IB-GAN: Disentangled Representation Learning with Information Bottleneck Generative Adversarial Networks

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 44
  • Download : 0
We 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 beta-VAEs and outperforms InfoGAN. Moreover, the visual quality and the diversity of samples generated by IB-GAN are often better than those by beta-VAEs and Info-GAN in terms of FID score on CelebA and 3D Chairs dataset.
Publisher
ASSOC ADVANCEMENT ARTIFICIAL INTELLIGENCE
Issue Date
2021-02
Language
English
Citation

35th AAAI Conference on Artificial Intelligence / 33rd Conference on Innovative Applications of Artificial Intelligence / 11th Symposium on Educational Advances in Artificial Intelligence, pp.7926 - 7934

ISSN
2159-5399
URI
http://hdl.handle.net/10203/288488
Appears in Collection
RIMS Conference Papers
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0