Generating new pictures in complex datasets with a simple neural network
arXiv:1810.12478
Abstract
We introduce a version of a variational auto-encoder (VAE), which can generate good perturbations of images, when trained on a complex dataset (in our experiments, CIFAR-10). The net is using only two latent generative dimensions per class, with uni-modal probability density. The price one has to pay for good generation is that not all training images are well reconstructed. An additional classifier is required to determine which training image is well reconstructed and generally the weights of training images. Only training images which are well reconstructed, can be perturbed. For good perturbations, we use the tentative empirical drifts of well reconstructed images. The construct is not predictive in the usual statistical sense.
arXiv comment13 pages plus some images
References in corpus (15)
- Auto-Encoding Variational Bayes
- Stochastic Backpropagation and Approximate Inference in Deep Generative Models
- Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
- Pixel Recurrent Neural Networks
- NICE: Non-linear Independent Components Estimation
- NIPS 2016 Tutorial: Generative Adversarial Networks
- Adversarially Learned Inference
- A note on the evaluation of generative models
- Batch Renormalization: Towards Reducing Minibatch Dependence in Batch-Normalized Models
- Do GANs actually learn the distribution? An empirical study
- Distribution Matching in Variational Inference
- A Generative Process for Sampling Contractive Auto-Encoders
- Linear Algebra and Duality of Neural Networks
- Symmetries and control in generative neural nets
- Towards universal neural nets: Gibbs machines and ACE