Learning to See by Moving

28 Jun 2018 in Studies on Deep Learning, Computer Vision

WHY?

Feature learning in Convolution Neural Network requires many hand labeled data. It would be useful if one can use other form of supervision. In nature world, organisms acquire many essential information regarding vision by moving itself(egomotion).

Adversarial Variational Bayes: Unifying Variational Autoencoder and Generative Adversarial Networks

18 Jun 2018 in Studies on Deep Learning, Generative Models

WHY?

In VAE framework, the quality of the generation relies on the expressiveness of inference model. Restricting hidden variables to Gaussian distribution with KL divergence limits the expressiveness of the model.