbitrary activation functions and its application to dyna mical systems,” IEEE transactions on neural networks , vol. 6, no. 4, pp. 911–917, 1995

    1995

  3. [8]

    Learnin g nonlinear operators via deeponet based on the universal approximation theorem of ope rators,

    L. Lu, P. Jin, G. Pang, Z. Zhang, and G. E. Karniadakis, “Learnin g nonlinear operators via deeponet based on the universal approximation theorem of ope rators,” Nature machine intelligence, vol. 3, no. 3, pp. 218–229, 2021

    2021

  4. [9]

    Fourier neural operator for parametric partial differen tial equations,

    Z. Li, N. Kovachki, K. Azizzadenesheli, B. Liu, K. Bhattacharya, A. Stuart, and A. Anand- kumar, “Fourier neural operator for parametric partial differen tial equations,” arXiv preprint arXiv:2010.08895 , 2020

    2010 arXiv

  5. [10]

    Neural operator: Learning maps between function spac es with applications to pdes,

    N. Kovachki, Z. Li, B. Liu, K. Azizzadenesheli, K. Bhattacharya , A. Stuart, and A. Anand- kumar, “Neural operator: Learning maps between function spac es with applications to pdes,” Journal of Machine Learning Research , vol. 24, no. 89, pp. 1–97, 2023

    2023

  6. [11]

    On universal appro ximation and error bounds for fourier neural operators,

    N. Kovachki, S. Lanthaler, and S. Mishra, “On universal appro ximation and error bounds for fourier neural operators,” Journal of Machine Learning Research , vol. 22, no. 290, pp. 1–76, 2021

    2021

  7. [12]

    Operator lea rning: Algorithms and analysis,

    N. B. Kovachki, S. Lanthaler, and A. M. Stuart, “Operator lea rning: Algorithms and analysis,” arXiv preprint arXiv:2402.15715 , 2024

    2024 arXiv

  8. [13]

    Error estimat es for deeponets: A deep learning framework in infinite dimensions,

    S. Lanthaler, S. Mishra, and G. E. Karniadakis, “Error estimat es for deeponets: A deep learning framework in infinite dimensions,” Transactions of Mathematics and Its Applica- tions, vol. 6, no. 1, p. tnac001, 2022

    2022

  9. [14]

    Deep nonparam etric estimation of operators between infinite dimensional spaces,

    H. Liu, H. Yang, M. Chen, T. Zhao, and W. Liao, “Deep nonparam etric estimation of operators between infinite dimensional spaces,” Journal of Machine Learning Research , vol. 25, no. 24, pp. 1–67, 2024. 15

    2024

  10. [16]

    Improved architecture s and training algorithms for deep operator networks,

    S. Wang, H. Wang, and P. Perdikaris, “Improved architecture s and training algorithms for deep operator networks,” Journal of Scientific Computing , vol. 92, no. 2, p. 35, 2022

    2022

  11. [18]

    Gradient descent fin ds global minima of deep neural networks,

    S. Du, J. Lee, H. Li, L. Wang, and X. Zhai, “Gradient descent fin ds global minima of deep neural networks,” in International conference on machine learning . PMLR, 2019, pp. 1675–1685

    2019

  12. [20]

    Gradient descent finds the global op tima of two-layer physics- informed neural networks,

    Y. Gao, Y. Gu, and M. Ng, “Gradient descent finds the global op tima of two-layer physics- informed neural networks,” in International Conference on Machine Learning . PMLR, 2023, pp. 10 676–10 707

    2023

  13. [21]

    Gin´ e and R

    E. Gin´ e and R. Nickl, Mathematical foundations of infinite-dimensional statist ical models. Cambridge university press, 2016, vol. 40

    2016

  14. [22]

    Moving beyond sub-g aussianity in high- dimensional statistics: Applications in covariance estimation and linea r regression,

    A. K. Kuchibhotla and A. Chakrabortty, “Moving beyond sub-g aussianity in high- dimensional statistics: Applications in covariance estimation and linea r regression,” In- formation and Inference: A Journal of the IMA , vol. 11, no. 4, pp. 1389–1456, 2022

    2022

  15. [23]

    A. W. Van Der Vaart, J. A. Wellner, A. W. van der Vaart, and J. A . Wellner, Weak convergence. Springer, 1996. 16 Appendix Before the proofs, we first define the events Ajr := {∃w : ‖w −wr(0)‖2 ≤R,I {wTyj ≥ 0} ⁄=I{wr(0)Tyj ≥ 0}} (24) and ˜Ai rk := {∃w : ‖w −wrk(0)‖2 ≤ ˜R,I {wTu ...

    1996

  16. [24]

    Proof of Lemma 1

    Proof of Continuous Time Analysis 7.1. Proof of Lemma 1. Proof. First, recall that H ∞ is a Kronecker product of H ∞ 1 and H ∞ 2 . The ( i,j )-th entry of H ∞ 1 is E[σ( ˜wTui)σ( ˜wTuj)] and the ( i1,j 1)-th entry of H ∞ 2 is E[yT i1yj1I{wTyi1 ≥ 0,wTyj1 ≥ 0}]. As we know, the K...

  17. [25]

    Proof of Lemma 4

    Proof of Descrete Time Analysis 8.1. Proof of Lemma 4. Proof. First, we can decompose Gt+1(ui)(yj) −Gt(ui)(yj) as follows. Gt+1(ui)(yj) −Gt(ui)(yj) =Gt+1(ui)(yj) −Gt(ui)(yj) − ⟨ ∂Gt(ui)(yj) ∂w ,w (t + 1) −w(t) ⟩ − ⟨ ∂Gt(ui)(yj) ∂ ˜w , ˜w(t + 1) − ˜w(t) ⟩ + ⟨ ∂Gt(ui)(yj) ∂w ,w ...

  18. [26]

    Let X ∼ N (0,σ 2), then for any t> 0, 2 3 t σ <P (|X| ≤ t)< 4 5 t σ

    Auxiliary Lemmas Lemma 15 (Anti-concentration of Gaussian distribution) . Let X ∼ N (0,σ 2), then for any t> 0, 2 3 t σ <P (|X| ≤ t)< 4 5 t σ. 49 Lemma 16 (Bernstein inequality, Theorem 3.1.7 in [ 21]). Let Xi, 1 ≤ i ≤ n be independent centered random variables a.s. bounded by...

Pith tools

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