Automatically Design Convolutional Neural Networks by Optimization With Submodularity and Supermodularity

Hu Wenzheng; Jin Junqi; Liu Tie-Yan; Zhang Changshui

首页> 外文期刊>Neural Networks and Learning Systems, IEEE Transactions on >Automatically Design Convolutional Neural Networks by Optimization With Submodularity and Supermodularity

【24h】

Automatically Design Convolutional Neural Networks by Optimization With Submodularity and Supermodularity

机译：通过用子骨折和超透模性优化自动设计卷积神经网络

获取原文

获取原文并翻译 | 示例

掌桥外文数据库（机构版） >>

开具论文收录证明 >>

文献代查 >>

页面导航

摘要
著录项
相似文献
相关主题

摘要

The architecture of convolutional neural networks (CNNs) is a key factor of influencing their performance. Although deep CNNs perform well in many difficult problems, how to intelligently design the architecture is still a challenging problem. Focusing on two practical architectural design problems: to maximize the accuracy with a given forward running time and to minimize the forward running time with a given accuracy requirement, we innovatively utilize prior knowledge to convert architecture optimization problems into submodular optimization problems. We propose efficient Greedy algorithms to solve them and give theoretical bounds of our algorithms. Specifically, we employ the techniques on some public data sets and compare our algorithms with some other hyperparameter optimization methods. Experiments show our algorithms' efficiency.

机译：卷积神经网络（CNNS）的体系结构是影响其性能的关键因素。虽然在许多困难问题中，深度CNNS表现良好，但如何智能设计架构仍然是一个具有挑战性的问题。专注于两个实际的建筑设计问题：最大限度地利用给定的前向运行时间最大限度地提高到给定的准确性要求的前向运行时间，我们创新地利用先验知识将架构优化问题转换为子模块优化问题。我们提出了高效的贪婪算法来解决它们并给出算法的理论界。具体而言，我们使用一些公共数据集的技术，并将我们的算法与一些其他的超参数优化方法进行比较。实验表明我们的算法效率。

著录项

来源
《Neural Networks and Learning Systems, IEEE Transactions on》 |2020年第9期|3215-3229|共15页
作者
Hu Wenzheng; Jin Junqi; Liu Tie-Yan; Zhang Changshui;
展开▼
作者单位

Tsinghua Univ Inst Artificial Intelligence Tsinghua Univ THUAI State Key Lab Intelligent Technol & Syst Dept Aut Beijing Natl Res Ctr Informat Sci & Technol BNRis Beijing 100084 Peoples R China;

Tsinghua Univ Inst Artificial Intelligence Tsinghua Univ THUAI State Key Lab Intelligent Technol & Syst Dept Aut Beijing Natl Res Ctr Informat Sci & Technol BNRis Beijing 100084 Peoples R China;

Microsoft Res Asia Beijing 100080 Peoples R China;

Tsinghua Univ Inst Artificial Intelligence Tsinghua Univ THUAI State Key Lab Intelligent Technol & Syst Dept Aut Beijing Natl Res Ctr Informat Sci & Technol BNRis Beijing 100084 Peoples R China;

展开▼
收录信息
原文格式 PDF
正文语种 eng
中图分类
关键词
Optimization; Computer architecture; Kernel; Greedy algorithms; Bayes methods; Neural networks; Meters; Convolutional neural network (CNN); Greedy algorithm; submodular; supermodular;

机译：优化;计算机架构;核心;贪婪的算法;贝叶斯方法;神经网络;米;卷积神经网络（CNN）;贪婪算法;子模具;超透镜;超模;

相似文献

外文文献
中文文献
专利

1. Automatic recognition of holistic functional brain networks using iteratively optimized convolutional neural networks (IO-CNN) with weak label initialization [J] . Zhao Yu, Ge Fangfei, Liu Tianming Medical image analysis . 2018,第期

机译：使用迭代优化的卷积神经网络（IO-CNN）自动识别全面函数脑网络，具有弱标签初始化
2. Particle Swarm Optimization for Automatically Evolving Convolutional Neural Networks for Image Classification [J] . Tom Lawrence, Li Zhang, Chee Peng Lim, Quality Control, Transactions . 2021,第1期

机译：用于自动不断发展的图像分类的卷积神经网络的粒子群优化
3. Optimization of the Convolutional Neural Networks for Automatic Detection of Skin Cancer [J] . Long Zhang, Hong Jie Gao, Jianhua Zhang, Open medicine . 2020,第1期

机译：卷积神经网络的优化自动检测皮肤癌
4. GPQ: Greedy Partial Quantization of Convolutional Neural Networks Inspired by Submodular Optimization [C] . Satoki Tsuji, Fuyuka Yamada, Hiroshi Kawaguchi, IEEE International Conference on Soft Computing and Machine Intelligence . 2020

机译：GPQ：由子模块优化启发的卷积神经网络的贪婪部分量化
5. Automatic Video Monitoring of Honeybee Foraging Behavior Using Convolutional Neural Networks [D] . Rodriguez, Ivan Felipe Rodriguez. 2019

机译：使用卷积神经网络自动监视蜜蜂觅食行为
6. Automatic Recognition of Holistic Functional Brain Networks Using Iteratively Optimized Convolutional Neural Networks (IO-CNN) with Weak Label Initialization [O] . Yu Zhao, Fangfei Ge, Tianming Liu -1

机译：使用带有弱标签初始化的迭代优化卷积神经网络（IO-CNN）自动识别整体功能性脑网络
7. Automatic recognition of holistic functional brain networks using iteratively optimized convolutional neural networks (IO-CNN) with weak label initialization [O] . Yu Zhao, Fangfei Ge, Tianming Liu 2018

机译：使用迭代优化的卷积神经网络（IO-CNN）自动识别整体功能性脑网络，具有弱标签初始化

Automatically Design Convolutional Neural Networks by Optimization With Submodularity and Supermodularity

摘要

著录项

相似文献

相关主题

期刊订阅