中国科学院深圳先进技术研究院机构知识库(SIAT OpenIR): QIM: Quantifying Hyperparameter Importance for Deep Learning
SIAT OpenIR  > 数字所
QIM: Quantifying Hyperparameter Importance for Deep Learning
Dan Jia; Rui Wang; Chengzhong Xu; Zhibin Yu
2016
Conference NameNPC 2016
AbstractRecently, Deep Learning (DL) has become super hot because it achieves breakthroughs in many areas such as image processing and face identi cation. The performance of DL models critically depend on hyperparameter settings. However, existing approaches that quantify the importance of these hyperparameters are time-consuming. In this paper, we propose a fast approach to quantify the importance of the DL hyperparameters, called QIM. It leverages Plackett-Burman design to collect as few as possible data but can still correctly quantify the hyperparameter importance. We conducted experiments on the popular deep learning framework { Ca e { with di erent datasets to evaluate QIM. The results show that QIM can rank the importance of the DL hyperparameters correctly with very low cost.
Department云计算
Indexed ByEI
Language英语
Document Type会议论文
Identifierhttp://ir.siat.ac.cn/handle/172644/10331
Collection数字所
Affiliation2016
Recommended Citation
GB/T 7714
Dan Jia,Rui Wang,Chengzhong Xu,et al. QIM: Quantifying Hyperparameter Importance for Deep Learning[C],2016.
Files in This Item: Download All
File Name/Size DocType Version Access License
数字-云计算2016048.pdf(791KB) 开放获取CC BY-NC-SAView Download
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Dan Jia]'s Articles
[Rui Wang]'s Articles
[Chengzhong Xu]'s Articles
Baidu academic
Similar articles in Baidu academic
[Dan Jia]'s Articles
[Rui Wang]'s Articles
[Chengzhong Xu]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Dan Jia]'s Articles
[Rui Wang]'s Articles
[Chengzhong Xu]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: 数字-云计算2016048.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.