Deep Q-Networks for Accelerating the Training of Deep Neural Networks

Jie Fu

In this paper, we propose a principled deep reinforcement learning (RL) approach that is able to accelerate the convergence rate of general deep neural networks (DNNs). With our ap…