Skip RNN: learning to skip state updates in recurrent neural networks

International Conferences 2018