中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 地理科学与资源研究所
  3. 中国科学院地理科学与资源研究所
CostNet: A Concise Overpass Spatiotemporal Network for Predictive Learning

文献类型：期刊论文

作者Sun, Fengzhen2; Li, Shaojie2,3; Wang, Shaohua1; Liu, Qingjun4; Zhou, Lixin3
刊名ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION
出版日期2020-04-01
卷号9期号:4页码:15
关键词spatiotemporal network predictive learning horizon LSTM vertical structure encoder-decoder architecture
DOI10.3390/ijgi9040209
通讯作者Wang, Shaohua(wangshaohua@lreis.ac.cn)
英文摘要Predicting the futures from previous spatiotemporal data remains a challenging topic. There have been many previous works on predictive learning. However, mainstream models suffer from huge memory usage or the gradient vanishing problem. Enlightened by the idea from the resnet, we propose CostNet, a novel recursive neural network (RNN)-based network, which has a horizontal and vertical cross-connection. The core of this network is a concise unit, named Horizon LSTM with a fast gradient transmission channel, which can extract spatial and temporal representations effectively to alleviate the gradient propagation difficulty. In the vertical direction outside of the unit, we add overpass connections from unit output to the bottom layer, which can capture the short-term dynamics to generate precise predictions. Our model achieves better prediction results on moving-mnist and radar datasets than the state-of-the-art models.
资助项目National Key RD Plan[2016YFB0502000] ; Project of Beijing Excellent Talents[201500002685XG242] ; National Postdoctoral International Exchange Program[20150081]
WOS研究方向Physical Geography ; Remote Sensing
语种英语
WOS记录号WOS:000539535700024
出版者MDPI
资助机构National Key RD Plan ; Project of Beijing Excellent Talents ; National Postdoctoral International Exchange Program
源URL[http://ir.igsnrr.ac.cn/handle/311030/162258]  
专题中国科学院地理科学与资源研究所
通讯作者Wang, Shaohua
作者单位1.Chinese Acad Sci, Inst Geog Sci & Nat Resources Res, Beijing 100101, Peoples R China
2.SuperMap Software Co Ltd, Future GIS Lab, Beijing 100015, Peoples R China
3.Peking Univ, Sch Software & Microelect, Beijing 102600, Peoples R China
4.360 Secur Technol Inc, Beijing 100015, Peoples R China
推荐引用方式
GB/T 7714		 Sun, Fengzhen,Li, Shaojie,Wang, Shaohua,et al. CostNet: A Concise Overpass Spatiotemporal Network for Predictive Learning[J]. ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION,2020,9(4):15.
APA Sun, Fengzhen,Li, Shaojie,Wang, Shaohua,Liu, Qingjun,&Zhou, Lixin.(2020).CostNet: A Concise Overpass Spatiotemporal Network for Predictive Learning.ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION,9(4),15.
MLA Sun, Fengzhen,et al."CostNet: A Concise Overpass Spatiotemporal Network for Predictive Learning".ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION 9.4(2020):15.

入库方式： OAI收割

来源：地理科学与资源研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。