RATM: Recurrent Attentive Tracking Model https://deeplearning.jp/wp-content/uploads/2015/10/dl_hacks20160115.pdf read more
On Learning to Think: Algorithmic Information Theory for Novel Combinations of Reinforcement Learning Controllers and Recurrent Neural World Models read more
Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks dl_hacks2015-10-29-kaneko read more