arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.15898cs.CVcs.AI
收藏

不同卷积神经网络激活函数的比较及集成构建方法

Comparison of different convolutional neural network activation functions and methods for building ensembles

University of Padua(帕多瓦大学);Missouri State University(密苏里州立大学);Tampere University(坦佩雷大学)

Loris Nanni, Gianluca Maguolo, Sheryl Brahnam, Michelangelo Paci

摘要:

最近,人们一直致力于寻找高效和强大的激活函数的CNN层。由于激活函数在影响性能的层之间注入不同的非线性,因此改变它们是构建CNN的鲁棒集成的一种方法。本研究的目的是检查使用不同激活函数制作的CNN集合的性能,包括这里提出的六个新函数:2D Mexican ReLU,TanELU,MeLU+GaLU,对称MeLU,对称GaLU和灵活MeLU。性能最高的集成是用具有不同激活层的CNN构建的,这些激活层随机取代了标准ReLU。在15个生物医学数据集代表各种分类任务的建议的方法进行了全面的评估。该方法在两种基本的CNN架构上进行了测试:Vgg 16和ResNet 50。结果表明,这种方法的性能优越。本研究的MATLAB源代码将在https://github.com/LorisNanni上提供。

英文摘要:

Recently, much attention has been devoted to finding highly efficient and powerful activation functions for CNN layers. Because activation functions inject different nonlinearities between layers that affect performance, varying them is one method for building robust ensembles of CNNs. The objective of this study is to examine the performance of CNN ensembles made with different activation functions, including six new ones presented here: 2D Mexican ReLU, TanELU, MeLU+GaLU, Symmetric MeLU, Symmetric GaLU, and Flexible MeLU. The highest performing ensemble was built with CNNs having different activation layers that randomly replaced the standard ReLU. A comprehensive evaluation of the proposed approach was conducted across fifteen biomedical data sets representing various classification tasks. The proposed method was tested on two basic CNN architectures: Vgg16 and ResNet50. Results demonstrate the superiority in performance of this approach. The MATLAB source code for this study will be available at https://github.com/LorisNanni.

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。

↑