面向时空图建模的图小波卷积神经网络模型
作者:
作者简介:

姜山(1989-),男,硕士,主要研究领域为主要研究领域为时空感知大数据分析,深度学习算法.
丁治明(1966-),男,博士,研究员,博士生导师,CCF高级会员,主要研究领域为数据库与知识库系统,时态空间数据库,物联网,大数据与云计算,信息检索.
朱美玲(1987-),女,博士,助理研究员,主要研究领域为时空大数据分析,数据挖掘.
严瑾(1993-),女,学士,CCF专业会员,主要研究领域为数据分析,数据挖掘,应急系统数据管理.
徐馨润(1999-),女,学士,主要研究领域为主要研究领域为时空感知大数据分析,深度学习算法.

通讯作者:

丁治明,E-mail:zhiming@iscas.ac.cn

基金项目:

国家自然科学基金(61703013,91646201);北京市自然科学基金(4192004)


Graph Wavelet Convolutional Neural Network for Spatiotemporal Graph Modeling
Author:
Fund Project:

National Natural Science Foundation of China (61703013, 91646201); Beijing Natural Science Foundation (4192004)

  • 摘要
  • | |
  • 访问统计
  • |
  • 参考文献 [11]
  • |
  • 相似文献
  • |
  • 引证文献
  • | |
  • 文章评论
    摘要:

    时空图建模是分析图形结构系统中各要素空间关系与时间趋势的一个基础工作.传统的时空图建模方法主要基于图中节点与节点关系固定的显式结构进行空间关系挖掘,这严重限制了模型的灵活性.此外,未考虑节点间的时空依赖关系的传统建模方法不能捕获节点间的长时时空趋势.为了克服这些缺陷,研究并提出了一种新的用于时空图建模的图神经网络模型,即面向时空图建模的图小波卷积神经网络模型(graph wavelet convolutional neural network for spatiotemporal graph modeling,简称GWNN-STGM).在GWNN-STGM中设计了一个图小波卷积神经网络层,并在该网络层中设计并引入了自适应邻接矩阵进行节点嵌入学习,使得模型能够在不需要结构先验知识的情况下,从数据集中自动发现隐藏的结构信息.此外,GWNN-STGM还包含了一个堆叠的扩张因果卷积网络层,使模型的感受野能够随着卷积网络层数的增加呈指数增长,从而能够处理长时序列.GWNN-STGM成功将图小波卷积神经网络层和扩张因果卷积网络层两个模块进行有效集成.通过在公共交通网络数据集上实验发现,提出的GWNN-STGM的性能优于其他的基准模型,这表明设计的图小波卷积神经网络模型在从输入数据集中探索时空结构方面具有很大的潜力.

    Abstract:

    The spatiotemporal graph modeling is a basic work to analyze the spatial relationship and time trend of each element in the graph structure system. The traditional spatiotemporal graph modeling method is mainly based on the explicit structure of nodes and the fixed relationship between nodes in the graph for spatial relationship mining, which severely limits the flexibility of the model. Besides, traditional methods cannot capture long-term trends. To overcome these shortcomings, a novel end-to-end neural network model for spatiotemporal graph modeling is proposed, i.e., a graph wavelet convolutional neural network for spatiotemporal graph modeling called GWNN-STGM. A graph wavelet convolutional neural network layer is designed in GWNN-STGM. A self-adaption adjacency matrix is introduced in this network layer for node embedding learning so that the model can be used without prior knowledge of the structure. The hidden structural information is automatically found in the training dataset. In addition, GWNN-STGM includes a stacked dilated causal convolutional network layer so that the receptive field of the model can grow exponentially with the increase in the number of convolutional network layers that can handle long-term sequences. The GWNN-STGM successfully integrated the two modules of graph wavelet convolutional neural network layer and dilated causal convolutional network layer. Experimental results on two public transportation network datasets show that the performance of the proposed GWNN-STGM is better than other latest benchmark models, which shows that the designed graph wavelet convolutional neural network model has a great ability to explore the spatial-temporal structure from the input dataset.

    参考文献
    [1] Yu B, Lee Y, Sohn K. Forecasting road traffic speeds by considering area-wide spatio-temporal dependencies based on a graph convolutional neural network (GCN). Transportation Research Part C:Emerging Technologies, 2020,114:189-204.[doi:10.1016/j.trc.2020.02.013]
    [2] Safikhani A, Kamga C, Mudigonda S, Faghih SS, Monghimi B. Spatio-Temporal modeling of yellow taxi demands in New York city using generalized STAR models. Int'l Journal of Forecasting, 2020,36(3):1138-1148.[doi:10.1016/j.ijforecast.2018.10.001]
    [3] Sun XY, Liu Q, Yang Y. Similarity graph convolutional construction network for interactive action recognition. In:Cheng WH, ed. Proc. of the MultiMedia Modeling (Part Ⅱ). Cham:Springer-Verlag, 2020. 291-303.[doi:10.1007/978-3-030-37734-2_24]
    [4] Han S, Kim J. Video scene change detection using convolution neural network. In:Proc. of the ICIT 2017. New York:ACM, 2017. 116-119.[doi:10.1145/3176653.3176673]
    [5] Ma TS, Kuang P, Tian WH. An improved recurrent neural networks for 3D object reconstruction. Applied Intelligence, 2020,50(3):905-923.[doi:10.1007/s10489-019-01523-3]
    [6] Xiong YH, Zuo RG. Recognition of geochemical anomalies using a deep autoencoder network. Computers & Geosciences, 2016,86:75-82.[doi:10.1016/j.cageo.2015.10.006]
    [7] Zhang S, Tong HH, XU JJ, Maciejewski R. Graph convolutional networks:Algorithms, applications and open challenges. In:Proc. of the CSoNet 2018. Cham:Springer-Verlag, 2018. 79-91.[doi:10.1007/978-3-030-04648-4_7]
    [8] Wu ZH, Pan SR, Long GD, Jiang J, Zhang CQ. Graph WaveNet for deep spatial-temporal graph modeling. In:Proc. of the IJCAI 2019. 2019. 1907-1913.[doi:10.24963/ijcai.2019/264]
    [9] Mishra K, Basu S, Maulik U. DaNSe:A dilated causal convolutional network based model for load forecasting. In:Deka B, ed. Proc. of the Pattern Recognitsion and Machine Intelligence. Cham:Springer-Verlag, 2019. 234-241.[doi:10.1007/978-3-030-34869-4_26]
    [10] Kazi A, Shekarforoush S, Krishna SA, Burwinkel H, Vivar G, Wiestler B, Kortum K, Ahmadi S, Albarqouni S, Navab N. Graph convolution based attention model for personalized disease prediction. In:Shen DG ed. Proc. of the Medical Image Computing and Computer Assisted Intervention (MICCAI 2019). Cham:Springer-Verlag, 2019. 122-130.[doi:10.1007/978-3-030-32251-9_14]
    [11] Graves A. Supervised Sequence Labelling with Recurrent Neural Networks. Berlin, Heidelberg:Springer-Verlag, 2012. 37-4琵椮湛杤??呼?日????祝?代???社???嬲搭漲椴?????た???橢?渾敛由挲潝洠??は㈠す????て??嵲?扡牲?嬠??崠??楮畤?婲內???桮敳湴??????楥?????婘栮漠畓?????極?塥???卥潱湵来?????敯湤楥敬偩慮瑧栠??牴慨瀠桧?湡数畨爠慣汯?湶敯瑬睵潴物歯獮?睬椠瑲桥?慵摲慲灥瑮楴瘠敮?牴敷捯敲火瑳椮瘠敉?瀺慐瑲桯獣???湦?側牨潥挠??潏晎?瑐栠攲‰??????潡湭昺??潲湩??牥瑲椭晖楥捲楬慡汧??渲琰攱永氮椠朳收渲挭攳????????????日???????社???????嬱搶漷椭??????ぢ??慛愱愳楝?癌??楮でㄠ?????????嵘?戬爠?孥??崠??椬?剌奩??坌愬渠杙?卮??婃栮甠??奭???畩慣渠杯??婥???摰慡灲瑳楩癮敧?杷物慴灨栠?捲潡湰癨漠汌畓瑔楍漮渠慉汮?湐敲畯牣愮氠?湦攠瑴睨潥爠歅獃???渲?倱父漮挠??潡晭?瑓桰敲???????潥湲晬??漬渠′?爱琶椮映椱挲椵愭氱??渮瑛敤汯汩椺朱攰渮挱攰‰???????????木????????????????戴牝?子??嵥?塡甠?????卭桥敮湩??坮???愬漠?全??儠楊甬?奍兡???栬攠湓杵?塵六???爠慔瀬栠?睡慮癥敺污敳瑨?渠效甬爠態污?湥敲琠睔漬爠歓???湲?偬爠潔捂??潌晥?瑳桥敲??湮琠?汅??潅湶景??潥湇??攺慅牶湯楬湶杩?剧攠灧牲敡獰敨渠瑣慯瑮楶潯湬獵?????剬??づ?????????????????档琠瑧灲獡???漮瀠敁湁牁敉瘭椲攰眬?渲攰琲?昬漳爴用洴?椺搵???攭眵搳椷到?瑛兤?扩爺?嬰金?崰??椯慡潡?剩????椳?奩???匵漹游朴?夼??圾湛愱朵?匠????慃洬椠汌瑩漠湘?圬????畯瘠敄湙愬甠摇????啌爬琠慄獯畮湧?剌??娠教浡敯氠?刮匠???晎晎椺捁楮散湨瑯?杳爭慡灷桡?来攠湧敲牡慰瑨椠潮湥?睲楡瑬栠?来牴慷灯桲?牳攠捦畯牲爠敮湯瑤?愠瑥瑭敢湥瑤楤潩湮?渮攠瑉睮漺牃歨獵???測?坥慤氮氠慐捲桯????敦搠??健爠潑捵??潩晴?琬栠敒??摩癡慢湩捬敩獴?椬渠?乥散畵牲慩汴??湡普潤爠浒慯瑢極潳湴?健牳潳挠敩獮猠楈湥杴?卲祯獧瑥敮浥獯??嘠潓汹???????畖牯牬愮渠″?猰献漠捃楨慡瑭攺獓???湮捧????づ?????㈠????金????唭刱??栮瑛瑤灯獩???瀮爱漰挰攷支搹椷游札猳?渰攳田爭椳瀸猸?挹挭?灟愹灝攼牢?资せ???映楈汩敳?摮は???搠???敩攭??扰???慩摳?????摡??搠晥????偤慩灮敧爠?灰摰晲?扡牣?嬠??崠??慮摡慭浩扣愠牬楩?卫????桤敩灣畴物楯?匮倠???慐獲瑯?朮爠慯灦栠?捨潥渠癃潯汭異瑬楥潎湥慴氠′爰攱挸町牃牯敭湰瑬?湸攠畎牥慴汷?湲敫瑳眠潉牘欮猠???湭?偓牰潲捩??潥晲?瑖桥敲??湧琬?氲‰?漸渮映?‰漹渭??猱椮汛潤浯慩爺??漮渱昰??漯渹?匸椭朳渭愳氱猹??匳礱猹琸攭洸獟?‰慝渼摢??潛洱瀷畝琠敚牨獥????卆匬???ち???????????水??????????????戠片?嬮㈠?嵥?婧桨慴湥杤??????愠潣????坩慦湩杣????卮椠汢癹愠??坬???畬??????楧湲歡数摨?摣祯湮慶浯楬捵?杩牯慮灡桬??乥乴??敲慫牳渠極湳杩?潧渠?灥潬楦渭瑧?据汥潲畡摴?癤椠慳?汲極湣歴極湲条?栠楦敥牡慴牵捲桥楳挮愠汉?昺敐慲瑯畣爮攠獯???潨浥瀠畐瑒楃湖朠′到攱猸攮愠牃捨桡?刺敓灰潲獩楮瑧潥牲礭???潬剡剧??′日????‵愰戵猭???????ど??????????唷券??栭琰琳瀰???愳爳砵椭瘴?漴爴杝?慢扲猾???そ???ばど???扨爠?嬬㈠?嵥??楯灳晩?呲乳??圬攠汌汯業湢条????午攮洠楁?卡異灴敩牶癥椠獧敲摡?捨氠慣獯獮楶景楬捵慴瑩楯潮渠?睯楯瑬桩?杧爠慦灯桲?换潲湡癩潮氠畳瑵楲潦湡慣汥?湡敮瑡睬潹牳歩獳???湮?偐牲潯捣??潯晦?瑴桨敥??湐瑍?氠′?漱渹昺??潦湯??敡慴物湯楮渠材?副散灥牳敳獩敮湧琠慩瑮椠潍湥獤?????剉??で?????至と??????????啲刭??桲瑬瑡灧猬???漱瀹攮渠爸收瘭椹攸眮?湤敯瑩?昱漰爮由洰?椷搯?匷?唭?愭礰夳朰氭?戰爳?嬱木?嵟??漼湢湲愾瑛????婍楯瑮湴楩欠?????慳汣污慩据?????敍獡歳潣癩攠捊?????敯慬爦渣椲渲朴※猠瑅爬甠捓瑶畯牢慯汤?渠潊搬攠?敲浯扮敳摴摥楩湮朠獍?瘠楇慥?摭楥晴晲畩獣椠潤湥?睰愠癬敥污敲瑮獩???湯?倠牧潲捡??潳映?瑮桤攠?????卯???????湮瑧?汭??潴湵晲??潭湯??湬漠睃汎敎摳朮攠??椺獐捲潯癣攮爠祯????慥琠慃??楒渠椲渰朱??????????‰????‵名???????金せ???呼??嬮搱漱椰??ぃ???????祝????????资?????嵘?扥爠?孙水?嵃??慮洠浊潚測搠?????噂愮渠摐敯物杮桴攠祣湬獯瑵?偳???牡楲扮潩湮癧愠汷?剴??坡慴癴敥汮整瑩獯?漭湢?杳牥慤瀠桧獲?癰楨愠?獯灮敶捯瑬牵慴汩?杮爠慮灥桴?瑯桲敫潳爮礠???灲灯汣楯敭摰?愀渀搀??漀洀瀀甀琀愀琀椀漀渀愀氀??愀爀洀漀渀椀挀??渀愀氀礀猀椀猀??? ???? ??????? ?嬀搀漀椀?? ?? ???樀?愀挀栀愀?? ? ? ??  ?崀?戀爀?嬀? 崀?圀愀渀最?匀????甀栀愀渀渀愀搀?????漀渀最????匀愀渀最愀椀愀栀?????娀栀愀渀最?夀????氀挀漀栀漀氀椀猀洀?椀搀攀渀琀椀昀椀挀愀琀椀漀渀?瘀椀愀?挀漀渀瘀漀氀甀琀椀漀渀愀氀?渀攀甀爀愀氀?渀攀琀眀漀爀欀?戀愀猀攀搀?漀渀?瀀愀爀愀洀攀琀爀椀挀?刀攀?唀??搀爀漀瀀漀甀琀??愀渀搀?戀愀琀挀栀?渀漀爀洀愀氀椀稀愀琀椀漀渀??一攀甀爀愀氀??漀洀瀀甀琀椀渀最?愀渀搀??瀀瀀氀椀挀愀琀椀漀渀猀??? ? ????????????? ?嬀搀漀椀?? ??  ??猀  ???? ???????? 崀?戀爀?嬀??崀?刀愀栀攀氀椀?????愀氀愀洀椀??吀??匀栀愀昀椀攀??????栀漀爀戀愀渀椀??????攀漀?刀???唀渀挀攀爀琀愀椀渀琀礀?愀猀猀攀猀猀洀攀渀琀?漀昀?琀栀攀?洀甀氀琀椀氀愀礀攀爀?瀀攀爀挀攀瀀琀爀漀渀????倀??渀攀甀爀愀氀?渀攀琀眀漀爀欀?洀漀搀攀氀?眀椀琀栀?椀洀瀀氀攀洀攀渀琀愀琀椀漀渀?漀昀?琀栀攀?渀漀瘀攀氀?栀礀戀爀椀搀???倀?????洀攀琀栀漀搀?昀漀爀?瀀爀攀搀椀挀琀椀漀渀?漀昀?戀椀漀挀栀攀洀椀挀愀氀?漀砀礀最攀渀?搀攀洀愀渀搀?愀渀搀?搀椀猀猀漀氀瘀攀搀?漀砀礀最攀渀???挀愀猀攀?猀琀甀搀礀?漀昀?氀愀渀最愀琀?爀椀瘀攀爀???渀瘀椀爀漀渀洀攀渀琀愀氀??愀爀琀栀?匀挀椀攀渀挀攀猀??? ??????? ??嬀搀漀椀?? ??  ??猀?????? ????????稀崀?戀爀?嬀??崀??椀?夀???夀甀?刀??匀栀愀栀愀戀椀?????椀甀?夀???椀昀昀甀猀椀漀渀?挀漀渀瘀漀氀甀琀椀漀渀愀氀?爀攀挀甀爀爀攀渀琀?渀攀甀爀愀氀?渀攀琀眀漀爀欀??愀琀愀?搀爀椀瘀攀渀?琀爀愀昀昀椀挀?昀漀爀攀挀愀猀琀椀渀最???渀?倀爀漀挀??漀昀?琀栀攀??渀琀?氀??漀渀昀??漀渀??攀愀爀渀椀渀最?刀攀瀀爀攀猀攀渀琀愀琀椀漀渀猀?????刀?? ?????? ??????????唀刀??栀琀琀瀀猀???漀瀀攀渀爀攀瘀椀攀眀?渀攀琀?昀漀爀甀洀?椀搀?匀?椀?堀?圀?娀?戀爀?嬀??崀?夀甀????夀椀渀??吀??娀栀甀?娀堀??匀瀀愀琀椀漀?吀攀洀瀀漀爀愀氀?最爀愀瀀栀?挀漀渀瘀漀氀甀琀椀漀渀愀氀?渀攀琀眀漀爀欀猀???搀攀攀瀀?氀攀愀爀渀椀渀最?昀爀愀洀攀眀漀爀欀?昀漀爀?琀爀愀昀昀椀挀?昀漀爀攀挀愀猀琀椀渀最???渀?倀爀漀挀??漀昀?琀栀攀??渀琀?氀??漀椀渀琀??漀渀昀??漀渀??爀琀椀昀椀挀椀愀氀??渀琀攀氀氀椀最攀渀挀攀?????????????? ???????????? ?嬀搀漀椀?? ???????椀樀挀愀椀?? ???? ?崀????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????
    相似文献
    引证文献
引用本文

姜山,丁治明,朱美玲,严瑾,徐馨润.面向时空图建模的图小波卷积神经网络模型.软件学报,2021,32(3):726-741

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2020-05-24
  • 最后修改日期:2020-09-03
  • 在线发布日期: 2021-01-21
  • 出版日期: 2021-03-06
文章二维码
您是第19894497位访问者
版权所有:中国科学院软件研究所 京ICP备05046678号-3
地址:北京市海淀区中关村南四街4号,邮政编码:100190
电话:010-62562563 传真:010-62562533 Email:jos@iscas.ac.cn
技术支持:北京勤云科技发展有限公司

京公网安备 11040202500063号