Information Theoretic-Learning Auto-Encoder
This work addresses the challenge of sample generation in machine learning, but it appears incremental as it builds on existing frameworks like variational autoencoders and generative adversarial networks.
The paper tackles the problem of generating sample data without explicitly defining a partition function by proposing Information Theoretic-Learning (ITL) divergence measures for variational regularization of neural networks, resulting in ITL-regularized autoencoders as an alternative to methods like variational autoencoders and generative adversarial networks.
We propose Information Theoretic-Learning (ITL) divergence measures for variational regularization of neural networks. We also explore ITL-regularized autoencoders as an alternative to variational autoencoding bayes, adversarial autoencoders and generative adversarial networks for randomly generating sample data without explicitly defining a partition function. This paper also formalizes, generative moment matching networks under the ITL framework.