arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1802.00212cs.LG
收藏

用幂线性单元训练神经网络(PoLUs)

Training Neural Networks by Using Power Linear Units (PoLUs)

School of Computing, Informatics, and Decision Systems Engineering(计算、信息与决策系统工程学院);Arizona State University(亚利桑那州立大学)

Yikang Li, Pak Lun Kevin Ding, Baoxin Li

摘要:

在本文中,我们介绍了“幂线性单元”(PoLU),它增加了神经网络的非线性能力,从而有助于提高其性能。PoLU采用了先前提出的激活函数的几个优点。首先,正输入的PoLU的输出被设计为恒等的,以避免梯度消失问题。其次,PoLU对于负输入具有非零输出,使得单元的输出均值接近于零,从而减少偏置移位效应。第三,PoLU的负部分存在饱和,这使得它对负输入更具噪声鲁棒性。此外,我们证明了PoLU能够通过使用幂函数将每层输入的更多部分映射到相同的空间,从而增加了神经网络的响应区域的数量。我们使用图像分类比较我们提出的激活函数与其他人。在实验中,MNIST,CIFAR-10,CIFAR-100,街景门牌号(SVHN)和ImageNet被用作基准数据集。我们实现的神经网络包括广泛使用的ELU网络,ResNet-50和VGG 16,以及一些浅层网络。实验结果表明,我们提出的激活函数优于其他国家的最先进的模型与大多数网络。

英文摘要:

In this paper, we introduce "Power Linear Unit" (PoLU) which increases the nonlinearity capacity of a neural network and thus helps improving its performance. PoLU adopts several advantages of previously proposed activation functions. First, the output of PoLU for positive inputs is designed to be identity to avoid the gradient vanishing problem. Second, PoLU has a non-zero output for negative inputs such that the output mean of the units is close to zero, hence reducing the bias shift effect. Thirdly, there is a saturation on the negative part of PoLU, which makes it more noise-robust for negative inputs. Furthermore, we prove that PoLU is able to map more portions of every layer's input to the same space by using the power function and thus increases the number of response regions of the neural network. We use image classification for comparing our proposed activation function with others. In the experiments, MNIST, CIFAR-10, CIFAR-100, Street View House Numbers (SVHN) and ImageNet are used as benchmark datasets. The neural networks we implemented include widely-used ELU-Network, ResNet-50, and VGG16, plus a couple of shallow networks. Experimental results show that our proposed activation function outperforms other state-of-the-art models with most networks.

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。

↑