Networks adjusting networks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Decoupled Neural Interfaces using Synthetic Gradients (2016, methodology): “Our synthetic gradient model is most analogous to a value function which is used for gradient ascent (Bax- ter & Bartlett, 2000) or critics for training neural networks (Schmidhuber, 1990).”
Reinforcement learning by backpropagation through an LSTM model/critic (2007, background): “If the state space is large and/or continuous and only partially observable, and the action space is large and/or continuous, the BTM approach based on recurrent neural networks is one of the few options to handle this very difficult case [13], [17].”
Making the World Di erentiable: On Using Self-Supervised Fully Recurrent Neural Networks for Dynamic Reinforcement Learning (1990, background): “(In [29] a pole balancing experiment with a non-recurrent, vector-valued adaptive critic is described.”
Reinforcement Learning with Long Short-Term Memory (2001, methodology): “An alternative, model-free approach, and the one used here, is to use the RNN to directly approximate the value function of a reinforcement learning algorithm [10, 6].”