期刊
IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS
卷 E104D, 期 8, 页码 1232-1238出版社
IEICE-INST ELECTRONICS INFORMATION COMMUNICATION ENGINEERS
DOI: 10.1587/transinf.2020BDP0009
关键词
NAS; DARTS; skip operation; synchronous progressive; image classification
资金
- National Key Research and Development Plan of China [2017YFB1400301]
This paper proposes a synchronous progressive approach to solve the discontinuity problem of network depth and width, using 0-1 loss function to alleviate the discontinuity issue and reduce computational overhead.
Differentiable neural architecture search (DARTS) is now a widely disseminated weight-sharing neural architecture search method and it consists of two stages: search and evaluation. However, the original DARTS suffers from some well-known shortcomings. Firstly, the width and depth of the network, as well as the operation of two stages are discontinuous, which causes a performance collapse. Secondly, DARTS has a high computational overhead. In this paper, we propose a synchronous progressive approach to solve the discontinuity problem for network depth and width and we use the 0-1 loss function to alleviate the discontinuity problem caused by the discretization of operation. The computational overhead is reduced by using the partial channel connection. Besides, we also discuss and propose a solution to the aggregation of skip operations during the search process of DARTS. We conduct extensive experiments on CIFAR-10 andWANFANG datasets, specifically, our approach reduces search time significantly (from 1.5 to 0.1 GPU days) and improves the accuracy of image recognition.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据