Deep Residual Attention Reinforcement Learning

Deep Residual Attention Reinforcement Learning
复制标题

深度残余注意力强化学习

DOI:
10.1109/taai48200.2019.8959896
复制
发表时间:
2019
期刊:
International Conference on Technologies and Applications of Artificial Intelligence
影响因子:
--
通讯作者:
Hanhua Zhu and Tomoyuki Kaneko
Hanhua Zhu and Tomoyuki Kaneko
中科院分区:
--
文献类型:
--
作者:
Hanhua Zhu and Tomoyuki Kaneko

文献摘要

参考文献

被引文献

相似文献

在强化学习中,更多地基于与给定视觉输入中的奖励密切相关的关键对象进行决策是有利的。在这项工作中,我们将基于注意力的结构纳入重要性加权演员-学习者架构(IMPALA)的网络结构中,以帮助模型找到关键对象,并提出深度剩余注意力强化学习(DRARL)。在Atari游戏和具有更多无关对象的特殊环境中的实验证明了DRARL在多对象环境中的优越性。此外,可视化的训练代理的注意力表明,额外的注意力机制有助于IMPALA集中在关键的对象,从而提高了IMPALA的性能。
Making decisions based more on the crucial objects which are closely connected to the reward in a given visual input is advantageous in reinforcement learning. In this work, we incorporate an attention-based structure into the network structure of Importance Weighted Actor-Learner Architecture (IMPALA) to help the model find out the crucial objects and propose Deep Residual Attention Reinforcement Learning (DRARL). Experiments in Atari games and special environments which have more irrelevant objects than usual demonstrate the superiority of DRARL in the multi-objects environment compared to the original IMPALA. Furthermore, the visualization of trained agents' attention indicates that the additional attention mechanism helps IMPALA concentrate on the crucial objects and therefore improves the performance of IMPALA.
DOI: --
发表时间: 2015-02
期刊: --
影响因子: --
作者:
Ke Xu;Jimmy Ba;Ryan Kiros;Kyunghyun Cho;Aaron C. Courville;R. Salakhutdinov;R. Zemel;Yoshua Bengio-Yoshua-Ben
通讯作者: Ke Xu;Jimmy Ba;Ryan Kiros;Kyunghyun Cho;Aaron C. Courville;R. Salakhutdinov;R. Zemel;Yoshua Bengio-Yoshua-Ben
视觉选择性注意与深度强化学习相结合的初步尝试
DOI: --
发表时间: 2018
期刊: arXiv.org
影响因子: --
作者:
Liu Yuezhang;Ruohan Zhang;D. Ballard
通讯作者: D. Ballard