Deep learning frameworks have often focused on either usability or speed...
In complex tasks, such as those with large combinatorial action spaces,
...
We formulate the problem of defogging as state estimation and future sta...
We present Value Propagation (VProp), a parameter-efficient differentiab...
We release a dataset of 65646 StarCraft replays that contains 1535 milli...
Recent studies have shown that deep neural networks (DNN) are vulnerable...
We present TorchCraft, a library that enables deep learning research on
...
We consider scenarios from the real-time strategy game StarCraft as new
...