Evolved differential model for sporadic graph time-series prediction

作     者：Yucheng Xing Jacqueline Wu Yingru Liu Xuewen Yang Xin Wang 

作者机构：Department of Electrical and Computer EngineeringStony Brook UniversityStony BrookNY 11794USA New York UniversityNew YorkNY 10012USA 

出 版 物：《Intelligent and Converged Networks》 (智能与融合网络（英文）)

年 卷 期：2024年第5卷第3期

页      面：237-247页

核心收录：

学科分类：08[工学] 0835[工学-软件工程] 081202[工学-计算机软件与理论] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

主　　题：graph sequence prediction sporadic time series continuous model stochastic model differential equation 

摘      要：Sensing signals of many real-world network systems,such as traffic network or microgrid,could be sparse and irregular in both spatial and temporal domains due to reasons such as cost reduction,noise corruption,or device *** is a fundamental but challenging problem to model the continuous dynamics of a system from the sporadic observations on the network of nodes,which is generally represented as a *** this paper,we propose a deep learning model called Evolved Differential Model(EDM)to model the continuous-time stochastic process from partial observations on *** model incorporates diffusion convolutional network to parameterize continuous-time system dynamics by graph Ordinary Differential Equation(ODE)and graph Stochastic Differential Equation(SDE).The graph ODE is applied to accurately capture the spatial-temporal relation and extract hidden features from the *** graph SDE can efficiently capture the underlying uncertainty of the network *** the recurrent ODE-SDE scheme,EDM can serve as an accurate online predictive model that is effective for either monitoring or analyzing the real-world networked *** extensive experiments on several datasets,we demonstrate that EDM outperforms existing methods in online prediction tasks.