arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1810.04038cs.LGcs.AIstat.ML
收藏

基于连续注意的人类活动识别递归网络的理解与改进

Understanding and Improving Recurrent Networks for Human Activity Recognition by Continuous Attention

Carnegie Mellon University(卡内基梅隆大学);The Norwegian University of Science and Technology(挪威科技大学);Google Brain(谷歌大脑)

Ming Zeng, Haoxiang Gao, Tong Yu, Ole J. Mengshoel, Helge Langseth, Ian Lane, Xiaobing Liu

摘要:

深度神经网络,包括递归网络,已经成功应用于人类活动识别。不幸的是,递归网络学习的最终表示可能会编码一些噪声(不相关的信号分量,不重要的传感器模态等)。此外,很难解释递归网络来深入了解模型的行为。为了解决这些问题,我们提出了两个人类活动识别的注意力模型:时间注意力和传感器注意力。这两种机制自适应地聚焦于重要信号和传感器模态。为了进一步提高可理解性和平均F1分数,我们增加了连续性约束,考虑到连续传感器信号比离散信号更鲁棒。我们在三个数据集上评估了这些方法,并获得了最先进的结果。此外,定性分析表明,模型学习到的注意力与人类的直觉一致。

英文摘要:

Deep neural networks, including recurrent networks, have been successfully applied to human activity recognition. Unfortunately, the final representation learned by recurrent networks might encode some noise (irrelevant signal components, unimportant sensor modalities, etc.). Besides, it is difficult to interpret the recurrent networks to gain insight into the models' behavior. To address these issues, we propose two attention models for human activity recognition: temporal attention and sensor attention. These two mechanisms adaptively focus on important signals and sensor modalities. To further improve the understandability and mean F1 score, we add continuity constraints, considering that continuous sensor signals are more robust than discrete ones. We evaluate the approaches on three datasets and obtain state-of-the-art results. Furthermore, qualitative analysis shows that the attention learned by the models agree well with human intuition.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑