753 citations · 1.4k across the 10 of their papers we have counts for
3 papers · 1 filter
Searching for Activation Functions
Prajit Ramachandran, Barret Zoph, Quoc V. Le
The choice of activation functions in deep networks has a significant effect on the training dynamics and task performance. Currently, the most successful and widely-used activatio…
Fast Generation for Convolutional Autoregressive Models
Prajit Ramachandran, Tom Le Paine, Pooya Khorrami +6
Convolutional autoregressive models have recently demonstrated state-of-the-art performance on a number of generation tasks. While fast, parallel training methods have been crucial…
Stein Variational Policy Gradient
Yang Liu, Prajit Ramachandran, Qiang Liu +1
Policy gradient methods have been successfully applied to many complex reinforcement learning problems. However, policy gradient methods suffer from high variance, slow convergence…