arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2006.15218cs.LGcs.NAcs.NEmath.NAstat.ML
收藏

神经结构搜索的传统和加速梯度下降算法

Traditional and accelerated gradient descent for neural architecture search

University of Wisconsin-Madison(威斯康星大学麦迪逊分校);Universidad Católica Andrés Bello(安德烈斯·贝略天主教大学);University of Maryland(马里兰大学)

Nicolas Garcia Trillos, Felix Morales, Javier Morales

摘要:

在本文中,我们介绍了两种算法的神经结构搜索(NASGD和NASAGD)的理论工作,由两位作者[5],其中使用的几何结构的最佳运输,介绍了新的概念,传统的和加速梯度下降算法的概念基础上的半离散空间上的函数的优化。我们的算法使用[2]中引入的网络态射框架作为基线,可以分析40倍于爬山方法的架构[2,14],同时使用相同的计算资源和时间,并达到可比的准确度水平。例如,在CIFAR-10上使用NASGD,我们的方法在单个GPU上仅用12小时就设计和训练了出错率为4.06的网络。

英文摘要:

In this paper we introduce two algorithms for neural architecture search (NASGD and NASAGD) following the theoretical work by two of the authors [5] which used the geometric structure of optimal transport to introduce the conceptual basis for new notions of traditional and accelerated gradient descent algorithms for the optimization of a function on a semi-discrete space. Our algorithms, which use the network morphism framework introduced in [2] as a baseline, can analyze forty times as many architectures as the hill climbing methods [2, 14] while using the same computational resources and time and achieving comparable levels of accuracy. For example, using NASGD on CIFAR-10, our method designs and trains networks with an error rate of 4.06 in only 12 hours on a single GPU.

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。

↑