An Initial Attempt of Combining Visual Selective Attention with Deep Reinforcement Learning

An Initial Attempt of Combining Visual Selective Attention with Deep Reinforcement Learning
复制标题

视觉选择性注意与深度强化学习相结合的初步尝试

DOI:
--
复制
发表时间:
2018
期刊:
arXiv.org
影响因子:
--
通讯作者:
D. Ballard
D. Ballard
中科院分区:
--
文献类型:
--
作者:
Liu Yuezhang;Ruohan Zhang;D. Ballard

文献摘要

参考文献

被引文献

相似文献

视觉注意是知觉系统中特征选择机制的一种手段。在布罗德本特的选择性注意漏失过滤模型的启发下,我们评估了这种机制是如何实现的,并影响了深度强化学习的学习过程。我们可视化并分析了DQN在一个玩具问题捕捉上的特征映射,并提出了一种将视觉选择性注意与深度强化学习相结合的方法。我们在Atari游戏上进行了基于光流的注意力和A2C的实验。实验结果表明,视觉选择性注意可以提高被测游戏的样本效率。还发现了注意力和批次标准化之间的有趣关系。
Visual attention serves as a means of feature selection mechanism in the perceptual system. Motivated by Broadbent's leaky filter model of selective attention, we evaluate how such mechanism could be implemented and affect the learning process of deep reinforcement learning. We visualize and analyze the feature maps of DQN on a toy problem Catch, and propose an approach to combine visual selective attention with deep reinforcement learning. We experiment with optical flow-based attention and A2C on Atari games. Experiment results show that visual selective attention could lead to improvements in terms of sample efficiency on tested games. An intriguing relation between attention and batch normalization is also discovered.
DOI: --
发表时间: --
期刊: --
影响因子: --
作者:
Christopher Kanan;Matthew H Tong;Lingyun Zhang;G. Cottrell
通讯作者: Christopher Kanan;Matthew H Tong;Lingyun Zhang;G. Cottrell