No Representation without Transformation
arXiv:1912.03845
Abstract
We extend the framework of variational autoencoders to represent transformations explicitly in the latent space. In the family of hierarchical graphical models that emerges, the latent space is populated by higher order objects that are inferred jointly with the latent representations they act on. To explicitly demonstrate the effect of these higher order objects, we show that the inferred latent transformations reflect interpretable properties in the observation space. Furthermore, the model is structured in such a way that in the absence of transformations, we can run inference and obtain generative capabilities comparable with standard variational autoencoders. Finally, utilizing the trained encoder, we outperform the baselines by a wide margin on a challenging out-of-distribution classification task.
Preprint. Accepted at BDL and PGR workshops at NeurIPS 2019
References in corpus (29)
- Auto-Encoding Variational Bayes
- Representation Learning with Contrastive Predictive Coding
- Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
- Prototypical Networks for Few-shot Learning
- Neural Discrete Representation Learning
- Stochastic Backpropagation and Approximate Inference in Deep Generative Models
- DRAW: A Recurrent Neural Network For Image Generation
- Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations
- Disentangling by Factorising
- Group Equivariant Convolutional Networks
- InfoVAE: Information Maximizing Variational Autoencoders
- Towards a Definition of Disentangled Representations
- Learning in Implicit Generative Models
- Gradient Estimation Using Stochastic Computation Graphs
- Towards a Neural Statistician
- Causality for Machine Learning
- Unsupervised Learning of Object Keypoints for Perception and Control
- Group Equivariant Capsule Networks
- Diagnosing and Enhancing VAE Models
- Learning Independent Causal Mechanisms
- Dynamics-Aware Unsupervised Discovery of Skills
- Weakly Supervised Disentanglement with Guarantees
- Explorations in Homeomorphic Variational Auto-Encoding
- Learning Independent Features with Adversarial Nets for Non-linear ICA
- Neural Predictive Belief Representations
- On multi-view feature learning
- Symmetry-Based Disentangled Representation Learning requires Interaction with Environments
- Only Bayes should learn a manifold (on the estimation of differential geometric structure from data)
- Associative Compression Networks for Representation Learning