IB-GAN: Disentangled Representation Learning with Information Bottleneck Generative Adversarial Networks
Keywords:Neural Generative Models & Autoencoders, Representation Learning, Adversarial Learning & Robustness, Unsupervised & Self-Supervised Learning
AbstractWe propose a new GAN-based unsupervised model for disentangled representation learning. The new model is discovered in an attempt to utilize the Information Bottleneck (IB) framework to the optimization of GAN, thereby named IB-GAN. The architecture of IB-GAN is partially similar to that of InfoGAN but has a critical difference; an intermediate layer of the generator is leveraged to constrain the mutual information between the input and the generated output. The intermediate stochastic layer can serve as a learnable latent distribution that is trained with the generator jointly in an end-to-end fashion. As a result, the generator of IB-GAN can harness the latent space in a disentangled and interpretable manner. With the experiments on dSprites and Color-dSprites dataset, we demonstrate that IB-GAN achieves competitive disentanglement scores to those of state-of-the-art β-VAEs and outperforms InfoGAN. Moreover, the visual quality and the diversity of samples generated by IB-GAN are often better than those by β-VAEs and Info-GAN in terms of FID score on CelebA and 3D Chairs dataset.
How to Cite
Jeon, I., Lee, W., Pyeon, M., & Kim, G. (2021). IB-GAN: Disentangled Representation Learning with Information Bottleneck Generative Adversarial Networks. Proceedings of the AAAI Conference on Artificial Intelligence, 35(9), 7926-7934. Retrieved from https://ojs.aaai.org/index.php/AAAI/article/view/16967
AAAI Technical Track on Machine Learning II