Trending

Bookmarks

Keep Up With Latest Trending Papers. Computer Science, AI and Machine Learning and more.Subscribe

Machine Learning

Artificial Intelligence

Convergence Guarantees for Min-Max GANs

Generative Minimization Networks: Training GANs Without Competition

We propose to address this problem by optimizing a different objective that circumvents the min-max structure of the generative model (gan) objective using the notion of duality gap from game theory.

We provide novelconvergence guarantees on this objective and demonstrate why the obtained limit point solves the problem better than known techniques.

Generative models, duality gap, game theory, convergence, convergence to a non-optimum cycle.

Authors

Paulina Grnarova, Yannic Kilcher, Kfir Y. Levy, Aurelien Lucchi, Thomas Hofmann

Related Topics

Generative models

Optimization techniques

Gradient-based techniques

Read the Paper

Read Paper

Trending

Bookmarks

◐ Recommended Follows

Founder 42papers.com. Also bui...

More

◐ Latest Activity

More

Convergence Guarantees for Min-Max GANs - 42Papers

Read Related Papers

Input Convex Gradient Networks

Adversarial AutoAugment

A Method for Evaluating the Capacity of Generative Adversarial Networks to Reproduce High-order Spatial Context

Adversarial Learning for deformable image registration

Latent Space Direction for Generative Adversarial Networks

Accelerating Deep Learning with Dynamic Data Pruning

A Deep generative Model for Video Generation

Transfer Learning with Gaussian Processes for Bayesian Optimization

No-Regret Dynamics in the Fenchel Game: A Unified Framework for Algorithmic Convex Optimization

Cycle Consistent Probability Divergences Across Different Spaces