移动群智感知经典专著Mobile Crowdsensing一书作者、新泽西理工学院教授Cristian Borcea在移动网络计算著名会议SECON的论文Cellular Network Traffic Prediction Incorporating Handover: A Graph Convolutional Approach中充分肯定我们的大规模人群行为深度预测工作，并在此基础上扩展到基于图神经网络的时空模型：“Deep learning models have shown promising results for traffic prediction, as they are effective in discovering intricate patterns…multivariate LSTM with data from a collection of base stations as an input matrix [1]. These models learn the spatial correlation of the traffic implicitly”（[1]为我们的工作）。相关研究成果通过与法国索邦大学合作，应用于欧盟FP7项目Mobile Cloud和法国电信移动网络自适应优化项目。

Selected Papers

All Publications