Pith. sign in

REVIEW 3 major objections 6 minor 2 cited by

Neural Operators for Predictor Feedback Control of Nonlinear Delay Systems

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper shows that replacing the exact predictor in delay-compensating feedback with a learned neural operator preserves closed-loop stability whenever the approximation error is below a computable threshold.

desk verdict A genuinely useful abstraction—black-box predictors with uniform error bounds give semiglobal practical stability—but the main theorem has a real regularity gap that needs to be fixed before the result is fully rigorous. read the letter →

arxiv 2411.18964 v3 pith:L4NXYP77 submitted 2024-11-28 eess.SY cs.LGcs.SYmath.DSmath.OC

Luke Bhan , Peijia Qin , Miroslav Krstic , Yuanyuan Shi This is my paper · ORCID
classification eess.SYcs.LGcs.SYmath.DSmath.OC MSC 93C2393D15
keywords neuraloperatorspredictorfeedbacknonlineardelaysystemsinputsemiglobalpracticalstabilitybacksteppingtransformationtransportPDEoperatorlearning
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 the predictor—the component of delay-compensating feedback that looks one delay ahead—can be replaced by a learned neural operator without sacrificing closed-loop stability. The authors prove that if the neural operator approximates the exact predictor uniformly within an error bound, the nonlinear system with input delay remains semiglobally practical stable: trajectories converge to a ball of radius proportional to the approximation error, and the set of admissible initial states can be enlarged by training a more accurate operator. Because the proof uses only the uniform error bound, the result covers any black-box predictor, not just neural operators. The practical payoff is a delay-compensating controller whose predictor is evaluated in milliseconds rather than by solving expensive numerical ODEs, demonstrated on a five-link robotic manipulator.

What carries the argument

The argument is carried by three objects. First, the predictor operator $\mathcal P(X,U)(s)$ defined by the integral equation (9); Lemma 4 proves it is Lipschitz continuous in $(X,U)$ when $f$ is Lipschitz, which combines with the nonlocal neural operator universal approximation theorem (Theorem 2) to give Theorem 5. Second, the input delay is rewritten as a transport PDE $u_t=u_x$ with boundary $u(D,t)=\kappa(\hat P(t))$, and the infinite-dimensional backstepping transformation $w=u-\kappa(p)$ maps the system to a target system whose $w$-PDE has the same transport dynamics but a boundary perturbation $\kappa(P)-\kappa(\hat P)$. Third, Lemma 7 is an ISS estimate for the sup norm of this perturbed transport PDE, $$\|w(t)\|_{L^\infty[0,D]} \le $e^{{c(D-t)}}$\|w(0)\|_{L^\infty[0,D]} + $e^{{cD}}$\sup_{0\le s\le t}|\kappa(P(s))-\kappa(\hat P(s))|,$$ which, combined with the ISS Lyapunov function for the $X$ subsystem and the inverse backstepping transformations, yields the semiglobal practical bound of Theorem 8.

What would settle it

Run the closed loop of Theorem 8 with a control signal that is continuous but not $C^1$ — for example a saturated, piecewise-linear torque — measure the largest uniform predictor error over the initial-condition set, and check whether inequality (25) holds; one trajectory exiting the predicted $\epsilon$-ball would show the theorem does not govern the implemented neural-operator controller.

Watch

Extended reading notes

Core claim

The paper's central claim is Theorem 8: if the neural-operator predictor $\hat P$ approximates the exact predictor $P$ uniformly on the relevant compact set with error $\epsilon$ below a threshold $\epsilon^*(B)$ that depends on the chosen state/control bound $B$, then the closed-loop system under $U(t)=\kappa(\hat P(t))$ is $B$-semiglobal and $\epsilon$-practical stable. Concretely, for every initial condition with $|X(0)|+\sup_{\theta\in[-D,0]}|U(\theta)|\le\Omega(\epsilon,B)$, the trajectory satisfies $$|X(t)|+\sup_{t-D\le\$\theta$\le t}|U(\$\theta$)|\le\beta_2(|X(0)|+\sup_{-D\le\$\theta$\le0}|U(\$\theta$)|,t)+\alpha_1(\epsilon)$$ for all $t\ge0$, with $\epsilon^*(B)=\alpha_1^{-1}(B)$ and $\Omega$ defined through $\bar\alpha_2$. In words, any black-box predictor whose worst-case error is small enough yields a closed loop that converges to a ball of radius $\alpha_1(\epsilon)$, and the initial-condition set can be enlarged by increasing $B$ at the price of a harder approximation problem. At $\epsilon=0$ the estimate recovers the stability of the exact predictor feedback.

Load-bearing premise

The load-bearing premise is that control histories are continuously differentiable, because the transport identity and the backstepping calculation are derived under that regularity, while the neural operator output used in the experiments is only continuous.

Editorial extensions

If this is right

  • The stability guarantee transfers to any black-box predictor with a uniform error bound, so recurrent networks, neural ODEs, and generative models inherit the same closed-loop guarantee without new analysis.
  • With perfect approximation ($\epsilon=0$) the estimate collapses to the stability bound of the exact predictor feedback design, making the neural-operator result a direct generalization rather than a separate design.
  • Enlarging the admissible set of initial conditions requires the uniform approximation to hold on a larger compact set, which makes the same error bound harder to achieve; the theorem makes this trade-off explicit through $\Omega(\epsilon,B)=\bar\alpha_2^{-1}(B-\alpha_1(\epsilon))$.
  • On the five-link manipulator, the FNO predictor runs about 30 times faster than one successive-approximation step and up to about 420 times faster at finer discretizations, so approximation error is traded directly for computational speedup while practical stability is preserved.

Reading between the lines

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

  • Going beyond the paper, the main unresolved technical step is the $C^1$-versus-$C^0$ regularity gap: if the transport and backstepping identities can be shown to hold for continuous or Lipschitz control histories, then Theorem 8 would directly cover the neural-operator controller actually deployed in the experiments.
  • Because the theorem assumes only a uniform error bound on the predictor, the same argument should extend to learned predictors inside output-feedback or predictor-observer delay compensators, though the paper does not analyze those settings.
  • A likely structural limitation is compactness: Theorem 5 only guarantees approximation on compact trajectory sets, so a fully global version would probably need the approximation error to vanish with the state or an extra assumption that annihilates the boundary perturbation, as in earlier numerical-predictor results.
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 paper proposes replacing the exact predictor state in Krstic-type predictor feedback for nonlinear input-delay systems with a neural operator approximation. It defines a predictor operator on X times C1 control histories, proves its continuity and hence universal approximability by neural operators (Theorem 5), then analyzes the closed loop through a transport-PDE/backstepping representation. The main result, Theorem 8, claims B-semiglobal and epsilon-practical stability under any approximate predictor satisfying a uniform error bound. Numerical experiments on a 5-link manipulator compare FNO and DeepONet variants with successive approximations and report large speedups.

Significance. If Theorem 8 holds as stated, the paper would provide a useful and fairly general template: a black-box approximate predictor with a uniform error bound degrades the exact predictor-feedback guarantee only by an epsilon-practical term, while the initial-condition set can be enlarged at the cost of harder training. The cleanest contributions are Lemma 4, with its explicit Lipschitz constant, and Lemma 7, the ISS-type estimate for the perturbed transport PDE. The paper ships code and data and reports reproducible training details, which is a strength. However, the main theorem is not fully supported as written because of two load-bearing gaps: a regularity mismatch between the C1 setting of the analysis and the C0 signals produced by neural operators, and the absence of a proof that the closed-loop trajectory remains in the compact set on which the uniform approximation bound is assumed.

major comments (3)
  1. [Definition 3, Lemma 6, Theorem 8] The stability analysis is carried out for control histories in C1([-D,0];R^m), but the implemented controller U(t)=kappa(hat P(t)) produces only continuous control histories. Definition 3 defines the predictor operator on C1 control histories, and the proof of Lemma 6 uses p_t=p_x, which requires differentiability of p and hence of the control history. Theorem 2 only guarantees a continuous neural operator, and the paper gives no argument that kappa composed with the neural operator output is C1 in time. Thus Theorem 8 does not strictly apply to the neural-operator controller used in Section 4. Please either prove the needed C1 regularity under additional assumptions on the neural operator and on kappa, or redo the transport/backstepping and ISS analysis for C0 or weak solutions.
  2. [Theorem 5, Theorem 8] The uniform approximation bound in Theorem 5 holds on a compact set K in X times C1([-D,0);U), but Theorem 8's proof uses the bound sup |kappa(P(t))-kappa(hat P(t))| <= alpha_13(epsilon) along the closed-loop trajectory without showing that the trajectory stays in K. Theorem 8 establishes only L-infinity bounds on |X(t)| and sup |U|, and L-infinity boundedness does not imply membership in the compact C1 set K. Moreover, Assumption 4 is stated on compact domains X and U, so f may not be Lipschitz outside them. The proof needs either a forward-invariance argument for K, an a priori C1 bound on the closed-loop signals, or a conditional statement that the stability estimate holds whenever the trajectory remains in the training domain.
  3. [Appendix B.5, Eq. (63)-(64)] The passage from Eq. (63) to Eq. (64) asserts the existence of beta_2 in KL and alpha_12 in K satisfying the displayed decomposition, but the composition of alpha_11 with beta_6 and with the class-K term inside the argument is only asserted, not proved. In particular, extracting a KL function from sums and compositions involving alpha_9 and alpha_11 requires a supporting lemma; class-K functions are not generally subadditive, so steps such as bounding alpha_11(a+b) need an explicit argument. Please supply the missing lemma or a detailed calculation.
minor comments (6)
  1. [Lemma 7] The final term in Lemma 7 reads sup_{0<=s<=t} |kappa(P(t))-kappa(hat P(t))|, but the integration variable inside the supremum should be s, not t; please correct to sup_{0<=s<=t} |kappa(P(s))-kappa(hat P(s))|.
  2. [Eq. (55)] There is a missing closing parenthesis in the argument of alpha_8 in Eq. (55), which makes the displayed inequality ambiguous.
  3. [Theorem 5] The notation C1([-D,0);U) is inconsistent with Definition 3, where the codomain is R^m and U denotes the control variable; please standardize the notation for the function space and for the bounded control domain.
  4. [Section 4 and Appendix C.2] Please specify how the continuous-time control U(t) is reconstructed from the discrete-time neural operator evaluation (e.g., zero-order hold or interpolation) and how this reconstruction interacts with the C1 regularity assumed in Definition 3.
  5. [Section 1 and Appendix] The paper states that detailed proofs are deferred to the extended online version; since the journal version contains a proof of Theorem 8 in Appendix B.5, the main text should either include all proofs or clearly state which theorems depend on the online appendix.
  6. [Figure 2] The bottom-right panel uses a logarithmic y-axis with values down to about 1e-9; please state explicitly which norm of the prediction error is plotted and whether the error is evaluated online during the closed-loop simulation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Theorem 8 follows from external universal-approximation and backstepping results, with the approximation error ε as a stated hypothesis rather than a fitted or self-defined quantity.

full rationale

The claimed derivation is self-contained in the sense required here. The predictor operator P in Definition 3 is defined by the integral equation (9); Lemma 4 proves its Lipschitz continuity directly from Assumption 4 and Gronwall's inequality, not from the stability conclusion. Theorem 5 then invokes an external universal approximation theorem (Lanthaler et al., 2024, Theorem 2.1), so the existence of an ε-close neural operator is independent support, not a self-citation. The stability analysis in Sections 3.2 and 3.3 is a standard backstepping argument: Lemma 6 derives the target system with boundary perturbation κ(P)-κ(hat P), Lemma 7 gives an ISS estimate for the transport PDE by a direct weighted-norm calculation, and Theorem 8 combines this with Assumption 3 and the inverse transformations of Krstic (2010), Lemmas 12 and 13; these are either external results or directly proven in the appendix. The quantity ε is not estimated from data and then renamed as a prediction; it is a hypothesis of Theorem 5 and appears as α1(ε) in the final estimate (25), which is the normal input-to-output dependence of a robustness result, not equivalence by construction. No uniqueness theorem from the authors' own prior work is invoked to forbid alternatives, and no fitted parameter is relabeled as a prediction. The paper's self-references (Bhan et al. 2024a, 2024b, 2025) are contextual or point to the extended version and do not carry the proof. The C1-regularity concern raised by the reader (Definition 3 and Lemma 6 require U in C1, while the implemented neural operator output is only C0) is a genuine validity gap in the proof chain, but it is a regularity gap, not a circular reduction: the theorem does not assume its own conclusion. Therefore the circularity score is 0.

Assumptions & free parameters 0 free parameters · 8 assumptions · 0 invented entities

The central claim rests on standard predictor-feedback assumptions (forward completeness, GAS delay-free stabilization, ISS) and on two external mathematical pillars: the universal approximation theorem for neural operators and Krstic's backstepping transformation. No free parameters are fitted; the approximation error ε is an input to the theorem. One ad hoc regularity assumption (C1 control histories) is present but not stated.

assumptions (8)
  • domain assumption Assumption 1: Strong forward completeness of x_dot = f(x, ω).
    Statified as Assumption 1; needed to ensure no finite escape time in the predictor ODE (3).
  • domain assumption Assumption 2: Existence of a delay-free control law κ that globally asymptotically stabilizes x_dot = f(x, κ(x)).
    Assumption 2; the controller in (2) is defined via κ.
  • domain assumption Assumption 3: The closed-loop system x_dot = f(x, κ(x)+ω) is ISS.
    Assumption 3; used to build the Lyapunov function in Appendix B.5.
  • domain assumption Assumption 4: f is Lipschitz on compact domains X × U.
    Assumption 4; guarantees continuity of the predictor operator (Lemma 4).
  • standard math Theorem 2 (Lanthaler et al., 2024): universal approximation theorem for nonlocal neural operators.
    Invoked to prove Theorem 5; taken as an external result.
  • standard math Krstic (2010) backstepping transformation and invertibility (Lemmas 12 and 13 in the appendix).
    Used to pass between u and w variables in the stability proof.
  • domain assumption Continuity of the stabilizing feedback κ, used to bound |κ(P)-κ(hat P)| by a class K∞ function of ε.
    Implicit in Assumption 2; required at the end of the proof of Theorem 8 (Appendix B.5).
  • ad hoc to paper Regularity of closed-loop signals: control histories are assumed C1, while neural operator outputs are only continuous.
    The predictor operator is defined on C1 histories and the transport property p_t = p_x in Lemma 6 is derived for this case; the paper does not justify the lower regularity of the implemented controller.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Neural Operators for Predictor Feedback Control of Nonlinear Delay Systems." pith.science (2026). https://pith.science/paper/L4NXYP77

@misc{pith2026241118964,
  author       = {Pith},
  title        = {Pith review of: Neural Operators for Predictor Feedback Control of Nonlinear Delay Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L4NXYP77}},
  note         = {Machine review of arXiv:2411.18964}
}
read the original abstract

Predictor feedback designs are critical for delay-compensating controllers in nonlinear systems. However, these designs are limited in practical applications as predictors cannot be directly implemented, but require numerical approximation schemes, which become computationally prohibitive when system dynamics are expensive to compute. To address this challenge, we recast the predictor design as an operator learning problem, and learn the predictor mapping via a neural operator. We prove the existence of an arbitrarily accurate neural operator approximation of the predictor operator. Under the approximated predictor, we achieve semiglobal practical stability of the closed-loop nonlinear delay system. The estimate is semiglobal in a unique sense - one can enlarge the set of initial states as desired, though this increases the difficulty of training a neural operator, which appears practically in the stability estimate. Furthermore, our analysis holds for any black-box predictor satisfying the universal approximation error bound. We demonstrate the approach by controlling a 5-link robotic manipulator with different neural operator models, achieving significant speedups compared to classic predictor feedback schemes while maintaining closed-loop stability.

Figures

Figures reproduced from arXiv: 2411.18964 by the authors.

Figure 1
Figure 1. Overview of learning neural operator predictors for compensating delays. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Example of a single trajectory of the robotic manipulator system. The fifth state (the last arm) and the fifth control input are displayed for clarity. The gray line in the top-left figure indicates the arrival of the first control signal at the delay of D = 0.5 seconds. The bottom-right figure shows the prediction error under the NeuralOP predictor. 5. Conclusion and Future Work In this work, we developed NeuralOP … view at source ↗
Figure 3
Figure 3. In the left table, we provide all the dataset generation parameters for the experiments in Section 4 and on the right figure, we provide an example trajectory using the noise injection as described in (b) above. Model parameters and training settings: Note, for each architecture, we perform a hyperparam￾eter search based on the relative L2 test error to obtain the best configurations. We terminate the training once … view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Worse case NeuralOP predictor from the 25 trajectories in the feedback loop. C.3. Comparison of computation time under various discretization sizes and delays As mentioned in the Section 4, NeuralOP approximated predictor scales well with respect to the discretization …

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. Delay-adaptive Control of Nonlinear Systems with Approximate Neural Operator Predictors

    eess.SY 2025-08 conditional novelty 6.0 of 10

    A neural operator approximate predictor with online delay adaptation yields semi-global practical stability for nonlinear systems with unknown constant input delay.

  2. Deep Operator Neural Network Model Predictive Control

    math.OC 2025-05 conditional novelty 6.0 of 10

    MS-DeepONet computes one-shot multi-step predictions for nonlinear MIMO systems in model predictive control, with a proven universal approximation property and better benchmark performance than the standard DeepONet.

Reference graph

Works this paper leans on

55 extracted references · 36 canonical work pages · cited by 2 Pith papers

  1. [1]

    Carrillo, and Niceto R

    Ignacio Abadía, Francisco Naveros, Eduardo Ros, Richard R. Carrillo, and Niceto R. Luque. A cerebellar-based solution to the nondeterministic time delay problem in robotic control. Science Robotics, 6 0 (58), 2021. doi:10.1126/scirobotics.abf2756. URL https://www.science.org/doi/abs/10.1126/scirobotics.abf2756

    work page doi:10.1126/scirobotics.abf2756 2021

  2. [2]

    Johansson

    Miguel Aguiar, Amritam Das, and Karl H. Johansson. Learning flow functions of spiking systems. In Alessandro Abate, Mark Cannon, Kostas Margellos, and Antonis Papachristodoulou, editors, Proceedings of the 6th Annual Learning for Dynamics & Control Conference, volume 242 of Proceedings of Machine Learning Research, pages 591--602. PMLR, 15--17 Jul 2024. U...

    work page 2024

  3. [3]

    Artstein

    Z. Artstein. Linear systems with delayed controls: A reduction. IEEE Transactions on Automatic Control, 27 0 (4): 0 869--879, 1982. doi:10.1109/TAC.1982.1103023

    work page arXiv 1982

  4. [4]

    Neural operators for accelerating scientific simulations and design

    Kamyar Azizzadenesheli, Nikola Kovachki, Zongyi Li, Miguel Liu-Schiaffini, Jean Kossaifi, and Anima Anandkumar. Neural operators for accelerating scientific simulations and design. Nature Reviews Physics, 6 0 (5): 0 320--328, 2024

    2024

  5. [5]

    Novel analytical and experimental trajectory optimization of a 7-DOF B axter robot: global design sensitivity and step size analyses

    Mostafa Bagheri and Peiman Naseradinmousavi. Novel analytical and experimental trajectory optimization of a 7-DOF B axter robot: global design sensitivity and step size analyses. The International Journal of Advanced Manufacturing Technology, 93: 0 4153--4167, 2017

    work page 2017

  6. [6]

    Feedback linearization based predictor for time delay control of a high- DOF robot manipulator

    Mostafa Bagheri, Peiman Naseradinmousavi, and Miroslav Krstić. Feedback linearization based predictor for time delay control of a high- DOF robot manipulator. Automatica, 108: 0 108485, 2019. ISSN 0005-1098. doi:https://doi.org/10.1016/j.automatica.2019.06.037. URL https://www.sciencedirect.com/science/article/pii/S0005109819303371

    work page doi:10.1016/j.automatica.2019.06.037 2019

  7. [7]

    Input-to-state stability in sup norms for hyperbolic systems with boundary disturbances

    Georges Bastin, Jean-Michel Coron, and Amaury Hayat. Input-to-state stability in sup norms for hyperbolic systems with boundary disturbances. Nonlinear Analysis, 208: 0 112300, 2021. ISSN 0362-546X. doi:https://doi.org/10.1016/j.na.2021.112300. URL https://www.sciencedirect.com/science/article/pii/S0362546X21000389

    work page arXiv 2021

  8. [8]

    Compensation of Time-Varying Input and State Delays for Nonlinear Systems

    Nikolaos Bekiaris-Liberis and Miroslav Krstic. Compensation of Time-Varying Input and State Delays for Nonlinear Systems . Journal of Dynamic Systems, Measurement, and Control, 134 0 (1): 0 011009, 12 2011. ISSN 0022-0434. doi:10.1115/1.4005278. URL https://doi.org/10.1115/1.4005278

    work page doi:10.1115/1.4005278 2011

Show all 55 references
  1. [9]

    Nonlinear Control Under Nonconstant Delays

    Nikolaos Bekiaris-Liberis and Miroslav Krstic. Nonlinear Control Under Nonconstant Delays. Society for Industrial and Applied Mathematics, Philadelphia, PA, 2013. doi:10.1137/1.9781611972856. URL https://epubs.siam.org/doi/abs/10.1137/1.9781611972856

    2013 doi

  2. [10]

    Predictor-feedback stabilization of multi-input nonlinear systems

    Nikolaos Bekiaris-Liberis and Miroslav Krstic. Predictor-feedback stabilization of multi-input nonlinear systems. IEEE Transactions on Automatic Control, 62 0 (2): 0 516--531, 2017. doi:10.1109/TAC.2016.2558293

    2017

  3. [11]

    Neural operators for predictor feedback control of nonlinear delay systems

    Luke Bhan, Peijia Qin, Miroslav Krstic, and Yuanyuan Shi. Neural operators for predictor feedback control of nonlinear delay systems. arXiv preprint arXiv:2411.18964, 2024 a

    2024 arXiv

  4. [12]

    Neural operators for bypassing gain and control computations in PDE backstepping

    Luke Bhan, Yuanyuan Shi, and Miroslav Krstic. Neural operators for bypassing gain and control computations in PDE backstepping. IEEE Transactions on Automatic Control, 69 0 (8): 0 5310--5325, 2024 b . doi:10.1109/TAC.2023.3347499

    2024

  5. [13]

    Adaptive control of reaction–diffusion PDEs via neural operator-approximated gain kernels

    Luke Bhan, Yuanyuan Shi, and Miroslav Krstic. Adaptive control of reaction–diffusion PDEs via neural operator-approximated gain kernels. Systems & Control Letters, 195: 0 105968, 2025. ISSN 0167-6911. doi:https://doi.org/10.1016/j.sysconle.2024.105968. URL https://www.scienced...

    2025

  6. [14]

    Delay-adaptive control for nonlinear systems

    Delphine Bresch-Pietri and Miroslav Krstic. Delay-adaptive control for nonlinear systems. IEEE Transactions on Automatic Control, 59 0 (5): 0 1203--1218, 2014. doi:10.1109/TAC.2014.2298711

    2014

  7. [15]

    Output feedback control of linear systems with input, state and output delays by chains of predictors

    Filippo Cacace and Alfredo Germani. Output feedback control of linear systems with input, state and output delays by chains of predictors. Automatica, 85: 0 455--461, 2017. ISSN 0005-1098. doi:https://doi.org/10.1016/j.automatica.2017.08.013. URL https://www.sciencedirect.com/...

    2017 doi

  8. [16]

    Ricky T. Q. Chen, Yulia Rubanova, Jesse Bettencourt, and David K Duvenaud. Neural ordinary differential equations. In S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett, editors, Advances in Neural Information Processing Systems, volume 31. Curra...

    2018

  9. [17]

    Learning stochastic dynamical system via flow map operator

    Yuan Chen and Dongbin Xiu. Learning stochastic dynamical system via flow map operator. Journal of Computational Physics, 508: 0 112984, 2024. ISSN 0021-9991. doi:https://doi.org/10.1016/j.jcp.2024.112984. URL https://www.sciencedirect.com/science/article/pii/S002199912400233X

    2024

  10. [18]

    On the properties of neural machine translation: Encoder -- decoder approaches

    Kyunghyun Cho, Bart van Merri \"e nboer, Dzmitry Bahdanau, and Yoshua Bengio. On the properties of neural machine translation: Encoder -- decoder approaches. In Dekai Wu, Marine Carpuat, Xavier Carreras, and Eva Maria Vecchi, editors, Proceedings of SSST -8, Eighth Workshop on...

    2014 doi

  11. [19]

    Physics-guided, physics-informed, and physics-encoded neural networks and operators in scientific computing: Fluid and solid mechanics

    Salah A Faroughi, Nikhil M Pawar, C \'e lio Fernandes, Maziar Raissi, Subasish Das, Nima K Kalantari, and Seyed Kourosh Mahjour. Physics-guided, physics-informed, and physics-encoded neural networks and operators in scientific computing: Fluid and solid mechanics. Journal of C...

    2024

  12. [20]

    Gnot: A general neural operator transformer for operator learning

    Zhongkai Hao, Zhengyi Wang, Hang Su, Chengyang Ying, Yinpeng Dong, Songming Liu, Ze Cheng, Jian Song, and Jun Zhu. Gnot: A general neural operator transformer for operator learning. In International Conference on Machine Learning, pages 12556--12569. PMLR, 2023

    2023

  13. [21]

    Time delay compensation for nonlinear processes

    Michael A Henson and Dale E Seborg. Time delay compensation for nonlinear processes. Industrial & engineering chemistry research, 33 0 (6): 0 1493--1500, 1994

    1994

  14. [22]

    Jankovic

    M. Jankovic. Control Lyapunov-Razumikhin functions and robust stabilization of time delay systems. IEEE Transactions on Automatic Control, 46 0 (7): 0 1048--1060, 2001. doi:10.1109/9.935057

    2001 doi

  15. [24]

    Recursive predictor design for state and output feedback controllers for linear time delay systems

    Mrdjan Jankovic. Recursive predictor design for state and output feedback controllers for linear time delay systems. Automatica, 46 0 (3): 0 510--517, 2010. ISSN 0005-1098. doi:https://doi.org/10.1016/j.automatica.2010.01.021. URL https://www.sciencedirect.com/science/article/...

    2010 doi

  16. [25]

    Finite-time global stabilization by means of time-varying distributed delay feedback

    Iasson Karafyllis. Finite-time global stabilization by means of time-varying distributed delay feedback. SIAM Journal on Control and Optimization, 45 0 (1): 0 320--342, 2006. doi:10.1137/040616383. URL https://doi.org/10.1137/040616383

    2006 doi

  17. [26]

    Predictor Feedback for Delay Systems: Implementations and Approximations

    Iasson Karafyllis and Miroslav Krstic. Predictor Feedback for Delay Systems: Implementations and Approximations. Birkh\" a user Cham, 2017

    2017

  18. [27]

    Input-to-State Stability for PDEs

    Iasson Karafyllis and Miroslav Krstic. Input-to-State Stability for PDEs. Springer Cham, 2018

    2018

  19. [28]

    Stability results for the continuity equation

    Iasson Karafyllis and Miroslav Krstic. Stability results for the continuity equation. Systems & Control Letters, 135: 0 104594, 2020. ISSN 0167-6911. doi:https://doi.org/10.1016/j.sysconle.2019.104594. URL https://www.sciencedirect.com/science/article/pii/S016769111930204X

    2020

  20. [29]

    On universal approximation and error bounds for Fourier neural operators

    Nikola Kovachki, Samuel Lanthaler, and Siddhartha Mishra. On universal approximation and error bounds for Fourier neural operators. J. Mach. Learn. Res., 22 0 (1), January 2021. ISSN 1532-4435

    2021

  21. [30]

    Costas Kravaris and Raymond A. Wright. Deadtime compensation for nonlinear processes. AIChE Journal, 35 0 (9): 0 1535--1542, 1989. doi:https://doi.org/10.1002/aic.690350914. URL https://aiche.onlinelibrary.wiley.com/doi/abs/10.1002/aic.690350914

    1989 doi

  22. [31]

    Input delay compensation for forward complete and strict-feedforward nonlinear systems

    Miroslav Krstic. Input delay compensation for forward complete and strict-feedforward nonlinear systems. IEEE Transactions on Automatic Control, 55 0 (2): 0 287--303, 2010. doi:10.1109/TAC.2009.2034923

    2010

  23. [32]

    Kokotovic, and Ioannis Kanellakopoulos

    Miroslav Krstic, Petar V. Kokotovic, and Ioannis Kanellakopoulos. Nonlinear and Adaptive Control Design. John Wiley & Sons, Inc., USA, 1st edition, 1995. ISBN 0471127329

    1995

  24. [33]

    Kwon and A

    W. Kwon and A. Pearson. Feedback stabilization of linear systems with delayed control. IEEE Transactions on Automatic Control, 25 0 (2): 0 266--269, 1980. doi:10.1109/TAC.1980.1102288

    1980

  25. [34]

    Adaptive neural-operator backstepping control of a benchmark hyperbolic PDE , 2024 a

    Maxence Lamarque, Luke Bhan, Yuanyuan Shi, and Miroslav Krstic. Adaptive neural-operator backstepping control of a benchmark hyperbolic PDE , 2024 a . URL https://arxiv.org/abs/2401.07862

    2024 arXiv

  26. [35]

    Gain scheduling with a neural operator for a transport PDE with nonlinear recirculation, 2024 b

    Maxence Lamarque, Luke Bhan, Rafael Vazquez, and Miroslav Krstic. Gain scheduling with a neural operator for a transport PDE with nonlinear recirculation, 2024 b . URL https://arxiv.org/abs/2401.02511

    2024 arXiv

  27. [36]

    Error estimates for DeepONets: a deep learning framework in infinite dimensions

    Samuel Lanthaler, Siddhartha Mishra, and George E Karniadakis. Error estimates for DeepONets: a deep learning framework in infinite dimensions . Transactions of Mathematics and Its Applications, 6 0 (1): 0 tnac001, 03 2022. ISSN 2398-4945. doi:10.1093/imatrm/tnac001. URL https...

    2022 doi

  28. [37]

    Samuel Lanthaler, Zongyi Li, and Andrew M. Stuart. Nonlocality and Nonlinearity Implies Universality in Operator Learning , 2024. URL https://arxiv.org/abs/2304.13221

    2024 arXiv

  29. [38]

    Fourier neural operator for parametric partial differential equations

    Zongyi Li, Nikola Borislavov Kovachki, Kamyar Azizzadenesheli, Burigede liu, Kaushik Bhattacharya, Andrew Stuart, and Anima Anandkumar. Fourier neural operator for parametric partial differential equations. In International Conference on Learning Representations, 2021. URL htt...

    2021

  30. [39]

    Decoupled weight decay regularization

    Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. In International Conference on Learning Representations, 2019. URL https://openreview.net/forum?id=Bkg6RiCqY7

    2019

  31. [40]

    Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators

    Lu Lu, Pengzhan Jin, Guofei Pang, Zhongqiang Zhang, and George Em Karniadakis. Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators. Nature Machine Intelligence, 3 0 (3): 0 218--229, Mar 2021. ISSN 2522-5839. doi:10.1038/s42256-02...

    2021 doi

  32. [41]

    Lum, Jacob Rosen, Hawkeye King, Diana C.W

    Mitchell J.H. Lum, Jacob Rosen, Hawkeye King, Diana C.W. Friedman, Thomas S. Lendvay, Andrew S. Wright, Mika N. Sinanan, and Blake Hannaford. Teleoperation in surgical robotics – network latency effects on surgical performance. In 2009 Annual International Conference of the IE...

    2009

  33. [42]

    Distributed optimal voltage control with asynchronous and delayed communication

    Sindri Magn \'u sson, Guannan Qu, and Na Li. Distributed optimal voltage control with asynchronous and delayed communication. IEEE Transactions on Smart Grid, 11 0 (4): 0 3469--3482, 2020

    2020

  34. [43]

    Mazenc and P.-A

    F. Mazenc and P.-A. Bliman. Backstepping design for time-delay nonlinear systems. In 42nd IEEE International Conference on Decision and Control (IEEE Cat. No.03CH37475), volume 5, pages 4551--4556 Vol.5, 2003. doi:10.1109/CDC.2003.1272269

    2003 arXiv

  35. [44]

    Subpredictor approach for event-triggered control of discrete-time systems with input delays

    Frédéric Mazenc, Michael Malisoff, Corina Barbalata, and Zhong-Ping Jiang. Subpredictor approach for event-triggered control of discrete-time systems with input delays. European Journal of Control, 68: 0 100664, 2022. ISSN 0947-3580. doi:https://doi.org/10.1016/j.ejcon.2022.10...

    2022

  36. [45]

    Neural operator learning for long-time integration in dynamical systems with recurrent neural networks, 2024

    Katarzyna Michałowska, Somdatta Goswami, George Em Karniadakis, and Signe Riemer-Sørensen. Neural operator learning for long-time integration in dynamical systems with recurrent neural networks, 2024. URL https://arxiv.org/abs/2303.02243

    2024 arXiv

  37. [46]

    Neural operators for PDE backstepping control of first-order hyperbolic PIDE with recycle and delay

    Jie Qi, Jing Zhang, and Miroslav Krstic. Neural operators for PDE backstepping control of first-order hyperbolic PIDE with recycle and delay. Systems & Control Letters, 185: 0 105714, 2024. ISSN 0167-6911. doi:https://doi.org/10.1016/j.sysconle.2024.105714. URL https://www.sci...

    2024

  38. [47]

    Gordon, and J

    Stephane Ross, Geoffrey J. Gordon, and J. Andrew Bagnell. A reduction of imitation learning and structured prediction to no-regret online learning, 2011. URL https://arxiv.org/abs/1011.0686

    2011 arXiv

  39. [48]

    Jacob H Seidman, Georgios Kissas, Paris Perdikaris, and George J. Pappas. NOMAD : Nonlinear manifold decoders for operator learning. In Alice H. Oh, Alekh Agarwal, Danielle Belgrave, and Kyunghyun Cho, editors, Advances in Neural Information Processing Systems, 2022. URL https...

    2022

  40. [49]

    Predictor-based compensation for electromechanical delay during neuromuscular electrical stimulation

    Nitin Sharma, Chris M Gregory, and Warren E Dixon. Predictor-based compensation for electromechanical delay during neuromuscular electrical stimulation. IEEE Trans. Neural Syst. Rehabil. Eng., 19 0 (6): 0 601--611, December 2011

    2011

  41. [50]

    Machine learning accelerated pde backstepping observers

    Yuanyuan Shi, Zongyi Li, Huan Yu, Drew Steeves, Anima Anandkumar, and Miroslav Krstic. Machine learning accelerated pde backstepping observers. In IEEE 61st Conference on Decision and Control, pages 5423--5428. IEEE, 2022

    2022

  42. [51]

    Development of a dynamics model for the baxter robot

    Alex Smith, Chenguang Yang, Chunxu Li, Hongbin Ma, and Lijun Zhao. Development of a dynamics model for the baxter robot. In 2016 IEEE International Conference on Mechatronics and Automation, pages 1244--1249, 2016. doi:10.1109/ICMA.2016.7558740

    2016

  43. [52]

    Closer control of loops with dead time

    Otto JM Smith. Closer control of loops with dead time. Chemical engineering progress, 53: 0 217--219, 1957

    1957

  44. [53]

    On characterizations of the input-to-state stability property

    Eduardo D Sontag and Yuan Wang. On characterizations of the input-to-state stability property. Systems & Control Letters, 24 0 (5): 0 351--359, 1995

    1995

  45. [54]

    Cascade structure predictive observer design for consensus control with applications to uavs formation flying

    Jianan Wang, Zhengyang Zhou, Chunyan Wang, and Zhengtao Ding. Cascade structure predictive observer design for consensus control with applications to uavs formation flying. Automatica, 121: 0 109200, 2020. ISSN 0005-1098. doi:https://doi.org/10.1016/j.automatica.2020.109200. U...

    2020

  46. [55]

    Neural operators for boundary stabilization of stop-and-go traffic

    Yihuai Zhang, Ruiguo Zhong, and Huan Yu. Neural operators for boundary stabilization of stop-and-go traffic. In Alessandro Abate, Mark Cannon, Kostas Margellos, and Antonis Papachristodoulou, editors, Proceedings of the 6th Annual Learning for Dynamics & Control Conference, vo...

    2024

  47. [56]

    Truncated Predictor Feedback for Time-Delay Systems

    Bin Zhou. Truncated Predictor Feedback for Time-Delay Systems. Springer Berlin, Heidelberg, 2014

    2014

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.