Video Prediction

Agents that can learn to imitate given video observation – without direct access to state or action information are more …

We introduce a variational inference model called VIM, for Variational Independent Modules, for sequential data that learns and infers …