Pith. sign in

REVIEW 4 major objections 4 minor 69 references

Truncated Differentiation Through Primal-Dual Solvers for Inverse Potential Mean-Field Games

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Truncated gradients through 25 PDHG steps match full backpropagation in inverse mean-field games at a quarter of the memory.

desk verdict Solid analysis of truncated differentiation through PDHG for inverse MFGs, with an honest but unresolved gap between the exact-proximal theory and the finite-Newton experiments. read the letter →

arxiv 2608.00217 v2 pith:UFUVPFSO submitted 2026-07-31 math.OC cs.NAmath.NA

Siting Liu , Yat Tin Chow , Samy Wu Fung This is my paper · ORCID
classification math.OCcs.NAmath.NA MSC 49N8065K1049M4147H05
keywords inversemean-fieldgamesJacobian-freebackpropagationprimal-dualhybridgradientmonotoneoperatorresolventimplicitdifferentiationtruncatedinverse-costmaprecovery
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

The paper asks whether an inverse potential mean-field game — recovering the unknown mobility cost that moves a population, from observed density data — can be trained by differentiating through only the tail of the forward solver instead of its entire trajectory, which would avoid both the large tape storage of full unrolling and the extra linear solves of adjoint-based implicit differentiation. It claims yes: run the full preconditioned primal-dual hybrid gradient (PDHG) solve, but record and backpropagate only the final $r$ iterations from a detached warm start, and the resulting JFB-$r$ direction converges to the exact implicit gradient $\nabla_\theta F(\theta_0)$ as the tracked depth grows. The load-bearing identity is that the dual-extrapolated PDHG map with exact proximal steps is the metric resolvent $(I + M^{-1}A_\theta)^{-1}$ of the game's maximal monotone KKT operator, which replaces the contraction condition that earlier Jacobian-free backpropagation proofs required and yields, under a locally fixed active set at an exact equilibrium detach point, geometric bias decay with depth. In experiments, moderate depths suffice: JFB-25 is statistically indistinguishable from full automatic differentiation under noisy, partial-in-time observation while using roughly a quarter of the peak memory and half the runtime, and JFB-10 reduces memory by up to a factor of nine in a multi-instance task.

What carries the argument

The carrying object is the metric resolvent identity $T_\theta = (I + M^{-1}A_\theta)^{-1}$ (Theorem 3.2): the dual-extrapolated, Schur-preconditioned PDHG update coincides with the unit resolvent of the maximal monotone KKT operator $A_\theta(x,\phi) = (\partial f_\theta(x) + K^\top\phi,\; b - Kx)$ in the weighted metric $M$ built from the primal and dual step operators. This converts the analysis from the contraction assumption that standard JFB guarantees require — which PDHG lacks, being merely averaged and nonexpansive — into firm nonexpansiveness in the $M$-inner product, and makes the map semi-differentiable with derivative resolvent $R_0 = (I + M^{-1}DA)^{-1}$. The derivative resolvent has trivial kernel (strong convexity of the forward objective plus the injective temporal block $D_t$), which forces grouped directional decay even at flux kinks without strict complementarity; under the locally fixed active set of Assumption 4.1 the resolvent becomes the linear map $L = (M+H)^{-1}M$ with spectral radius below one, so the tracked sensitivities obey $u_{k+1} = Lu_k + P$ and JFB-$r$ is the truncated adjoint iteration of (4.8).

What would settle it

Measure the relative bias $\lVert d^{\mathrm{JFB}}_r - d^{\mathrm{ID}} \rVert / \lVert d^{\mathrm{ID}} \rVert$ against a cold-solved implicit reference on a problem with degenerate flux slots and persistent active-set changes: if increasing $r$ from 10 to hundreds does not reduce the bias toward zero, or if cosine alignment with the implicit gradient stays below about 0.5 at large $r$, Theorem 4.5's consistency is contradicted where it was meant to apply. A cheaper check: on a task where the warm start is far from equilibrium (large detach residual), increasing $r$ should still improve recovery if the practical claim holds; failure would show the near-equilibrium detachment is indispensable.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is a sharp trade-off: in inverse potential mean-field games the differentiation depth $r$ can be truncated severely without losing gradient quality. The paper proves (Theorem 4.5) that for the exact-proximal map, with exact equilibrium detachment and a locally fixed active set, $d^{\mathrm{JFB}}_r(\theta_0) \to \nabla_\theta F(\theta_0)$ as $r \to \infty$, with error $-P^\top (L^\top)^r (I - L^\top)^{-1} g$; the truncated direction is exactly $r$ metric-preconditioned iterations toward the implicit adjoint $H^{-\top}g$. The practical demonstration is that moderate depths suffice: JFB-25 matches full unrolling within three-seed variability in the hardest setting, using roughly a quarter of AD's peak memory and half its runtime, and JFB-10 runs 3.5x faster than AD in a multi-instance task with one-ninth the memory. The authors state the scope precisely: the consistency theorem is local and holds for the exact-proximal map, not directly for the implemented finite-Newton map with changing active sets.

Load-bearing premise

The proof of large-depth consistency assumes three things the implementation only approximates: the proximal solves inside PDHG are exact (the code uses 12 damped Newton iterations), the detachment point is exactly the equilibrium (the measured residual has median $2\times 10^{-7}$), and the set of binding flux constraints stays fixed during differentiation (in practice active sets shift in bursts of up to 48 slots during training). The authors state plainly that the local consistency theorem therefore does not apply directly to their experiments.

Editorial extensions

If this is right

  • Inverse-MFG training can recover the mobility map with $r = 10$–$25$ tracked steps, reaching errors statistically equal to full unrolling while using roughly a quarter to a ninth of the peak memory and half or less of the runtime.
  • At a $128^2$ spatial grid plain full unrolling exceeds a 48 GB GPU, while JFB-10 finishes at 17.5 GB and about 5.2x faster than checkpointed AD, at a modest recovery-error cost.
  • For any differentiable tracked map JFB-$r$ is the exact gradient of an $r$-step surrogate, so it is always a valid descent direction for that surrogate — but $r=1$ can badly misalign (negative cosine in 13 of 15 checkpoints), which explains why moderate depths matter.
  • JFB-$r$ needs no equilibrium-adjoint linear solve: it is entirely $r$ metric-preconditioned backward-Euler iterations toward the adjoint system $H^\top w = g$, with geometric bias decay in depth.

Reading between the lines

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

  • The metric-resolvent argument is not specific to mean-field games, since it only uses maximal monotonicity of the KKT operator and metric positivity: the same truncated-differentiation guarantee should carry over to other PDHG-solved saddle problems such as dynamic optimal transport, imaging inverse problems, and convex games, a transfer the paper mentions as future work only in passing.
  • If the grouped-decay bound on $L$ is taken literally, a principled adaptive schedule suggests itself: grow $r$ early in training when the warm-start detach residual is large, and shrink it once the model is near equilibrium — a testable extension that would automate the manual $r = 25$ choice.
  • The analysis predicts that JFB-$r$ becomes more reliable as training converges, because the detached warm start approaches a fixed point; one could verify this by tracking alignment of JFB-$r$ with the implicit reference across outer iterations, which the reported diagnostics already show improving in the descent phase.
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

4 major / 4 minor

Summary. The manuscript studies inverse potential mean-field games, in which an unknown spatial inverse-cost map is inferred from observed densities, and proposes a truncated differentiation scheme called JFB-r applied to a preconditioned primal-dual hybrid gradient (PDHG) forward solver. The main theoretical contributions are: (i) the exact-proximal dual-extrapolated PDHG map is identified as a metric resolvent of the maximal monotone KKT operator (Theorem 3.2); (ii) directional stability of the derivative resolvent is established without assuming flux strict complementarity (Theorem 3.6); and (iii) under a locally fixed active set with strict complementarity and exact equilibrium detachment, the JFB-r direction is shown to converge to the exact implicit gradient as the tracked depth r tends to infinity (Theorem 4.5). Numerically, the paper reports four inverse-MFG experiments in which moderate depths (r=10 or 25) reduce peak memory and runtime relative to full unrolling while achieving recovery errors close to those of full AD. The paper is explicit that the implemented solver uses an approximate proximal map, warm detached starts, and changing active sets, so the exact-proximal fixed-active-set theorem does not apply directly to the experiments.

Significance. If the results hold, the resolvent identification is a valuable theoretical contribution: it extends Jacobian-free backpropagation-style analysis beyond contractive fixed-point maps and gives a clean interpretation of JFB-r as a truncated adjoint iteration. The mathematical derivations in Sections 3-4 and Appendices C-F appear carefully executed, and the paper is unusually transparent about the gap between its assumptions and its implementation. The numerical experiments are useful and suggest that the method has practical merit. However, the central quantitative claim about moderate depths is not connected to the theory by any available bound: the theorem applies to a different map, at exact detachment, with a fixed active set, and the only finite-depth bound is conditional on an unverified contraction certificate. As it stands, the theoretical significance for the actually implemented method is therefore not yet established, while the empirical contribution stands as a separate, weaker claim.

major comments (4)
  1. [Theorem 4.5 and Section 5.1] Theorem 4.5 assumes the exact-proximal map T_theta, a locally fixed flux active set with strict complementarity (Assumption 4.1), and exact equilibrium detachment z-bar = z_0. Each of these hypotheses is violated in Algorithm 2.1 and in the experiments: the implemented map eT_theta replaces exact proximal solves by 12 damped Newton iterations (Section 2.4, Appendix A.4), Appendix G.7 reports active-set changes in bursts of up to 48 slots and 16 degenerate flux slots at the true-cost equilibrium, and the detach residual has median 2e-7 rather than zero. The paper concedes in Sections 4 and 5.1 that Theorem 4.5 does not apply directly. Because this theorem is the only theoretical result connecting tracked depth to the implicit gradient, the moderate-depth speedup claims in Sections 5.2 and 5.3 are not underwritten by the analysis. The manuscript should either provide a quantitative extension covering approximate prox and warm detachment, or clearly separate the asymptotic exact-regime theorem from the empirical approximate-regime results and soften the corresponding contribution claims.
  2. [Section 5.1, Table 2] The descent-phase alignment diagnostics show median relative gradient bias of 0.64 at r=10 and 0.60 at r=20 against the branch-matched implicit reference; the bias remains of order one at every tested depth. The only finite-depth error bound, Corollary E.2, is conditional on a numerically verified grouped contraction q_M<1 that is not supplied. Consequently, the paper provides no quantitative statement that the directions used at r=10, r=20, or r=25 are accurate surrogates for the true gradient. The recovery comparisons in Tables 4 and 5 are therefore empirical observations rather than consequences of the theory. The text should state this limitation explicitly and, ideally, add a direct attempt to estimate the relevant finite-depth operator norms or a sensitivity study that bridges the exact-proximal regime.
  3. [Section 4.3 and Appendix F] The adjoint-iteration interpretation in (4.8) and Appendix F is elegant, but the alignment is established in the M-geometry of the adjoint space, while the parameter-space gradient is obtained by a final compression through B B^T. Appendix F acknowledges that descent can be lost in this compression, and Table 2 confirms that the parameter-space bias is large in practice. Since the paper's practical claim is about parameter-space optimization, the M-alignment result does not by itself explain the observed recovery behavior. The authors should either prove a parameter-space bound under additional assumptions on B, or avoid phrasing that suggests the adjoint alignment transfers to the gradient used in training.
  4. [Corollary E.2 and Section 5.1] Corollary E.2 is the only quantitative finite-depth statement in the paper, but it is conditional and non-constructive: the existence of (r*, q_M) follows from Gelfand's formula, with no explicit estimate of r* or q_M. For a paper whose headline is that moderate tracked depths suffice, a non-constructive existence result does not provide the needed quantitative certificate at the tested depths. The authors should report an attempt to verify the grouped contraction numerically at representative checkpoints, or state plainly that no finite-depth theoretical guarantee is available for the depths used in the experiments.
minor comments (4)
  1. [Section 5.1, Table 1 discussion] The sentence 'JFB-10 gives up about 0.2 percentage points of accuracy for a 2.4x speedup' is ambiguous; it should read that JFB-10 is about 0.2 percentage points less accurate than AD, since the reported errors are 1.13% for JFB-10 versus 0.95% for AD.
  2. [Section 5.1, Table 2 and Appendix G.2.2] The alignment table evaluates r=5,10,20 retrospectively at states visited by JFB-10 trajectories, so the depth comparison is not made along each depth's own optimization trajectory. This caveat is present in the appendix but should be stated more prominently next to Table 2.
  3. [Appendix G.2.2, Table 7] Over all 63 checkpoints the depth ordering by median cosine is non-monotone, with r=5 having median cosine -0.53 and r=1 having +0.21. The main text should not overinterpret the descent-phase ordering without acknowledging this full-set variability.
  4. [Appendix G] The repository link in Appendix G says the code 'will be released'; for a submission claiming reproducibility, a permanent and accessible code artifact should be provided at review time.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivation chain is self-contained and the practical claims are explicitly separated from the non-applicable theoretical assumptions.

full rationale

The paper's central analysis is a self-contained mathematical derivation, not a reduction of the conclusion to its inputs. Theorem 3.2 proves that the dual-extrapolated PDHG map is the metric resolvent of the KKT operator by direct blockwise verification in Appendix C, and Theorem 3.6 derives directional decay from Minty's theorem, firm nonexpansiveness, and a kernel argument in Appendix D. Theorem 4.5 then follows from the explicit sensitivity recurrence (4.7) under the stated assumptions of Assumptions 2.2, 3.4, and 4.1, with the exact error formula d_JFB_r - ∇_θF = -P^T(L^T)^r(I-L^T)^{-1}g proved in Appendix E. Proposition 4.3, which identifies d_JFB_r with the gradient of the finite-trajectory surrogate, is a transparent chain-rule identity; the paper labels the object a surrogate and does not disguise it as a prediction of the true objective. The numerical claims are also not circular: the alignment diagnostics compare JFB-r against a separately cold-solved implicit reference (Section 5.1, Appendix G.2.2), and recovery is measured against a known ground-truth mobility map. The paper explicitly concedes that Theorem 4.5 does not apply directly to the experiments because the implemented map uses finite Newton prox steps, the detach residual is nonzero, and active sets change during training (Section 5.1, Appendix G.7); this is an honest validity gap or correctness risk, not a circularity. Self-citations to prior JFB and PDHG work are contextual and are not load-bearing here, since the present analysis supplies its own resolvent-based argument instead of invoking the contraction guarantees of prior JFB papers. No fitted parameter is renamed as a prediction, and no uniqueness theorem is imported from the authors' own prior work. Accordingly, no specific circular step can be exhibited, and the appropriate finding is no significant circularity.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

No fitted constants are introduced to force the main theorem; the free parameters are solver step sizes, regularization weights, and training hyperparameters. The load-bearing axioms are the explicit regularity assumptions (Assumptions 2.2, 3.4, 4.1) plus standard monotone operator theory. The exact-proximal idealization is an unverified modeling assumption separating theory from the implemented solver. No new physical or mathematical entities are invented.

free parameters (7)
  • step sizes tau and sigma = tau=sigma=0.99
    Solver step sizes chosen by hand in all experiments (Section 5). The theory only requires tau*sigma<1 (Assumption 2.2), so the central derivation does not depend on the specific value; the experiments do.
  • entropy weight gamma_I = 0.01
    Regularization weight chosen by hand (Section 5); theory requires gamma_I>0 (Assumption 3.4).
  • terminal weight gamma_T = 0.5 (Exp 1); 0.1 (Exp 2-4)
    Per-experiment hyperparameter from Table 6; theory requires gamma_T>0.
  • MLP width = 32 (Exp 1); 64 (Exp 2-4)
    Architecture choice; affects expressivity and recovery but not the theoretical claims.
  • outer iterations = 500 (Exp 1); 300 (Exp 2); 1000 (Exp 3); 2000 (Exp 4)
    Training budget chosen per experiment.
  • batch size = full batch (Exp 1-3); 2 (Exp 4)
    Experiment 4 uses mini-batches, making JFB-r a biased stochastic-gradient method.
  • learning rate schedule = cosine 1e-2 to 1e-3
    Optimizer schedule; standard choice.
assumptions (5)
  • domain assumption Assumption 2.2: the step-size operators satisfy ||S^{1/2} K T_x^{1/2}||_2^2 < 1, i.e., tau*sigma<1; this makes the metric M positive definite.
    Invoked to define the M-inner product and the resolvent (Theorem 3.2, Appendix C.1).
  • domain assumption Assumption 3.4: density bounded away from zero, uniform bounds on flux and cost map, positive entropy and terminal weights, positive terminal target.
    Used to prove local strong convexity of E_theta (Lemma 3.5), which underpins the kernel argument in Theorem 3.6 and the invertibility in Lemma 4.2.
  • domain assumption Assumption 4.1: locally fixed flux active set with strict complementarity, and density slots always retained.
    Gives local C^1 structure of the reduced KKT system, the implicit gradient formula (4.5), and the convergence theorem 4.5. The paper itself notes this assumption is restrictive and does not hold across the whole training trajectory.
  • standard math Classical results from monotone operator theory: Minty's theorem, sum theorem for maximal monotone operators, firm nonexpansiveness of resolvents, Krasnosel'skii-Mann convergence, Dini's theorem, implicit function theorem, polyhedral reduction lemma.
    Used throughout Sections 3-4 and appendices C-F; these are standard and not proved in the paper.
  • ad hoc to paper The exact-proximal map T_theta (2.15) is a faithful model of the implemented map eT_theta with 12 damped Newton steps.
    All theory analyzes T_theta, while all experiments use eT_theta; the paper acknowledges this gap and does not prove closeness of the two maps. This is an idealization rather than a derived equivalence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Truncated Differentiation Through Primal-Dual Solvers for Inverse Potential Mean-Field Games." pith.science (2026). https://pith.science/paper/UFUVPFSO

@misc{pith2026260800217,
  author       = {Pith},
  title        = {Pith review of: Truncated Differentiation Through Primal-Dual Solvers for Inverse Potential Mean-Field Games},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UFUVPFSO}},
  note         = {Machine review of arXiv:2608.00217}
}
abstract

We study inverse potential mean-field games (MFGs), in which an unknown spatial inverse-cost (mobility) map is inferred from observed population densities. We solve the forward MFG with a preconditioned primal-dual hybrid gradient (PDHG) method and develop Jacobian-free backpropagation (JFB-$r$), which records only the final $r$ iterations from a detached warm start while retaining the full forward solve. To analyze this truncated differentiation method, we show that the exact-proximal dual-extrapolated PDHG map is a metric resolvent of the maximal monotone KKT operator. This resolvent view shows that JFB-$r$ exactly differentiates a finite-trajectory surrogate and, under a locally fixed active set at an exact equilibrium detach point, converges to the implicit gradient as the tracked depth increases. Across several inverse-MFG settings, numerical experiments show that JFB-r at moderate tracked depths can achieve recovery accuracy comparable to full unrolling while reducing memory and runtime.

Figures

Figures reproduced from arXiv: 2608.00217 by the authors.

Figure 1
Figure 1. Experiment 3 noisy, partial-in-time setting for a representative seed. Top: ground truth and AD, JFB-10, and JFB-25 reconstructions. Bottom: pointwise errors on a shared scale. The depth dependence in [PITH_FULL_IMAGE:figures/full_fig_p017_1.png] view at source ↗
Figure 2
Figure 2. shows the final reconstructions [PITH_FULL_IMAGE:figures/full_fig_p018_2.png] view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

69 extracted references · 54 canonical work pages

  1. [1]

    Achdou, F

    Y. Achdou, F. Camilli, and I. Capuzzo-Dolcetta , Mean field games: Numerical methods for the planning problem , SIAM Journal on Control and Optimization, 50 (2012), pp. 77--109, https://doi.org/10.1137/100790069

    doi:10.1137/100790069 2012

  2. [2]

    Achdou and I

    Y. Achdou and I. Capuzzo-Dolcetta , Mean field games: numerical methods , SIAM Journal on Numerical Analysis, 48 (2010), pp. 1136--1162

    work page 2010

  3. [3]

    Achdou and M

    Y. Achdou and M. Lauri\` e re , Mean field games and applications: numerical aspects , in Mean Field Games, vol. 2281 of Lecture Notes in Mathematics, Springer, Cham, 2020, pp. 249--307

    work page 2020

  4. [4]

    Agrawal, W

    S. Agrawal, W. Lee, S. Wu Fung, and L. Nurbekyan , Random features for high-dimensional nonlocal mean-field games , Journal of Computational Physics, 459 (2022), p. 111136

    work page 2022

  5. [5]

    Andrade, G

    F. Andrade, G. Peyr \'e , and C. Poon , Sparsistency for inverse optimal transport , in International Conference on Learning Representations, vol. 2024, 2024, pp. 16575--16606

    work page 2024

  6. [6]

    S. Bai, J. Z. Kolter, and V. Koltun , Deep equilibrium models , in Proceedings of the 33rd International Conference on Neural Information Processing Systems, 2019, pp. 690--701

    work page 2019

  7. [7]

    H. H. Bauschke and P. L. Combettes , Convex Analysis and Monotone Operator Theory in Hilbert Spaces , CMS Books in Mathematics, Springer, Cham, 2nd ed., 2017

    work page 2017

  8. [8]

    Benamou and Y

    J.-D. Benamou and Y. Brenier , A computational fluid mechanics solution to the M onge- K antorovich mass transfer problem , Numerische Mathematik, 84 (2000), pp. 375--393

    work page 2000

Show all 69 references
  1. [9]

    Benamou and G

    J.-D. Benamou and G. Carlier , Augmented L agrangian methods for transport optimization, mean field games, and degenerate elliptic equations , Journal of Optimization Theory and Applications, 167 (2015), pp. 1--26

    2015

  2. [10]

    Benamou, G

    J.-D. Benamou, G. Carlier, and F. Santambrogio , Variational mean field games , in Active Particles, Volume 1: Advances in Theory, Models, and Applications, Birkh\" a user, Cham, 2017, pp. 141--171

    2017

  3. [11]

    L. M. Brice \ n o-Arias, D. Kalise, Z. Kobeissi, M. Lauri\` e re, A. Mateos Gonz\' a lez, and F. J. Silva , On the implementation of a primal-dual algorithm for second order time-dependent mean field games with local couplings , ESAIM: Proceedings and Surveys, 65 (2019), pp. 330--348

    2019

  4. [12]

    L. M. Brice \ n o-Arias, D. Kalise, and F. J. Silva , Proximal methods for stationary mean field games with local couplings , SIAM Journal on Control and Optimization, 56 (2018), pp. 801--836, https://doi.org/10.1137/16M1095615

    2018 doi

  5. [13]

    Chambolle and T

    A. Chambolle and T. Pock , A first-order primal-dual algorithm for convex problems with applications to imaging , Journal of Mathematical Imaging and Vision, 40 (2011), pp. 120--145, https://doi.org/10.1007/s10851-010-0251-1

    2011 doi

  6. [14]

    Chambolle and T

    A. Chambolle and T. Pock , On the ergodic convergence rates of a first-order primal--dual algorithm , Mathematical Programming, 159 (2016), pp. 253--287, https://doi.org/10.1007/s10107-015-0957-3

    2016 doi

  7. [15]

    Y. T. Chow, S. Wu Fung, S. Liu, L. Nurbekyan, and S. Osher , A numerical algorithm for inverse problem from partial boundary measurement arising from mean field game problem , Inverse Problems, 39 (2023), p. 014001

    2023

  8. [16]

    L. Condat , A primal--dual splitting method for convex optimization involving L ipschitzian, proximable and linear composite terms , Journal of Optimization Theory and Applications, 158 (2013), pp. 460--479, https://doi.org/10.1007/s10957-012-0245-9

    2013 doi

  9. [17]

    Daosud, P

    W. Daosud, P. Thitiyasook, A. Arpornwichanop, P. Kittisupakorn, and M. A. Hussain , Neural network inverse model-based controller for the control of a steel pickling process , Computers & Chemical Engineering, 29 (2005), pp. 2110--2119

    2005

  10. [18]

    L. Ding, W. Li, S. Osher, and W. Yin , A mean field game inverse problem , Journal of Scientific Computing, 92 (2022), p. 7

    2022

  11. [19]

    A. L. Dontchev and R. T. Rockafellar , Implicit Functions and Solution Mappings: A View from Variational Analysis , Springer, New York, 2nd ed., 2014

    2014

  12. [20]

    El Ghaoui, F

    L. El Ghaoui, F. Gu, B. Travacca, A. Askari, and A. Tsai , Implicit deep learning , SIAM Journal on Mathematics of Data Science, 3 (2021), pp. 930--958

    2021

  13. [21]

    Gelphman, D

    E. Gelphman, D. Verma, N. T. Yang, S. Osher, and S. Wu Fung , End-to-end training of high-dimensional optimal control with implicit H amiltonians via J acobian-free backpropagation , arXiv preprint arXiv:2510.00359, (2025)

    2025

  14. [22]

    Gelphman, D

    E. Gelphman, D. Verma, N. T. Yang, S. Osher, and S. Wu Fung , On the convergence of J acobian-free backpropagation for optimal control problems with implicit H amiltonians , arXiv preprint arXiv:2602.00921, (2026)

    2026 arXiv

  15. [23]

    Geng, X.-Y

    Z. Geng, X.-Y. Zhang, S. Bai, Y. Wang, and Z. Lin , On training implicit models , Advances in neural information processing systems, 34 (2021), pp. 24247--24260

    2021

  16. [24]

    Gonz \'a lez-Sanz, M

    A. Gonz \'a lez-Sanz, M. Groppe, and A. Munk , Nonlinear inverse optimal transport: Identifiability of the transport cost from its marginals and optimal values , SIAM Journal on Mathematical Analysis, 56 (2024), pp. 7808--7829

    2024

  17. [25]

    J. Guo, C. Mou, X. Yang, and C. Zhou , Decoding mean field games from population and environment observations by G aussian processes , Journal of Computational Physics, 508 (2024), p. 112978, https://doi.org/10.1016/j.jcp.2024.112978

    2024

  18. [26]

    He and X

    B. He and X. Yuan , Convergence analysis of primal-dual algorithms for a saddle-point problem: From contraction perspective , SIAM Journal on Imaging Sciences, 5 (2012), pp. 119--149

    2012

  19. [27]

    Heaton and S

    H. Heaton and S. Wu Fung , Explainable AI via learning to optimize , Scientific Reports, 13 (2023), p. 10103

    2023

  20. [28]

    Heaton, S

    H. Heaton, S. Wu Fung, A. Gibali, and W. Yin , Feasibility-based fixed point networks , Fixed Point Theory and Algorithms for Sciences and Engineering, 2021 (2021), p. 21

    2021

  21. [29]

    Huang, J

    H. Huang, J. Yu, T. Chen, and R. Lai , Joint inference of trajectory and obstacle in mean-field games via bilevel optimization , arXiv preprint arXiv:2507.19344, (2025)

    2025 arXiv

  22. [30]

    Huang, R

    M. Huang, R. P. Malham \'e , and P. E. Caines , Large population stochastic dynamic games: closed-loop McKean-Vlasov systems and the N ash certainty equivalence principle , Communications in Information and Systems, 6 (2006), pp. 221--252

    2006

  23. [31]

    Imanuvilov, H

    O. Imanuvilov, H. Liu, and M. Yamamoto , Lipschitz stability for determination of states and inverse source problem for the mean field game equations , Inverse Problems and Imaging, 18 (2024), pp. 824--859, https://doi.org/10.3934/ipi.2023057

    2024 doi

  24. [32]

    Jacobs, F

    M. Jacobs, F. L \'e ger, W. Li, and S. Osher , Solving large-scale optimization problems with a convergence rate independent of grid size , SIAM Journal on Numerical Analysis, 57 (2019), pp. 1100--1123

    2019

  25. [33]

    Knutson, A

    B. Knutson, A. C. Rabeendran, M. Ivanitskiy, J. Pettyjohn, C. D. Behn, S. Wu Fung, and D. McKenzie , On logical extrapolation for mazes with recurrent and implicit networks , in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 40, 2026, pp. 22635--22643

    2026

  26. [34]

    Lasry and P.-L

    J.-M. Lasry and P.-L. Lions , Mean field games , Japanese journal of mathematics, 2 (2007), pp. 229--260

    2007

  27. [35]

    Levine and V

    S. Levine and V. Koltun , Continuous inverse optimal control with locally optimal examples , in International Conference on Machine Learning (ICML), 2012

    2012

  28. [36]

    X. Li, K. Kan, D. Verma, K. Kumar, S. Osher, and S. Wu Fung , End-to-end learning of safe optimal feedback control in high dimensions with control barrier function layers , arXiv preprint arXiv:2607.20674, (2026)

    2026 arXiv

  29. [37]

    A. T. Lin, S. Wu Fung, W. Li, L. Nurbekyan, and S. J. Osher , Alternating the population and control neural networks to solve high-dimensional stochastic mean-field games , Proceedings of the National Academy of Sciences, 118 (2021), p. e2024713118

    2021

  30. [38]

    H. Liu, C. Mou, and S. Zhang , Inverse problems for mean field games , Inverse Problems, 39 (2023), p. 085003

    2023

  31. [39]

    Liu and S

    H. Liu and S. Zhang , On an inverse boundary problem for mean field games , arXiv preprint arXiv:2212.09110, (2022)

    2022 arXiv

  32. [40]

    S. Liu, M. Jacobs, W. Li, L. Nurbekyan, and S. J. Osher , Computational methods for first-order nonlocal mean field games with applications , SIAM Journal on Numerical Analysis, 59 (2021), pp. 2639--2668

    2021

  33. [41]

    Liu and L

    S. Liu and L. Nurbekyan , Splitting methods for a class of non-potential mean field games , Journal of Dynamics and Games, 8 (2021), pp. 467--486

    2021

  34. [42]

    McKenzie, H

    D. McKenzie, H. Heaton, Q. Li, S. Wu Fung, S. Osher, and W. Yin , Three-operator splitting for learning to predict equilibria in convex games , SIAM Journal on Mathematics of Data Science, 6 (2024), pp. 627--648

    2024

  35. [43]

    McKenzie, H

    D. McKenzie, H. Heaton, and S. Wu Fung , Differentiating through integer linear programs with quadratic regularization and D avis- Y in splitting , Transactions on Machine Learning Research, (2024)

    2024

  36. [44]

    T. Meng, S. Liu, S. Wu Fung, and S. Osher , Recent advances in numerical solutions for Hamilton--Jacobi PDEs , Communications on Applied Mathematics and Computation, (2026), https://doi.org/10.1007/s42967-026-00570-1. arXiv:2502.20833

    2026

  37. [45]

    G. J. Minty , Monotone (nonlinear) operators in H ilbert space , Duke Mathematical Journal, 29 (1962), pp. 341--346

    1962

  38. [46]

    Onken, L

    D. Onken, L. Nurbekyan, X. Li, S. Wu Fung, S. Osher, and L. Ruthotto , A neural network approach applied to multi-agent optimal control , in 2021 European Control Conference (ECC), IEEE, 2021, pp. 1036--1041

    2021

  39. [47]

    Onken, L

    D. Onken, L. Nurbekyan, X. Li, S. Wu Fung, S. Osher, and L. Ruthotto , A neural network approach for high-dimensional optimal control applied to multiagent path finding , IEEE Transactions on Control Systems Technology, 31 (2023), pp. 235--251

    2023

  40. [48]

    Onken, S

    D. Onken, S. Wu Fung, X. Li, and L. Ruthotto , OT-Flow : Fast and accurate continuous normalizing flows via optimal transport , in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 35, 2021, pp. 9223--9232

    2021

  41. [49]

    Papadakis, G

    N. Papadakis, G. Peyr \'e , and E. Oudet , Optimal transport with proximal splitting , SIAM Journal on Imaging Sciences, 7 (2014), pp. 212--238

    2014

  42. [50]

    Y. Park, E. Gelphman, S. Osher, and S. Wu Fung , Implicit neural optimal transport via fixed-point optimization , arXiv preprint arXiv:2605.10792, (2026)

    2026 arXiv

  43. [51]

    K. Ren, N. Soedjak, and S. Tong , A policy iteration method for inverse mean field games , Research in the Mathematical Sciences, 13 (2026), p. 34

    2026

  44. [52]

    K. Ren, N. Soedjak, K. Wang, and H. Zhai , Reconstructing a state-independent cost function in a mean-field game model , Inverse Problems, 40 (2024), p. 105010, https://doi.org/10.1088/1361-6420/ad7497

    2024 doi

  45. [53]

    R. T. Rockafellar , Monotone operators and the proximal point algorithm , SIAM Journal on Control and Optimization, 14 (1976), pp. 877--898

    1976

  46. [54]

    R. T. Rockafellar , Proto-differentiability of set-valued mappings and its applications in optimization , Annales de l'Institut Henri Poincar\'e, Analyse Non Lin\'eaire, 6 (1989), pp. 449--482

    1989

  47. [55]

    R. T. Rockafellar and R. J.-B. Wets , Variational Analysis , vol. 317 of Grundlehren der mathematischen Wissenschaften, Springer, Berlin, 1998

    1998

  48. [56]

    Ruthotto, S

    L. Ruthotto, S. J. Osher, W. Li, L. Nurbekyan, and S. Wu Fung , A machine learning framework for solving high-dimensional mean field game and mean field control problems , Proceedings of the National Academy of Sciences, 117 (2020), pp. 9183--9193

    2020

  49. [57]

    E. K. Ryu and W. Yin , Large-scale convex optimization: algorithms & analyses via monotone operators , Cambridge University Press, 2022

    2022

  50. [58]

    Shaban, C.-A

    A. Shaban, C.-A. Cheng, N. Hatch, and B. Boots , Truncated back-propagation for bilevel optimization , in The 22nd international conference on artificial intelligence and statistics, PMLR, 2019, pp. 1723--1732

    2019

  51. [59]

    A. M. Stuart and M.-T. Wolfram , Inverse optimal transport , SIAM Journal on Applied Mathematics, 80 (2020), pp. 599--619

    2020

  52. [60]

    Vidal, S

    A. Vidal, S. Wu Fung, S. Osher, L. Tenorio, and L. Nurbekyan , Kernel expansions for high-dimensional mean-field control with non-local interactions , in 2025 American Control Conference (ACC), IEEE, 2025, pp. 4164--4171

    2025

  53. [61]

    Vidal, S

    A. Vidal, S. Wu Fung, L. Tenorio, S. Osher, and L. Nurbekyan , Taming hyperparameter tuning in continuous normalizing flows using the JKO scheme , Scientific Reports, 13 (2023), p. 4501

    2023

  54. [62]

    B. C. V \ u , A splitting algorithm for dual monotone inclusions involving cocoercive operators , Advances in Computational Mathematics, 38 (2013), pp. 667--681, https://doi.org/10.1007/s10444-011-9254-8

    2013 doi

  55. [63]

    X. Wang, S. Wu Fung, and L. Nurbekyan , A primal-dual price-optimization method for computing equilibrium prices in mean-field games models , Dynamic Games and Applications, (2025), https://doi.org/10.1007/s13235-025-00688-6. Online first

    2025 doi

  56. [64]

    Wu Fung and B

    S. Wu Fung and B. Berkels , A generalization bound for a family of implicit networks , Neurocomputing, 678 (2026), p. 133136

    2026

  57. [65]

    Wu Fung, H

    S. Wu Fung, H. Heaton, Q. Li, D. McKenzie, S. Osher, and W. Yin , JFB : J acobian-free backpropagation for implicit networks , in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 36, 2022, pp. 6648--6656

    2022

  58. [66]

    W. Yin, D. McKenzie, and S. Wu Fung , Learning to optimize: Where deep learning meets optimization and inverse problems , SIAM News, (2022)

    2022

  59. [67]

    Yu, J.-G

    J. Yu, J.-G. Liu, and H. Zhao , Equilibrium correction iteration for a class of mean-field game inverse problems , Inverse Problems, 41 (2025), p. 125009, https://doi.org/10.1088/1361-6420/ae2b09

    2025 doi

  60. [68]

    J. Yu, Q. Xiao, T. Chen, and R. Lai , A bilevel optimization method for inverse mean-field games , Inverse Problems, 40 (2024), p. 105016

    2024

  61. [69]

    Zhang, X

    J. Zhang, X. Yang, C. Mou, and C. Zhou , Learning surrogate potential mean field games via G aussian processes: A data-driven approach to ill-posed inverse problems , Journal of Computational Physics, 543 (2025), p. 114412, https://doi.org/10.1016/j.jcp.2025.114412

    2025

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.