weixin_40485472的博客

论文的Motivation是大部分效果比较好的RL算法都是基于大量的人工特征，而Deep Neural Network（CNN等）已经在图像等领域被证明具有很好的直接从原始数据中提取高层特征的能力，因此很自然的想到将Deep Neural Network与RL算法结合。 但是Deep learning与reinforcement learning有很大的不同： 1. deep learning...

提出问题： To use reinforcement learning successfully insituationsapproaching real-world complexity, however, agents are confrontedwith a difficult task: theymust derive efficient representations of the...