LOGAN: Latent Optimisation for Generative Adversarial Networks

By Y. Wu, Jeff Donahue, David Balduzzi, Karen Simonyan, Timothy P. Lillicrap. 2019

92 citations recorded by the source corpus.

Training generative adversarial networks requires balancing of delicate adversarial dynamics. Even with careful tuning, training may diverge or end up in a bad equilibrium with dropped modes. In this work, we improve CS-GAN with natural gradient-based latent optimisation and show that it improves adversarial dynamics by enhancing interactions between the discriminator and the generator. Our experiments demonstrate that latent optimisation can significantly improve GAN training, obtaining state-of-the-art performance for the ImageNet ($128 \times 128$) dataset. Our model achieves an Inception Score (IS) of $148$ and an Frechet Inception Distance (FID) of $3.4$, an improvement of $17\%$ and $32\%$ in IS and FID respectively, compared with the baseline BigGAN-deep model with the same architecture and number of parameters.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Generative Models

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Generative Models

Citations in the corpus

Your GAN is Secretly an Energy-based Model and You Should use Discriminator Driven Latent Sampling (2020, background): “Moreover, it has recently been found in LOGAN [20] that optimizing D on p t rather than p g can lead to better performance.”
Diffusion Models Beat GANs on Image Synthesis (2021, background): “GANs [16] currently hold the state-of-the-art on most image generation tasks [5, 60, 22] as measured by sample quality metrics such as FID [18], Inception Score [47] and Precision [26].”
DeepSMOTE: Fusing Deep Learning and SMOTE for Imbalanced Data (2021, background): “GANs have achieved impressive results in the computer vision arena [70], [71].”
Improved Consistency Regularization for GANs (2020, methodology): “Generative Adversarial Networks (GANs; Goodfellow et al., 2014) are a powerful class of deep generative models, but are known for training difﬁculties (Salimans et al., 2016).”
A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications (2020, background): “There are many papers related to GANs [126]–[131] such as CSGAN [132] and LOGAN [133].”
Refining Deep Generative Models via Discriminator Gradient Flow (2020, background): “Latent Optimization GAN (LOGAN) (Wu et al., 2020) optimizes a latent vector via natural gradient descent as part of the GAN training process.”
ContraGAN: Contrastive Learning for Conditional Image Generation (2020, methodology): “Also, we think that exploring advanced regularization techniques [8, 9, 22, 23] is necessary to understand ContraGAN further.”
Minimax Theorem for Latent Games or: How I Learned to Stop Worrying about Mixed-Nash and Love Neural Nets (2020, background): “At the same time, remarkable progress in generative modeling of images (Wu et al., 2019) and speech synthesis (Bi´nkowski et al., 2020) has resulted from zero-sum games designed speciﬁcally to facilitate via carefully constructed arms races (Goodfellow et al., 2014).”
Instance Selection for GANs (2020, background): “Latent Optimisation for Generative Adversarial Networks (LOGAN) [28] optimizes latent samples each iteration at the cost of an additional forward and backward pass.”
Generalized Energy Based Models (2020, background): “While there has been recent interest in using the discriminator to improve the quality of the generator during sampling (Azadi et al., 2019; Turner et al., 2019; Neklyudov et al., 2019; Grover et al., 2019; Tanaka, 2019; Wu et al., 2019b), our approach emerges naturally from the model we consider.”