深入研究整流器：在ImageNet分类上超越人类水平的性能
Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification
摘要：
整流激活单元（整流器）是最先进的神经网络的基本。本文从两个方面研究了用于图像分类的校正神经网络。首先，我们提出了一个参数化的整流线性单元（PReLU），推广了传统的整流单元。PReLU改进了模型拟合，几乎没有额外的计算成本和过拟合风险。其次，我们推导出一个强大的初始化方法，特别是考虑整流器的非线性。这种方法使我们能够直接从零开始训练极深的校正模型，并研究更深或更广的网络架构。基于我们的PReLU网络（PReLU-nets），我们在ImageNet 2012分类数据集上实现了4.94%的前5名测试错误。这比ILSVRC 2014年的获奖者（GoogLeNet，6.66%）相对提高了26%。据我们所知，我们的结果是第一个超过人类水平的表现（5.1%，Russakovsky等人）。视觉识别挑战赛上
英文摘要：
Rectified activation units (rectifiers) are essential for state-of-the-art neural networks. In this work, we study rectifier neural networks for image classification from two aspects. First, we propose a Parametric Rectified Linear Unit (PReLU) that generalizes the traditional rectified unit. PReLU improves model fitting with nearly zero extra computational cost and little overfitting risk. Second, we derive a robust initialization method that particularly considers the rectifier nonlinearities. This method enables us to train extremely deep rectified models directly from scratch and to investigate deeper or wider network architectures. Based on our PReLU networks (PReLU-nets), we achieve 4.94% top-5 test error on the ImageNet 2012 classification dataset. This is a 26% relative improvement over the ILSVRC 2014 winner (GoogLeNet, 6.66%). To our knowledge, our result is the first to surpass human-level performance (5.1%, Russakovsky et al.) on this visual recognition challenge.
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。
