Authors: Hyunjik Kim , Andriy Mnih - International Conference on Machine Learning, ICML 2018 cited by 1,574
We define and address the problem of unsupervised learning of disentangled representations on data generated from independent factors of variation. We propose FactorVAE, a method that disentangles by encouraging the distribution of representations to be factorial and hence independent across the dimensions. We show that it improves upon $\beta$-VAE by providing a better trade-off between disentanglement and reconstruction quality. Moreover, we highlight the problems of a commonly used disentanglement metric and introduce a new metric that does not suffer from them.
✨ Checking sign-in… PDF Cited by View BibTeX Hide BibTeX View BibTeX Cite
Learning Factorial Codes by Predictability… Learning Factorial Codes by Predictability Minimization Unsupervised Feature Learning and Deep… Unsupervised Feature Learning and Deep Learning: A Review and New Perspectives Learning to Disentangle Factors of Variation… Learning to Disentangle Factors of Variation with Manifold Interaction Deep Convolutional Inverse Graphics Network Deep Convolutional Inverse Graphics Network Adversarial Autoencoders Adversarial Autoencoders Batch Normalization: Accelerating Deep… Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift InfoGAN: Interpretable Representation Learning… InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets beta-VAE: Learning Basic Visual Concepts with a… beta-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework Learning Independent Features with… Learning Independent Features with Adversarial Nets for Non-linear ICA Wasserstein Generative Adversarial Networks Wasserstein Generative Adversarial Networks A Framework for the Quantitative Evaluation… A Framework for the Quantitative Evaluation of Disentangled Representations Distribution Matching in Variational Inference Distribution Matching in Variational Inference Relevance Factor VAE: Learning and Identifyin… Relevance Factor VAE: Learning and Identifying Disentangled Factors Multi-Object Representation Learning… Multi-Object Representation Learning with Iterative Variational Inference Hierarchical Disentanglement of… Hierarchical Disentanglement of Discriminative Latent Features for Zero-Shot Learning Weakly Supervised Disentanglement with… Weakly Supervised Disentanglement with Guarantees OOGAN: Disentangling GAN with One-Hot Sampling… OOGAN: Disentangling GAN with One-Hot Sampling and Orthogonal Regularization Variational Autoencoders and Nonlinear ICA: A… Variational Autoencoders and Nonlinear ICA: A Unifying Framework Independent Subspace Analysis for… Independent Subspace Analysis for Unsupervised Learning of Disentangled Representations Independence Promoted Graph Disentangled… Independence Promoted Graph Disentangled Networks Lifelong Teacher-Student Network Learning Lifelong Teacher-Student Network Learning Commutative Lie Group VAE for Disentanglement… Commutative Lie Group VAE for Disentanglement Learning GroupifyVAE: from Group-based Definition… GroupifyVAE: from Group-based Definition to VAE-based Unsupervised Representation Disentanglement Interpretable Machine Learning: Fundamental… Interpretable Machine Learning: Fundamental Principles and 10 Grand Challenges Disentangling by Factorising Disentangling by Factorising Earlier references Focus paper Citing papers Older Newer Click a node to pin it, click the empty canvas to go back to this paper, or hover to preview. Open a node’s page from its title.
