Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Characterizing Nash Equilibria in Zero-Sum Games: A Physics-Inspired, Parallelizable Approach with a Linear Number of Gradient Queries

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Modulo degeneracy, a linear number of alternating gradient descent iterations — 2k, 2k+1, or k, depending on what is observed — yields linear equations whose solution is the Nash equilibrium of a zero-sum or coordination game.

desk verdict The finite-iteration characterization is real but only works for games with a unique Nash equilibrium; the unproven 'modulo degeneracy' caveat is actually the main limitation and needs to be stated up front. read the letter →

arxiv 2507.11366 v2 pith:ZQB6A3M6 submitted 2025-07-15 cs.GT cs.LG

Taemin Kim , James P. Bailey This is my paper · ORCID
classification cs.GTcs.LG MSC 91A0591A1091A2690C47
keywords zero-sumgamesNashequilibriumalternatinggradientdescentHamiltoniandynamicsinvariantenergyfunctiononlineoptimizationparallelizablealgorithmsarbitrarylearningrates
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 attempts to establish that alternating gradient descent in zero-sum and coordination games does more than converge slowly: every update step yields one linear equation in the unknown Nash equilibrium, so after a number of steps linear in the dimension ($2k$, $2k+1$, or $k$, depending on what is observed) the accumulated equations form a linear system whose solution is the equilibrium. The statement carries an explicit caveat, "modulo degeneracy": the system must be non-singular for the equilibrium to be its unique solution, and the paper assumes rather than proves this rank condition. If the claim is right, it replaces asymptotic time-average or last-iterate guarantees with a direct solve after a finite number of gradient queries, works with arbitrary learning rates, and parallelizes across independent runs. The authors report that the parallelized version reaches about $10^{-10}$ relative error in under a second in dimensions up to 20 (and stays accurate at dimension 500), where time-average alternating gradient descent reaches only about $10^{-6}$ after five minutes in dimension 7.

What carries the argument

The load-bearing object is the discrete-time invariant energy function, called the perturbed energy — for zero-sum games $h^-_t = \|x_t - x^*\|^2/\eta_1 + \|y_t - y^*\|^2/\eta_2 + \langle x_t, Ay_t - b_1\rangle + \langle y_t, b_2\rangle$, with an analogous $h^+_t$ for coordination games in which the agent-2 squared-distance and inner-product terms enter with opposite signs. Theorems 1 and 2 prove $h^-_t = h^-_0$ and $h^+_t = h^+_0$ for every $t$ under alternating gradient descent, using only the unbounded-setting first-order conditions $Ay^* = b_1$ and $Bx^* = b_2$. The mechanism that carries the argument is the difference $h^-_t - h^-_{t+1} = 0$: expanding it cancels the quadratic terms and leaves one linear equation in the unknown equilibrium per iteration, each row of the eventual linear system. The physical motivation — alternating gradient descent as a symplectic Störmer-Verlet discretization of Hamiltonian dynamics whose energy is squared distance to the NE — explains why such an invariant exists, but the linearity of the expanded differences is what does the mathematical work.

What would settle it

For the one-dimensional zero-sum game with payoff $a$, costs $b_1, b_2$, and learning rates $\eta_1, \eta_2$, derive the two Section 4.2 equations from three consecutive alternated updates, view them as a $2\times 2$ linear system in $(x^*, y^*)$, and compute the determinant symbolically as a function of $a$, $b_1$, $b_2$, $\eta_1$, $\eta_2$, and the starting point. If any non-degenerate game (unique equilibrium, $a \neq 0$) has a starting point where the determinant vanishes, the finite-iteration characterization fails outside the degeneracies the paper acknowledges; if the determinant is nonzero on a dense set of parameters, that calculation shows exactly how large the 'modulo degeneracy' caveat is.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the time-invariant "perturbed energy" conserved exactly by alternating gradient descent in a zero-sum game (with a sibling energy in coordination games) is not only a convergence tool: expanding its constancy across two consecutive iterations gives an equation that is linear in the unknown equilibrium $(x^*, y^*)$ and the linear cost terms, with every coefficient observable from the agents' strategies and gradients, without knowing the payoff matrix. Three models trade information for iteration count: observing both agents' strategies recovers $(x^*, y^*)$ from $2k$ equations; observing only the opponent's strategy norm recovers $(x^*, b_1)$ from $2k+1$ equations; observing only the opponent's gradient norm recovers $x^*$ from $k$ equations. The authors are careful to state that they do not claim the dynamics themselves solve for the equilibrium in finite time; they claim that the dynamics supply the linear system $A'(x^*, y^*) = b'$, and that solving that system, when it is non-singular, characterizes the NE exactly. They further claim this is the first finite-iteration characterization in online optimization in games, and the first such method that is parallelizable and supports arbitrary learning rates.

Load-bearing premise

For the method to work, the linear equations collected across alternating-gradient iterations must be independent enough that the unknown equilibrium is their unique solution; the paper assumes this full-rank condition, flagging it as "modulo degeneracy" in Sections 4.1 and 5, and does not prove that random parallel initialization makes it hold.

Editorial extensions

If this is right

  • An equilibrium can be computed from $2k$ (or, in the reduced models, $2k+1$ or $k$) gradient queries made by an online algorithm, with no need to wait for time-averaging or last-iterate convergence.
  • The procedure works with arbitrary learning rates, so the stability condition $\sqrt{\eta_1 \eta_2} < 2/\|A\|$ required for convergence of the underlying dynamics is not a restriction on the characterization.
  • The computation parallelizes: each equation may come from an independent initial condition, and the experiments indicate the assembled system is then well-conditioned across dimensions.
  • The reported comparison shows roughly $10^{-10}$ relative error in under a second for the new method versus roughly $10^{-6}$ after five minutes for time-average alternating gradient descent in dimension 7.
  • Via a variable substitution that preserves zero-sum and coordination structure, the characterization transfers to games on probability simplices when the equilibrium is fully mixed, and to polymatrix zero-sum games through standard reductions.

Reading between the lines

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

  • My inference: the consecutive-energy-difference trick should transfer to any discrete-time dynamic that conserves or nearly conserves a quadratic invariant in the unknown equilibrium, such as extragradient or optimistic variants, converting each step into a row of a linear system; the paper does not explore this generalization.
  • My inference: the practical reach of the method depends on how often the 'modulo degeneracy' caveat binds, and the experiments do not prove it is rare — a concentration or genericity result showing random independent initializations are full-rank with high probability would convert the observed conditioning into a guarantee.
  • My inference: since only one or two updates per trajectory are needed, the method could serve as a probing device that decouples informativeness from convergence, deliberately using large learning rates to collect equations from trajectories that would diverge if run longer — a mode of use the convergence literature would not predict.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper studies alternating gradient descent (AGD) in two-agent unbounded zero-sum and coordination games with linear costs. It shows that the discrete-time AGD dynamics admit an invariant energy (a perturbed distance to the Nash equilibrium). Taking differences of this energy at consecutive iterations yields equations that are linear in the unknown equilibrium strategy and/or cost vectors, provided the game is zero-sum or coordination. The paper proposes three models requiring observation of different information (both strategies; one strategy plus the opponent's norm; one strategy plus the opponent's gradient norm), and claims that, modulo a non-degeneracy condition, 2k, 2k+1, or k iterations suffice to generate a linear system whose solution is the Nash equilibrium. Numerical experiments show that solving this system directly is numerically unstable, but a parallelized version using fresh random initial conditions for each equation yields accurate solutions very quickly for random invertible payoff matrices.

Significance. If the rank question were resolved, the core idea would be valuable: it gives a finite-time, equation-based alternative to asymptotic time-average or last-iterate convergence for finding Nash equilibria, and the parallelization and learning-rate independence are appealing. The algebraic proofs of the invariant energy and of the linear forms appear correct and are presented self-containedly, with the zero-sum invariant re-proved in the appendix. The experiments provide reproducible evidence of high accuracy for random invertible matrices. However, the central theoretical claim currently rests on an uncharacterized 'modulo degeneracy' condition, and the method implicitly excludes games with non-unique equilibria. These gaps substantially temper the novelty as stated.

major comments (3)
  1. [Theorems 5–12; Sections 4.1, 5, 6] The theorems prove only that every Nash equilibrium satisfies the displayed linear equations; they never prove that the assembled 2k×2k (or k×k) system has full rank. The phrase 'modulo degeneracy' (Section 4.1, Section 5) is never made precise: no condition on A, the learning rates, or the initial conditions is identified that guarantees non-singularity. Table 2 shows that for a single trajectory the determinant is already about 10^-7 at dimension 3, so degeneracy is not a rare pathology, and Section 6's parallelization is supported only by experiments. Without a genericity or probability bound, the central claim of a finite-iteration characterization is not established for any specific game.
  2. [Section 4.2, Theorem 5; NE conditions] For any two Nash equilibria (x*, y*) and (x*+Δx, y*+Δy), the difference of the two equalities in Theorem 5 shows the row vector at time t is orthogonal to (Δx, Δy). Consequently, whenever the NE set has positive dimension (which occurs exactly when A is singular and the NE conditions are consistent), the 2k×2k system matrix necessarily has a nontrivial nullspace, so the method cannot 'characterize the set of NE' for such games. This limitation is not stated in the abstract or in the model description; the paper should either restrict all claims to unique-NE games or explain how degeneracy is resolved.
  3. [Section 6] The claim that parallelization 'fundamentally changes the structure of linear systems' and yields well-conditioned matrices is not accompanied by any theoretical justification. Since each row is an affine function of the initial condition, one can compute the linear part of the map from initial conditions to rows; its kernel is null(A^T) × null(A), which would establish generic full rank for invertible A and also confirm the obstruction for singular A. This argument is absent, and without it the parallelized method is an empirical heuristic rather than a theorem.
minor comments (6)
  1. [Abstract; Section 1.1] The claims 'a first in online optimization' and 'both firsts in algorithmic game theory' are difficult to verify and are not supported by a thorough literature comparison; they should be qualified or removed.
  2. [Section 5.3.1, Table 5] Table 5 reports relative errors near 1.0 for dimensions 100–500 even with the least-squares method, which is at odds with the narrative that least squares 'stabilizes' the system; this discrepancy is not discussed in the text.
  3. [Section 4.1] The term 'modulo degeneracy' is used informally; the paper should give a formal definition, for example 'the linear system constructed from the displayed equations has full row rank,' and state what happens to the solution set when this fails.
  4. [Sections 4.3 and 4.4] The abstract and introduction should be consistent with the fact that the models of Sections 4.3 and 4.4 characterize only x* (and b1) and require a symmetric model to obtain y*; as written, some passages suggest the full equilibrium is recovered.
  5. [Theorems 8 and 10] The proof of Theorem 10 is only sketched in the main text, with no appendix; the paper should either provide the full proof or state that it follows from Theorem 9 with the same cancellation as in Theorem 8.
  6. [Throughout] There are several typographical issues, e.g., 'V on Neumann' in the Introduction and inconsistent subscript formatting in the appendices; a careful proofread is recommended.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the invariant-based linear equations are re-derived in the paper and solved for the NE; the only caveat is an unproven non-singularity condition, which is a correctness gap, not circularity.

full rationale

The paper's central derivation is self-contained. The invariant functions h^t_- and h^t_+ are defined using the unknown NE (x*, y*), but the paper proves their time-invariance from the Alternating GD update rule and the NE gradient conditions alone, and then forms equations h^t - h^{t+1} = 0 that are linear in the unknowns. Solving those equations does not assume the NE as an input; it derives constraints that every NE must satisfy. The paper explicitly states the results for zero-sum and coordination games were previously shown in the authors' earlier work, but it also says 'To keep the paper self-contained, we include the proofs here' and provides the proofs in the main text and appendices, so the prior results are not merely imported by citation. The Hamiltonian-dynamics motivation from [6] is heuristic and not load-bearing for the theorems. No parameter is fitted to data and no 'prediction' is statistically forced. The main unproven premise is that the constructed linear system is non-singular, which the paper flags as 'modulo degeneracy' in Sections 4.1 and 5; this is a gap in the theoretical guarantee rather than a circular reduction, because the equations would still be valid constraints even if the system were singular. The experimental comparisons use independently generated games and standard solvers, so there is no circular benchmarking. Overall, the derivation does not reduce to its inputs by construction.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

No free parameters are fitted to data; the learning rates are arbitrary and the method is parameter-free. The Tikhonov regularization parameter lambda in the experiments is a numerical stabilizer, not part of the central claim. Two axioms are listed: the non-degeneracy assumption that is essential but unproven, and the standard gradient characterization of NE in unbounded games.

assumptions (2)
  • ad hoc to paper The linear system formed from energy differences is non-singular (full rank), so the NE is the unique solution.
    Stated as 'modulo degeneracy' in Sections 4.1 and 5. The paper never proves this and instead relies on random parallel initializations to make the system well-conditioned in experiments.
  • domain assumption In the unbounded strategy space, a Nash equilibrium is equivalent to both players' gradients being zero.
    Used throughout (Section 2) to derive the NE conditions and the energy invariant. This is standard for unconstrained convex-concave games.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Characterizing Nash Equilibria in Zero-Sum Games: A Physics-Inspired, Parallelizable Approach with a Linear Number of Gradient Queries." pith.science (2026). https://pith.science/paper/ZQB6A3M6

@misc{pith2026250711366,
  author       = {Pith},
  title        = {Pith review of: Characterizing Nash Equilibria in Zero-Sum Games: A Physics-Inspired, Parallelizable Approach with a Linear Number of Gradient Queries},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZQB6A3M6}},
  note         = {Machine review of arXiv:2507.11366}
}
read the original abstract

We study online optimization methods for zero-sum games, a fundamental problem in adversarial learning in machine learning, economics, and many other domains. Traditional methods approximate Nash equilibria (NE) using either regret-based methods (time-average convergence) or contraction-map-based methods (last-iterate convergence). We propose a new method based on Hamiltonian dynamics in physics and prove that it can characterize the set of NE in a finite (linear) number of iterations of alternating gradient descent in the unbounded setting, modulo degeneracy, a first in online optimization. Unlike standard methods for computing NE, our proposed approach can be parallelized and works with arbitrary learning rates, both firsts in algorithmic game theory. Experimentally, we support our results by showing our approach drastically outperforms standard methods.

Figures

Figures reproduced from arXiv: 2507.11366 by the authors.

Figure 1
Figure 1. Let (q(t), p(t)) represent the position and momentum of a unit mass on a frictionless spring with spring constant k = 1, and let (x1(t), x2(t)) represent agent strategies in the zero-sum game maxx1∈R minx2∈R⟨x1, x2⟩ where agents update strategies with Continuous-time GD with learning rate η = 1. [6] provides the framework to show (q(t), p(t)) = (x1(t), x2(t)) ∀t if (q(0), p(0)) = (x1(0), x2(0)), i.e., Continuous-tim… view at source ↗
Figure 2
Figure 2. The Continuous-time GD dynamics cycle around the set of NE on a closed orbit. We show that Alternating GD closely [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Characterizing the set of NE via Alternating GD in dimension 1. In contrast, the time-average of the strategies are still [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Spiral trajectory of time-average strategies towards the NE via Alternating GD [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Implicit Midpoint Gradient Descent: Fast and Learning rate free convergence for Zero-Sum Games

    cs.GT 2026-07 accept novelty 6.0 of 10

    Implicit midpoint gradient descent preserves distance to every Nash equilibrium and achieves O(1/(ηT)+1/T) time-average convergence in unconstrained bilinear zero-sum games for any η>0.

Reference graph

Works this paper leans on

33 extracted references · 31 canonical work pages · cited by 1 Pith paper

  1. [1]

    Last-iterate convergence rates for min-max optimization: Convergence of hamiltonian gradient descent and consensus optimization

    Jacob Abernethy, Kevin A Lai, and Andre Wibisono. Last-iterate convergence rates for min-max optimization: Convergence of hamiltonian gradient descent and consensus optimization. In Algorithmic Learning Theory , pages 3–47. PMLR, 2021

    work page 2021

  2. [2]

    The equivalence of linear programs and zero-sum games

    Ilan Adler. The equivalence of linear programs and zero-sum games. International Journal of Game Theory , 42:165–177, 2013

    work page 2013

  3. [3]

    $O\left(1/T\right)$ Time-Average Convergence in a Generalization of Multiagent Zero-Sum Games

    James P Bailey. O(1/t) time-average convergence in a generalization of multiagent zero-sum games. arXiv preprint arXiv:2110.02482, 2021

    work page Pith review arXiv 2021

  4. [4]

    Bailey, Gauthier Gidel, and Georgios Piliouras

    James P. Bailey, Gauthier Gidel, and Georgios Piliouras. Finite regret and cycles with fixed step-size via alter- nating gradient descent-ascent. In Jacob Abernethy and Shivani Agarwal, editors, Proceedings of Thirty Third Conference on Learning Theory , volume 125 of Proceedings of Machine Learning Research , pages 391–407. PMLR, 09–12 Jul 2020

    work page 2020

  5. [5]

    Bailey and Georgios Piliouras

    James P. Bailey and Georgios Piliouras. Multiplicative weights update in zero-sum games. In Proceedings of the 2018 ACM Conference on Economics and Computation, EC ’18, page 321–338, New York, NY , USA, 2018. Association for Computing Machinery

    work page 2018

  6. [6]

    Bailey and Georgios Piliouras

    James P. Bailey and Georgios Piliouras. Multi-agent learning in network zero-sum games is a hamiltonian system. In Edith Elkind, Manuela Veloso, Noa Agmon, and Matthew E. Taylor, editors,Proceedings of the 18th International Conference on Autonomous Agents and MultiAgent Systems, AAMAS ’19, Montreal, QC, Canada, May 13-17, 2019 , pages 233–241. Internatio...

    work page 2019

  7. [7]

    Large scale gan training for high fidelity natural image synthesis

    Andrew Brock, Jeff Donahue, and Karen Simonyan. Large scale gan training for high fidelity natural image synthesis. International Conference on Learning Representations, 2018. 19 A PREPRINT - J UNE 17, 2026

    work page 2018

  8. [8]

    Zero-sum polymatrix games: A generalization of minmax

    Yang Cai, Ozan Candogan, Constantinos Daskalakis, and Christos Papadimitriou. Zero-sum polymatrix games: A generalization of minmax. Mathematics of Operations Research, 41(2):648–655, 2016

    work page 2016

Show all 33 references
  1. [9]

    Accelerated algorithms for constrained nonconvex- nonconcave min-max optimization and comonotone inclusion

    Yang Cai, Argyris Oikonomou, and Weiqiang Zheng. Accelerated algorithms for constrained nonconvex- nonconcave min-max optimization and comonotone inclusion. In Ruslan Salakhutdinov, Zico Kolter, Katherine Heller, Adrian Weller, Nuria Oliver, Jonathan Scarlett, and Felix Berken...

    2024

  2. [10]

    Tikhonov regularization and the l-curve for large discrete ill-posed problems

    Daniela Calvetti, Serena Morigi, Lothar Reichel, and Fiorella Sgallari. Tikhonov regularization and the l-curve for large discrete ill-posed problems. Journal of computational and applied mathematics , 123(1-2):423–446, 2000

    2000

  3. [11]

    Prediction, learning, and games

    Nicolo Cesa-Bianchi and G ´abor Lugosi. Prediction, learning, and games. Cambridge university press, 2006

    2006

  4. [12]

    V ox2vox: 3d-gan for brain tumour segmentation

    Marco Domenico Cirillo, David Abramian, and Anders Eklund. V ox2vox: 3d-gan for brain tumour segmentation. In Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries: 6th International Workshop, BrainLes 2020, Held in Conjunction with MICCAI 2020, Lima, Pe...

    2020

  5. [13]

    Last-iterate convergence: Zero-sum games and constrained min-max opti- mization

    C Daskalakis and Ioannis Panageas. Last-iterate convergence: Zero-sum games and constrained min-max opti- mization. In 10th Innovations in Theoretical Computer Science (ITCS) conference, ITCS 2019, 2019

    2019

  6. [14]

    Near-optimal no-regret algorithms for zero- sum games

    Constantinos Daskalakis, Alan Deckelbaum, and Anthony Kim. Near-optimal no-regret algorithms for zero- sum games. In Proceedings of the twenty-second annual ACM-SIAM symposium on Discrete Algorithms, pages 235–254. SIAM, 2011

    2011

  7. [15]

    Near-optimal no-regret learning in general games, 2021

    Constantinos Daskalakis, Maxwell Fishelson, and Noah Golowich. Near-optimal no-regret learning in general games, 2021

    2021

  8. [16]

    Training gans with optimism

    Constantinos Daskalakis, Andrew Ilyas, Vasilis Syrgkanis, and Haoyang Zeng. Training gans with optimism. arXiv preprint arXiv:1711.00141, 2017

    2017 arXiv

  9. [17]

    The limit points of (optimistic) gradient descent in min-max optimization

    Constantinos Daskalakis and Ioannis Panageas. The limit points of (optimistic) gradient descent in min-max optimization. Advances in neural information processing systems, 31, 2018

    2018

  10. [18]

    Multiagent online learning in time-varying games

    Benoit Duvocelle, Panayotis Mertikopoulos, Mathias Staudigl, and Dries Vermeulen. Multiagent online learning in time-varying games. Mathematics of Operations Research, 48(2):914–941, 2023

    2023

  11. [19]

    Last iterate is slower than averaged iterate in smooth convex-concave saddle point problems

    Noah Golowich, Sarath Pattathil, Constantinos Daskalakis, and Asuman Ozdaglar. Last iterate is slower than averaged iterate in smooth convex-concave saddle point problems. In Jacob Abernethy and Shivani Agarwal, editors, Proceedings of Thirty Third Conference on Learning Theor...

    2020

  12. [20]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In Advances in neural information processing systems, pages 2672–2680, 2014

    2014

  13. [21]

    Long-time Energy Conservation , page 162–180

    Ernst Hairer. Long-time Energy Conservation , page 162–180. London Mathematical Society Lecture Note Series. Cambridge University Press, 2006

    2006

  14. [22]

    Precomputed real-time texture synthesis with markovian generative adversarial networks

    Chuan Li and Michael Wand. Precomputed real-time texture synthesis with markovian generative adversarial networks. In Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part III 14, pages 702–716. Springer, 2016

    2016

  15. [23]

    Cycles in adversarial regularized learning

    Panayotis Mertikopoulos, Christos Papadimitriou, and Georgios Piliouras. Cycles in adversarial regularized learning. In Proceedings of the Twenty-Ninth Annual ACM-SIAM Symposium on Discrete Algorithms , SODA ’18, page 2703–2717, USA, 2018

    2018

  16. [24]

    Convergence rate of O(1/k) for optimistic gradient and extragradient methods in smooth convex-concave saddle point problems

    Aryan Mokhtari, Asuman E Ozdaglar, and Sarath Pattathil. Convergence rate of O(1/k) for optimistic gradient and extragradient methods in smooth convex-concave saddle point problems. SIAM Journal on Optimization , 30(4):3230–3251, 2020

    2020

  17. [25]

    Arkadi Nemirovski. Prox-method with rate of convergence o (1/t) for variational inequalities with lipschitz con- tinuous monotone operators and smooth convex-concave saddle point problems.SIAM Journal on Optimization, 15(1):229–251, 2004. 20 A PREPRINT - J UNE 17, 2026

    2004

  18. [26]

    Excessive gap technique in nonsmooth convex minimization

    Yu Nesterov. Excessive gap technique in nonsmooth convex minimization. SIAM Journal on Optimization , 16(1):235–249, 2005

    2005

  19. [27]

    An online mechanism for resource allocation in networks

    Shi Pu, J Joaquin Escudero-Garz ´as, Alfredo Garcia, and Shahin Shahrampour. An online mechanism for resource allocation in networks. IEEE Transactions on Control of Network Systems, 7(3):1140–1150, 2020

    2020

  20. [28]

    Twenty lectures on algorithmic game theory

    Tim Roughgarden. Twenty lectures on algorithmic game theory. Cambridge University Press, 2016

    2016

  21. [29]

    On the Theory of Games of Strategy

    John von Neumann. Zur Theorie der Gesellschaftsspiele. Mathematische Annalen, 100:295–320, 1928. Trans- lated by S. Bargmann as “On the Theory of Games of Strategy” in A. Tucker and R. D. Luce, editors, Contri- butions to the Theory of Games IV, volume 40 of Annals of Mathemat...

    1928

  22. [30]

    Accelerated algorithms for smooth convex-concave minimax problems with o (1/kˆ 2) rate on squared gradient norm

    TaeHo Yoon and Ernest K Ryu. Accelerated algorithms for smooth convex-concave minimax problems with o (1/kˆ 2) rate on squared gradient norm. In International Conference on Machine Learning, pages 12098–12109. PMLR, 2021

    2021

  23. [31]

    Deform-gan: An unsupervised learning model for deformable registration

    Xiaoyue Zhang, Weijian Jian, Yu Chen, and Shihting Yang. Deform-gan: An unsupervised learning model for deformable registration. arXiv preprint arXiv:2002.11430, 2020

    2002 arXiv

  24. [32]

    Bayesian conditional gan for mri brain image synthesis

    Gengyan Zhao, Mary E Meyerand, and Rasmus M Birn. Bayesian conditional gan for mri brain image synthesis. arXiv preprint arXiv:2005.11875, 2020

    2005 arXiv

  25. [33]

    3d high resolution generative deep-learning network for fluorescence microscopy imaging

    Hang Zhou, Ruiyao Cai, Tingwei Quan, Shijie Liu, Shiwei Li, Qing Huang, Ali Ert ¨urk, and Shaoqun Zeng. 3d high resolution generative deep-learning network for fluorescence microscopy imaging. Optics letters , 45(7):1695–1698, 2020. A Proof of Theorem 1 Theorem 1. The energy f...

    2020

Pith tools

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