Parametrizing filters of a CNN with a GAN
This addresses the need for better invariance modeling in ML, though it appears incremental as it builds on existing CNN and GAN frameworks.
The paper tackled the problem of modeling richer, high-level invariances in machine learning by introducing a method to parametrize CNN filters with a GAN's latent space, and demonstrated its ability to capture non-linear invariances through visualizations in data space.
It is commonly agreed that the use of relevant invariances as a good statistical bias is important in machine-learning. However, most approaches that explicitly incorporate invariances into a model architecture only make use of very simple transformations, such as translations and rotations. Hence, there is a need for methods to model and extract richer transformations that capture much higher-level invariances. To that end, we introduce a tool allowing to parametrize the set of filters of a trained convolutional neural network with the latent space of a generative adversarial network. We then show that the method can capture highly non-linear invariances of the data by visualizing their effect in the data space.