Controller Design for Bilinear Neural Feedback Loops

Dhruv Shah; Jorge Cort\'es

REVIEW 4 major objections 4 minor 21 references

Controller Design for Bilinear Neural Feedback Loops

T0 review · 4 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 proves that a class of discrete-time bilinear systems with a neural network in the loop can be locally exponentially stabilized by gains computed from a single linear matrix inequality feasibility problem, with a certified…

desk verdict A plausible LMI synthesis framework for bilinear neural feedback loops, with clean reformulation but a load-bearing well-posedness step left unverified. read the letter →

arxiv 2505.24106 v1 pith:TOYSPGZL submitted 2025-05-30 eess.SY cs.SY

Dhruv Shah , Jorge Cortés This is my paper

classification eess.SYcs.SY

keywords bilinearsystemsneuralfeedbackloopsimplicitnetworkslinearfractionalrepresentationsquadraticconstraintsmatrixinequalitieslocalexponentialstabilityparametervaryingcontrol

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 attacks a practical bottleneck: when a neural network sits inside a control loop, the closed loop usually comes with no formal stability guarantee. It treats a discrete-time bilinear system whose nonlinear terms are learned neural networks, and shows that this system can be recast as a linear system connected to an uncertainty block. The paper then proves that if a certain system of linear matrix inequalities is feasible, the resulting gains make the equilibrium locally exponentially stable, and the ellipsoid $Z_{\mathrm{ROA}}=\{z: z^T P^{-1} z\le 1\}$ is a forward-invariant safe region contained in the desired operating region. Because the inequalities are convex, the gains can be computed offline and the online controller amounts to evaluating an implicit algebraic equation. The contribution is a convex certificate that gives neural feedback loops quantifiable stability and safety margins.

What carries the argument

The load-bearing object is the coupled LMI system (21) from Theorem 4.5. The transformation that makes it possible is the linear fractional representation: the paper rewrites the bilinear-neural closed loop as a nominal linear interconnection whose feedback path is an uncertain matrix $\Delta_c = \mathrm{diag}(\Delta_m, \mathrm{diag}_l(\Delta_{k_\psi}), \Delta_\phi, \Delta_\psi)$, with the neural activations, the bilinear products $(z\otimes I_m)u$, and the terms $(z\otimes I_{k_\psi})s_\psi$ all absorbed into $\Delta_c$. The uncertainty sets are carved out by quadratic constraints: slope-restricted activations satisfy the incremental QCs of Lemma 2.2, and the bilinear terms satisfy Lemma 4.3 whenever the state stays in $Z$. With those sets in place, the synthesis LMIs are derived from a full-block multiplier condition for robust quadratic performance; feasibility gives the gains and, simultaneously, the Lyapunov matrix $P$ that defines $Z_{\mathrm{ROA}}$.

What would settle it

For the 4D example in Section IV, after the LMIs are declared feasible, evaluate the controller equation (19) by fixed-point iteration from two different starting points at a dense grid of states inside $Z_{\mathrm{ROA}}$; a state with two distinct fixed points, or a trajectory starting in $Z_{\mathrm{ROA}}$ that leaves $Z_{\mathrm{ROA}}$, would contradict Theorem 4.5.

Watch

Extended reading notes

Core claim

The paper's core discovery is that robust control machinery—linear fractional representations, quadratic-constraint uncertainty sets, and full-block multiplier LMIs—can be applied to a generalized bilinear system with neural networks in the loop. Concretely, Theorem 4.5 asserts that if the LMIs in (21) are feasible, then the controller gains $K_z, K_u, K_{w_\psi}, K_\phi, K_\psi$ make the closed-loop system (20) well-posed and locally exponentially stable, with region of attraction $Z_{\mathrm{ROA}}=\{z : z^T P^{-1} z \le 1\}$ forward invariant and contained in the user-specified region of interest $Z$. The proof constructs $V(z)=z^T P^{-1} z$ as a Lyapunov function that works against every matrix in the combined uncertainty set $\Delta_c$, then uses the second LMI to force $Z_{\mathrm{ROA}}$ inside $Z$ by an S-procedure argument. This turns a nonlinear, non-affine-in-input stabilization problem into an offline convex feasibility check.

Load-bearing premise

Everything rests on the claim that the implicit equation (19), which defines the control input $u$ as a fixed point of an expression involving $u$ on both sides, has a unique solution for every state in $Z_{\mathrm{ROA}}$; the paper asserts this follows from LMI feasibility but provides no dedicated well-posedness argument.

Editorial extensions

If this is right

A feasible LMI computation yields controller gains that carry a provable local exponential stability guarantee and a forward-invariant safe set, so the approach replaces heuristic tuning with a certificate.
Because the synthesis is convex and offline, the same framework can be rerun when the neural weights change or when the region of interest is tightened, without changing the online control law's structure.
The reformulation covers the broader system class (9) with neural terms in $\phi_u,\phi_z,\Psi_z,\Psi_u,\Psi$, so the method is not tied to the particular bilinear structure used in the statement.
For neural feedback loops arising from lifted representations or deep-learning-based system identification, the LMI certificate provides stability and robustness analysis that the raw learned model does not provide.

Reading between the lines

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

Beyond the paper, the well-posedness certificate for the implicit controller suggests that a simple fixed-point iteration $u_{k+1}=K_z z+K_u(z\otimes I_m)u_k+\cdots$ should converge quickly inside $Z_{\mathrm{ROA}}$; testing contraction rates numerically would give an implementable real-time solver.
Beyond the paper, the same quadratic-constraint abstraction could be run in reverse: given an LMI-infeasible instance, a dual certificate might identify a specific destabilizing activation nonlinearity, extending infeasibility-duality results known for linear neural loops to bilinear ones.
Beyond the paper, the certified ellipsoid $Z_{\mathrm{ROA}}$ could be used as a safety filter in a perception–control stack: an upstream planner only needs to keep the lifted state inside the ellipsoid to inherit the stability guarantee.

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.

Desk Editor's Note

read the letter

Plainly, the useful thing here is a convex offline synthesis for a system class nobody had done before: bilinear dynamics with neural networks in the loop. The LFR reformulation in Prop 4.1 is clean, and the shift to the equilibrium is handled carefully. The QC-based uncertainty sets for the bilinear terms (Lemma 4.3) are a nice adaptation of the gain-scheduling idea, and the simulation against the baseline controller shows the expected benefit. That part I would take seriously.

The soft spots are real but not fatal. Lemma 4.4 — the inverse multiplier formula — is stated without proof, and the whole LMI construction depends on it. The alpha=0 (ReLU) case, which is what the simulation actually uses, gets only a sketch in Remark 4.6; that needs a proper derivation. The bigger issue is well-posedness of the implicit controller (19). The proof of Theorem 4.5 disposes of it with a single citation to [15, Thm 2] and no verification that the hypotheses of that theorem hold for this particular D and uncertainty set. The stress-test note is right: if the algebraic loop in u has multiple solutions for some z in ZROA, the nominal closed-loop system does not exist and the stability certificate is void. It may well be that LMI feasibility implies well-posedness, but the paper does not show the pointwise argument needed to conclude that.

So the central claim is plausible, the framework is genuinely useful, and the gaps are about rigor and completeness rather than a demonstrated counterexample. The right audience is people working on LMI-based certification of learning-controlled systems; they will want to see the full proof before building on it. I would send it to peer review, but require the omitted proof and a dedicated well-posedness argument before acceptance.

Referee Report

4 major / 4 minor

Summary. The paper considers discrete-time generalized bilinear systems with neural-network nonlinearities in the feedback loop. It rewrites the system as a linear fractional representation (LFR) with the nonlinearities absorbed into structured uncertainties, describes the uncertainty sets by quadratic constraints, and proposes an LMI-based static state-feedback controller that is defined implicitly through a fixed-point equation. The main result, Theorem 4.5, claims that if the LMIs in (21) are feasible, the closed loop is well posed and locally exponentially stable with a forward-invariant ellipsoidal region of attraction contained in a prescribed set Z. A four-dimensional numerical example with ReLU MLP approximations is used to illustrate the approach.

Significance. If the claims can be made fully rigorous, the framework would extend LMI-based robust-control synthesis from linear or gain-scheduled plants to a class of neural feedback loops with non-affine input dependence, while providing an explicit region of attraction. The general architecture is plausible, the use of LFRs and quadratic constraints is standard, and the paper correctly identifies the conservatism introduced by over-approximation. However, the manuscript currently contains several load-bearing gaps: an ordering inconsistency in the definition of the combined uncertainty signals, an unproved multiplier-inverse lemma on which the LMI derivation depends, an unverified well-posedness assertion for the implicit controller, and a simulation example that is not covered by the theorem's hypotheses because it uses ReLU activations.

major comments (4)

[§IV-B, Eq. (18) and §IV-C, Eq. (20)] The combined uncertainty vectors are misordered. In §IV-B, q = vec(wu, wψ, sϕ, sψ) and p = vec(u, sϕ, vϕ, vψ) are defined, but the feedback (14b) gives wu = Δm u, wψ = diagl(Δkψ) sψ, sϕ = Δϕ vϕ, and sψ = Δψ vψ. Hence the second component of p must be sψ, not sϕ; sϕ is an output of Δϕ, not an input to Δc. The matrix D in (20) and the block definitions in the LMI use the correct order [u; sψ; vϕ; vψ]. As written, the identity q = Δc p does not represent (14b), and the block alignments in Q, S, R and in the LMI variables are not well defined. This ordering must be corrected and all subsequent block matrices checked against the corrected ordering.
[§IV-C, Lemma 4.4] Lemma 4.4 is stated with 'the proof is omitted for space reasons', but it is used directly in the proof of Theorem 4.5 to define the inverse multiplier matrices and the factorizations SL, SR that enter the synthesis LMI. Without a proof or a complete reference that covers this exact block-diagonal and tensor-product structure, the LMI derivation is unsupported. Please include the proof, or state precisely which result in the literature establishes these inverse formulas.
[§IV-C, end of proof of Theorem 4.5] The final sentence of the proof asserts that 'well posedness of the closed-loop system and hence of the implicit equations (19) is ensured if the LMI (21) is feasible, cf. [15, Thm 2]', but the paper does not verify that the hypotheses of [15, Thm 2] hold for this particular algebraic loop. In particular, no argument is given that I - D Δc is invertible for every Δc in the combined uncertainty set Δc. Equation (19) is a nonlinear fixed-point equation in u, so non-uniqueness or non-existence of solutions for some z in ZROA would void the Lyapunov argument. A dedicated well-posedness proof is needed; a citation alone is not sufficient here.
[§IV-C, Remark 4.6, and §IV-D simulation] Theorem 4.5 assumes αβ < 0, but the numerical example uses ReLU activations, for which α = 0. Remark 4.6 only sketches that the case α = 0 can be handled 'by removing the last kϕ + lkψ rows and columns' from the left LMI in (21), without giving the resulting LMI or its proof. Since the simulation is the main demonstration of the method, either the α = 0 case must be stated and proved as a theorem with an explicit LMI, or the simulation must be presented as not covered by the theorem. The current gap between the theorem's hypotheses and the example is load-bearing.

minor comments (4)

[§III, Eq. (7)] The dimension given for \tilde{D} is R^{m×lm}, but since \tilde{D}(z ⊗ u) must be in R^l, it should be R^{l×lm}.
[§IV-C] The controller parameters are introduced as Kz, Ku, Kwψ, Ksϕ, Ksψ in the text, but the theorem and synthesis statements use Kϕ and Kψ; please make the notation consistent.
[§V] The conclusions refer to 'ICQs' while Remark 4.7 uses 'IQCs'; please unify the abbreviation.
[§I] The introductory claim that the controller is 'richer than a vanilla linear controller with certified performance' is not quantified in the paper; the formal results certify local exponential stability and a region of attraction, but no performance index is defined beyond that.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the LMI-derived stability certificate rests on external robust-control theorems, and no fitted quantity is renamed as a prediction.

full rationale

The central claim of the paper is a sufficient LMI condition for local exponential stability of the closed-loop LFR (20) with an over-approximated uncertainty set Δc. The derivation chain is: reformulate the bilinear-NFL system as an LFR (Proposition 4.1), shift the equilibrium (Lemma 4.2), characterize the uncertainty sets via quadratic constraints (Lemma 4.3 and equation (18)), and then apply the external full-block multiplier machinery of Scherer [15] to obtain the synthesis LMIs (21). No parameter is fitted to data and then presented as a prediction; the simulated trajectories merely illustrate the certified controller, and the LMI feasibility is checked numerically rather than inferred from closed-loop behavior. The only self-citation, [9] (Haseli and Cortés), appears in the introduction as motivational background for why bilinear neural feedback loops arise from Koopman-based lifted representations, and it is not used to justify the load-bearing stability theorem. The proof of Theorem 4.5 relies on [15, Thm 2] and [15, Lemma A.1], which are independent external results. Even the reviewer's identified weakness — the assertion that feasibility of (21) ensures well-posedness of the implicit controller equation (19), cited to [15, Thm 2] with no dedicated verification of its hypotheses — is a correctness or robustness concern about an external-theorem application, not a circularity. There is no self-definitional step, no fitted input renamed as a prediction, and no load-bearing self-citation chain. The paper is therefore assigned circularity score 0.

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

No data-fitted constants appear in the synthesis; the LMI decision variables P, Lambda, T, L, and nu are optimization variables chosen by the solver to satisfy sufficient conditions, not calibrated to data. The uncertainty matrices Delta_m, Delta_kpsi, Delta_phi, Delta_psi are analytical constructs used to abstract nonlinearities, not new physical entities. The main axioms are the stabilizability/equilibrium assumption, the slope-restricted INN structure, the cited bilinear uncertainty lemma, and the standard full-block multiplier machinery.

assumptions (5)

domain assumption System (7) is stabilizable and an equilibrium (z*,u*) exists (Section III).
Without this, the controller design problem is undefined; the paper assumes it without proof.
domain assumption Neural network nonlinearities phi, Psi are implicit neural networks with the same componentwise activation xi slope-restricted on [alpha,beta] (Sections III and IV-A).
The QCs and LMI derivation rely on this slope bound.
standard math Lemma 4.3: for Z as in (8), Delta in Delta_m iff Delta = z otimes Im for some z in Z (Lemma 4.3, cited from [10]).
External theorem from prior literature; the paper does not prove it, and it is needed for the bilinear uncertainty sets.
standard math The full-block multiplier stability conditions [15, Thm 2] and dualization lemma [15, Lemma A.1] hold for the shifted closed-loop LFR (Theorem 4.5 proof).
External robust control results used to derive the LMIs; assumed valid.
domain assumption LMI (21) is feasible (Theorem 4.5 assumption).
The theorem is conditional on feasibility; infeasibility means no guaranteed controller from this method.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Controller Design for Bilinear Neural Feedback Loops." pith.science (2026). https://pith.science/paper/TOYSPGZL

@misc{pith2026250524106,
  author       = {Pith},
  title        = {Pith review of: Controller Design for Bilinear Neural Feedback Loops},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TOYSPGZL}},
  note         = {Machine review of arXiv:2505.24106}
}

read the original abstract

This paper considers a class of bilinear systems with a neural network in the loop. These arise naturally when employing machine learning techniques to approximate general, non-affine in the input, control systems. We propose a controller design framework that combines linear fractional representations and tools from linear parameter varying control to guarantee local exponential stability of a desired equilibrium. The controller is obtained from the solution of linear matrix inequalities, which can be solved offline, making the approach suitable for online applications. The proposed methodology offers tools for stability and robustness analysis of deep neural networks interconnected with dynamical systems.

Figures

Figures reproduced from arXiv: 2505.24106 by the authors.

**Figure 1.** Time evolution of (24) using the controller (19), which accounts for the nonlinearities ϕ1 and ϕ2. Results are compared with the controller from [10], which does not take into account these nonlinearities. Remark 4.7: (Computational Complexity and LMI Feasibility): LMI-based controller synthesis can become computationally expensive for high-dimensional systems. This can be mitigated by neural network pruning [19]… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

21 extracted references · 16 canonical work pages

[1]

Reachability analysis of neural feedback loops,

M. Everett, G. Habibi, C. Sun, and J. P. How, “Reachability analysis of neural feedback loops,” IEEE Access, vol. 9, pp. 163938–163953, 2021

work page 2021
[2]

Reach-SDP: Reach- ability analysis of closed-loop systems with neural network controllers via semidefinite programming,

H. Hu, M. Fazlyab, M. Morari, and G. J. Pappas, “Reach-SDP: Reach- ability analysis of closed-loop systems with neural network controllers via semidefinite programming,” in IEEE Conf. on Decision and Control , (Jeju Island, Republic of Korea), pp. 5929–5934, Dec. 2020

work page 2020
[3]

ReachNN: Reachability analysis of neural-network controlled systems,

C. Huang, J. Fan, W. Li, X. Chen, and Q. Zhu, “ReachNN: Reachability analysis of neural-network controlled systems,” ACM Transactions on Embedded Computing Systems , vol. 18, no. 5s, pp. 1–22, 2019

work page 2019
[4]

Reachability analysis for neural feedback systems using regressive polynomial rule inference,

S. Dutta, X. Chen, and S. Sankaranarayanan, “Reachability analysis for neural feedback systems using regressive polynomial rule inference,” in Hybrid Systems: Computation and Control , pp. 157–168, 2019

work page 2019
[5]

Robust guarantees for perception-based control,

S. Dean, N. Matni, B. Recht, and V . Ye, “Robust guarantees for perception-based control,” in Learning for Dynamics and Control , pp. 350–360, PMLR, 2020

work page 2020
[6]

Stability Margins of Neural Network Controllers

N. Junnarkar, M. Arcak, and P. Seiler, “Stability margins of neural network controllers,” arXiv preprint arXiv:2409.09184 , 2024

work page Pith review arXiv 2024
[7]

Synthesizing neural network controllers with closed-loop dissipativity guarantees,

N. Junnarkar, M. Arcak, and P. Seiler, “Synthesizing neural network controllers with closed-loop dissipativity guarantees,” arXiv preprint arXiv:2404.07373, 2024

arXiv 2024
[8]

Deep networks for system identification: a survey,

G. Pillonetto, A. Aravkin, D. Gedon, L. Ljung, A. H. Ribeiro, and T. B. Sch¨on, “Deep networks for system identification: a survey,” Automatica, vol. 171, p. 111907, 2025

work page 2025

Show all 21 references

[9]

Modeling nonlinear control systems via Koop- man control family: universal forms and subspace invariance proximity,

M. Haseli and J. Cort ´es, “Modeling nonlinear control systems via Koop- man control family: universal forms and subspace invariance proximity,” Available at https://arxiv.org/abs/2307.15368, 2023

2023 arXiv
[10]

Control of bilinear systems using gain-scheduling: Stability and performance guarantees,

R. Str ¨asser, J. Berberich, and F. Allg ¨ower, “Control of bilinear systems using gain-scheduling: Stability and performance guarantees,” in IEEE Conf. on Decision and Control , (Singapore), pp. 4674–4681, Dec. 2023

2023
[11]

Implicit deep learning,

L. El Ghaoui, F. Gu, B. Travacca, A. Askari, and A. Tsai, “Implicit deep learning,” SIAM Journal on Mathematics of Data Science , vol. 3, no. 3, pp. 930–958, 2021

2021
[12]

Efficient and accurate estimation of Lipschitz constants for deep neural networks,

M. Fazlyab, A. Robey, H. Hassani, M. Morari, and G. Pappas, “Efficient and accurate estimation of Lipschitz constants for deep neural networks,” Advances in Neural Information Processing Systems , vol. 32, 2019

2019
[13]

Zhou and J

K. Zhou and J. C. Doyle, Essentials of Robust Control, vol. 104. Prentice hall Upper Saddle River, NJ, 1998

1998
[14]

Linear matrix inequalities in control,

C. Scherer and S. Weiland, “Linear matrix inequalities in control,” Lecture Notes, Dutch Institute for Systems and Control, Delft, The Netherlands, vol. 3, no. 2, 2000

2000
[15]

LPV control and full block multipliers,

C. W. Scherer, “LPV control and full block multipliers,” Automatica, vol. 37, no. 3, pp. 361–375, 2001

2001
[16]

Extremal characterizations of the Schur complement and resulting inequalities,

C. K. Li and R. Mathias, “Extremal characterizations of the Schur complement and resulting inequalities,” SIAM Review , vol. 42, no. 2, pp. 233–246, 2000

2000
[17]

CVXPY: A python-embedded modeling lan- guage for convex optimization,

S. Diamond and S. Boyd, “CVXPY: A python-embedded modeling lan- guage for convex optimization,” Journal of Machine Learning Research , vol. 17, no. 83, pp. 1–5, 2016

2016
[18]

Mosek optimizer API for python,

M. ApS, “Mosek optimizer API for python,” V ersion, vol. 9, no. 17, pp. 6–4, 2022

2022
[19]

What is the state of neural network pruning?,

D. Blalock, J. J. Gonzalez Ortiz, J. Frankle, and J. Guttag, “What is the state of neural network pruning?,” Proceedings of Machine Learning and Systems , vol. 2, pp. 129–146, 2020

2020
[20]

A structure exploiting SDP solver for robust controller synthesis,

D. Gramlich, T. Holicki, C. W. Scherer, and C. Ebenbauer, “A structure exploiting SDP solver for robust controller synthesis,” IEEE Control Systems Letters , vol. 7, pp. 1831–1836, 2023

2023
[21]

On dual of LMIs for absolute stability analysis of nonlinear feedback systems with static O’Shea-Zames-Falb Multipliers,

H. Gyotoku, T. Yuno, Y . Ebihara, V . Magron, D. Peaucelle, and S. Tarbouriech, “On dual of LMIs for absolute stability analysis of nonlinear feedback systems with static O’Shea-Zames-Falb Multipliers,” arXiv preprint arXiv:2411.14339 , 2024

2024 arXiv

Pith tools

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

[1] [1]

Reachability analysis of neural feedback loops,

M. Everett, G. Habibi, C. Sun, and J. P. How, “Reachability analysis of neural feedback loops,” IEEE Access, vol. 9, pp. 163938–163953, 2021

work page 2021

[2] [2]

Reach-SDP: Reach- ability analysis of closed-loop systems with neural network controllers via semidefinite programming,

H. Hu, M. Fazlyab, M. Morari, and G. J. Pappas, “Reach-SDP: Reach- ability analysis of closed-loop systems with neural network controllers via semidefinite programming,” in IEEE Conf. on Decision and Control , (Jeju Island, Republic of Korea), pp. 5929–5934, Dec. 2020

work page 2020

[3] [3]

ReachNN: Reachability analysis of neural-network controlled systems,

C. Huang, J. Fan, W. Li, X. Chen, and Q. Zhu, “ReachNN: Reachability analysis of neural-network controlled systems,” ACM Transactions on Embedded Computing Systems , vol. 18, no. 5s, pp. 1–22, 2019

work page 2019

[4] [4]

Reachability analysis for neural feedback systems using regressive polynomial rule inference,

S. Dutta, X. Chen, and S. Sankaranarayanan, “Reachability analysis for neural feedback systems using regressive polynomial rule inference,” in Hybrid Systems: Computation and Control , pp. 157–168, 2019

work page 2019

[5] [5]

Robust guarantees for perception-based control,

S. Dean, N. Matni, B. Recht, and V . Ye, “Robust guarantees for perception-based control,” in Learning for Dynamics and Control , pp. 350–360, PMLR, 2020

work page 2020

[6] [6]

Stability Margins of Neural Network Controllers

N. Junnarkar, M. Arcak, and P. Seiler, “Stability margins of neural network controllers,” arXiv preprint arXiv:2409.09184 , 2024

work page Pith review arXiv 2024

[7] [7]

Synthesizing neural network controllers with closed-loop dissipativity guarantees,

N. Junnarkar, M. Arcak, and P. Seiler, “Synthesizing neural network controllers with closed-loop dissipativity guarantees,” arXiv preprint arXiv:2404.07373, 2024

arXiv 2024

[8] [8]

Deep networks for system identification: a survey,

G. Pillonetto, A. Aravkin, D. Gedon, L. Ljung, A. H. Ribeiro, and T. B. Sch¨on, “Deep networks for system identification: a survey,” Automatica, vol. 171, p. 111907, 2025

work page 2025