Hyperprior Induced Unsupervised Disentanglement of Latent Representations
arXiv:1809.04497
Abstract
We address the problem of unsupervised disentanglement of latent representations learnt via deep generative models. In contrast to current approaches that operate on the evidence lower bound (ELBO), we argue that statistical independence in the latent space of VAEs can be enforced in a principled hierarchical Bayesian manner. To this effect, we augment the standard VAE with an inverse-Wishart (IW) prior on the covariance matrix of the latent code. By tuning the IW parameters, we are able to encourage (or discourage) independence in the learnt latent dimensions. Extensive experimental results on a range of datasets (2DShapes, 3DChairs, 3DFaces and CelebA) show our approach to outperform the -VAE and is competitive with the state-of-the-art FactorVAE. Our approach achieves significantly better disentanglement and reconstruction on a new dataset (CorrelatedEllipses) which introduces correlations between the factors of variation.
arXiv commentAAAI-2019
References in corpus (15)
- Adam: A Method for Stochastic Optimization
- Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
- Improved Training of Wasserstein GANs
- Tutorial on Variational Autoencoders
- InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets
- Deep Convolutional Inverse Graphics Network
- Disentangling by Factorising
- Adversarial Autoencoders
- Unsupervised Learning of Disentangled Representations from Video
- Unsupervised Learning of Disentangled and Interpretable Representations from Sequential Data
- Learning Disentangled Representations with Semi-Supervised Deep Generative Models
- Implicit Reparameterization Gradients
- Disentangling Factors of Variation via Generative Entangling
- Stick-Breaking Variational Autoencoders
- InfoCatVAE: Representation Learning with Categorical Variational Autoencoders