arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1709.06247cs.CV
收藏

重新思考ReLU以训练更好的CNN

Rethink ReLU to Training Better CNNs

CASIA(中国科学院自动化研究所);University of Chinese Academy of Sciences(中国科学院大学);CAS Center for Excellence in Brain Science and Intelligence Technology(中国科学院脑科学与智能技术卓越创新中心);Huazhong University of Science and Technology(华中科技大学);Microsoft Research(微软研究院)

Gangming Zhao, Zhaoxiang Zhang, He Guan, Peng Tang, Jingdong Wang

摘要:

大多数卷积神经网络都有一个共同的特点:每个卷积层后面都有一个非线性激活层,其中最广泛使用的是整流线性单元(ReLU)。在本文中,我们认为,所设计的结构之间的这两个层的比例可能不是最好的选择,因为它可能会导致泛化能力差。因此,我们试图研究一种更合适的方法来使用ReLU来探索更好的网络架构。具体来说,我们提出了一个比例模块来保持卷积和ReLU量之间的比例为N:M(N>M)。比例模块可以应用于几乎所有的网络,而不需要额外的计算成本,以提高性能。综合实验结果表明,该方法在不同网络架构的基准测试中均取得了较好的性能,从而验证了本文工作的优越性。

英文摘要:

Most of convolutional neural networks share the same characteristic: each convolutional layer is followed by a nonlinear activation layer where Rectified Linear Unit (ReLU) is the most widely used. In this paper, we argue that the designed structure with the equal ratio between these two layers may not be the best choice since it could result in the poor generalization ability. Thus, we try to investigate a more suitable method on using ReLU to explore the better network architectures. Specifically, we propose a proportional module to keep the ratio between convolution and ReLU amount to be N:M (N>M). The proportional module can be applied in almost all networks with no extra computational cost to improve the performance. Comprehensive experimental results indicate that the proposed method achieves better performance on different benchmarks with different network architectures, thus verify the superiority of our work.

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。

↑