- Dream Team Machine Here we want to share the experience we gained and challenges we faced when creating the machine for deep learning tasks.
- Deep Attention Recurrent Q-Network Description of our paper at the first-ever Deep Reinforcement Learning Workshop at NIPS 2015 in Montréal, Canada.
- Pretrain DQN with ISA In this post we explain how to extend DQN model by adding the unsupervised pre-training stage for the first convolutional layer.