IB-GAN: Disentangled Representation Learning with Information Bottleneck GANDownload PDF

27 Sept 2018, 22:39 (edited 05 Apr 2022)ICLR 2019 Conference Blind SubmissionReaders: Everyone
  • Keywords: Unsupervised disentangled representation learning, GAN, Information Bottleneck, Variational Inference
  • TL;DR: Inspired by Information Bottleneck theory, we propose a new architecture of GAN for a disentangled representation learning
  • Abstract: We present a novel architecture of GAN for a disentangled representation learning. The new model architecture is inspired by Information Bottleneck (IB) theory thereby named IB-GAN. IB-GAN objective is similar to that of InfoGAN but has a crucial difference; a capacity regularization for mutual information is adopted, thanks to which the generator of IB-GAN can harness a latent representation in disentangled and interpretable manner. To facilitate the optimization of IB-GAN in practice, a new variational upper-bound is derived. With experiments on CelebA, 3DChairs, and dSprites datasets, we demonstrate that the visual quality of samples generated by IB-GAN is often better than those by β-VAEs. Moreover, IB-GAN achieves much higher disentanglement metrics score than β-VAEs or InfoGAN on the dSprites dataset.
  • Code: [![github](/images/github_icon.svg) insuj3on/IB-GAN](https://github.com/insuj3on/IB-GAN) + [![Papers with Code](/images/pwc_icon.svg) 1 community implementation](https://paperswithcode.com/paper/?openreview=ryljV2A5KX)
  • Data: [CelebA](https://paperswithcode.com/dataset/celeba), [Chairs](https://paperswithcode.com/dataset/chairs), [dSprites](https://paperswithcode.com/dataset/dsprites)
12 Replies