Symmetries and control in generative neural nets
arXiv:1511.02841
Abstract
We study generative nets which can control and modify observations, after being trained on real-life datasets. In order to zoom-in on an object, some spatial, color and other attributes are learned by classifiers in specialized attention nets. In field-theoretical terms, these learned symmetry statistics form the gauge group of the data set. Plugging them in the generative layers of auto-classifiers-encoders (ACE) appears to be the most direct way to simultaneously: i) generate new observations with arbitrary attributes, from a given class, ii) describe the low-dimensional manifold encoding the "essence" of the data, after superfluous attributes are factored out, and iii) organically control, i.e., move or modify objects within given observations. We demonstrate the sharp improvement of the generative qualities of shallow ACE, with added spatial and color symmetry statistics, on the distorted MNIST and CIFAR10 datasets.
References in corpus (11)
- Auto-Encoding Variational Bayes
- Stochastic Backpropagation and Approximate Inference in Deep Generative Models
- Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
- Deep Unsupervised Learning using Nonequilibrium Thermodynamics
- NICE: Non-linear Independent Components Estimation
- Theano: new features and speed improvements
- DRAW: A Recurrent Neural Network For Image Generation
- Multiple Object Recognition with Visual Attention
- Deep Generative Stochastic Networks Trainable by Backprop
- Multi-column Deep Neural Networks for Image Classification
- Towards universal neural nets: Gibbs machines and ACE