mlsanityData to 7 Oct 2026

Generative Minimization Networks: Training GANs Without Competition

arXiv 2103.12685 ↗First posted 23 Mar 2021cs.LG cs.AIcited by 3

Not on the map. Why?

Abstract (arXiv)

Many applications in machine learning can be framed as minimization problems and solved efficiently using gradient-based techniques. However, recent applications of generative models, particularly GANs, have triggered interest in solving min-max games for which standard optimization techniques are often not suitable. Among known problems experienced by practitioners is the lack of convergence guarantees or convergence to a non-optimum cycle. At the heart of these problems is the min-max structure of the GAN objective which creates non-trivial dependencies between the players. We propose to address this problem by optimizing a different objective that circumvents the min-max structure using the notion of duality gap from game theory. We provide novel convergence guarantees on this objective and demonstrate why the obtained limit point solves the problem better than known techniques.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 37 of its 49 references that are in the map. First posted on arXiv 23 Mar 2021.

Builds on

37 of 49 references.

References · 37

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2016Improved Techniques for Training GANscited by 2,938
  3. 2017Improved Training of Wasserstein GANscited by 2,223
  4. 2015Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networkscited by 2,048
  5. 2018Spectral Normalization for Generative Adversarial Networkscited by 1,618
  6. 2016Least Squares Generative Adversarial Networkscited by 1,000
  7. 2018Which Training Methods for GANs do actually Converge?cited by 672
  8. 2017Wasserstein GANcited by 635
  9. 2016f-GAN: Training Generative Neural Samplers using Variational Divergence Minimizationcited by 504
  10. 2016Adversarial Feature Learningcited by 350
  11. 2014Identifying and attacking the saddle point problem in high-dimensional non-convex optimizationcited by 286
  12. 2017Are GANs Created Equal? A Large-Scale Studycited by 267
25 more
  1. 2019Large Scale Adversarial Representation Learningcited by 238
  2. 2017Training GANs with Optimismcited by 217
  3. 2016Adversarially Learned Inferencecited by 216
  4. 2016Unrolled Generative Adversarial Networkscited by 208
  5. 2019Solving a Class of Non-Convex Min-Max Games Using Iterative First Order Methodscited by 148
  6. 2018A Variational Inequality Perspective on Generative Adversarial Networkscited by 144
  7. 2013Stochastic Gradient Descent, Weighted Sampling, and the Randomized Kaczmarz algorithmcited by 132
  8. 2017Non-convex Optimization for Machine Learningcited by 132
  9. 2017BEGAN: Boundary Equilibrium Generative Adversarial Networkscited by 121
  10. 2017Stabilizing Training of Generative Adversarial Networks through Regularizationcited by 118
  11. 2018The Mechanics of n-Player Differentiable Gamescited by 118
  12. 2019Reducing Noise in GAN Training with Variance Reduced Extragradientcited by 61
  13. 2018Negative Momentum for Improved Game Dynamicscited by 59
  14. 2017Online to Offline Conversions, Universality and Adaptive Minibatch Sizescited by 39
  15. 2020Non-convex Min-Max Optimization: Applications, Challenges, and Recent Theoretical Advancescited by 37
  16. 2019On Solving Minimax Optimization Locally: A Follow-the-Ridge Approachcited by 34
  17. 2019Last-iterate convergence rates for min-max optimizationcited by 25
  18. 2019A Closer Look at the Optimization Landscapes of Generative Adversarial Networkscited by 25
  19. 2018A Convex Duality Framework for GANscited by 15
  20. 2019Implicit competitive regularization in GANscited by 14
  21. 2017An Online Learning Approach to Generative Adversarial Networkscited by 13
  22. 2019Investigating Under and Overfitting in Wasserstein Generative Adversarial Networkscited by 7
  23. 2017Dualing GANscited by 4
  24. 2018Primal-Dual Wasserstein GANcited by 4
  25. 2018Training Generative Adversarial Networks via Primal-Dual Subgradient Methods: A Lagrangian Perspective on GANcited by 3

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 1; 2023: 1.

3 papers.

  1. 2021Stochastic Projective Splitting: Solving Saddle-Point Problems with Multiple Regularizerscited by 1
  2. 2022Finding mixed-strategy equilibria of continuous-action games without gradients using randomized policy networkscited by 0
  3. 2023ApproxED: Approximate exploitability descent via learned best responsescited by 0