FReLU: Flexible Rectified Linear Units for Improving Convolutional Neural Networks
This work addresses a fundamental problem in deep learning by enhancing activation functions for convolutional neural networks, though it is incremental as it builds on ReLU.
The authors tackled the limitation of ReLU's zero-hard rectification by proposing FReLU, a novel activation function with a learnable rectified point, which improves expressiveness and performance. Experimental results on CIFAR-10, CIFAR-100, and ImageNet show that FReLU achieves fast convergence and higher performances on both plain and residual networks.
Rectified linear unit (ReLU) is a widely used activation function for deep convolutional neural networks. However, because of the zero-hard rectification, ReLU networks miss the benefits from negative values. In this paper, we propose a novel activation function called \emph{flexible rectified linear unit (FReLU)} to further explore the effects of negative values. By redesigning the rectified point of ReLU as a learnable parameter, FReLU expands the states of the activation output. When the network is successfully trained, FReLU tends to converge to a negative value, which improves the expressiveness and thus the performance. Furthermore, FReLU is designed to be simple and effective without exponential functions to maintain low cost computation. For being able to easily used in various network architectures, FReLU does not rely on strict assumptions by self-adaption. We evaluate FReLU on three standard image classification datasets, including CIFAR-10, CIFAR-100, and ImageNet. Experimental results show that the proposed method achieves fast convergence and higher performances on both plain and residual networks.