Autonomous Visions – Which training methods for GANs do actually converge?

training methods for GANs

Recent work has shown local convergence of training methods for GANs for absolutely continuous data and generator distributions. In this paper, we show that the requirement of absolute continuity is necessary: we describe a simple yet prototypical counterexample showing that in the more realistic case of distributions that are not absolutely continuous, un-regularized GAN training is not always convergent.

We are interested in computer vision and machine learning with a focus on 3D scene understanding, parsing, reconstruction, material and motion estimation for autonomous intelligent systems such as self-driving cars or household robots.

For further reading

Click here to connect with us

Spread the word

Related posts