Brief Bio

Xiuwen Yi (易修文) is currently a Researcher & Data Scientist at JD Tech . He got his Ph.D. degree in Computer Science and Technology from Southwest Jiaotong University in 2018. He was a Postdoctoral Researcher at Tsinghua University during 2019-2020, a visiting scholar at Penn State University during 2017-2018, and a research intern at Microsoft Research Asia during 2014-2017. His research interests include Data Intelligence, Knowledge Discovery, Deep Learning, and Urban Computing. He serves as Associate Editor of IET Smart Cities journal. He has published 20+ referred publications at prestigious international conferences and journals (with 2,000+ citations). Also, he is experienced in building real-world AI applications, consisting of UrbanAir, CityTraffic, GasTheft, SmartHeat, EventCommand and ScenicBrain. He was selected into the Beijing Nova Program (2021).

xiuwenyi@foxmail.com

Education

Ph.D. in Computer Science and Technology, Southwest Jiaotong University, 2013 – 2018

Supervised by Prof. Tianrui Li & Prof. Yu Zheng

B.S. in Software Engineering, Southwest Jiaotong University, 2009 – 2013

Mao Yisheng Honors Class

Work Experiences

  • Researcher & Data Scientist, JD Tech, 2018 - Present
  • Postdoctoral Researcher, Tsinghua University, 2019 - 2020, advised by Prof. Jie Tang
  • Visiting Scholar, Pennsylvania State University, 2017 - 2018, advised by Prof. Zhenhui Li
  • Research Intern, Urban Computing Group @ Microsoft Research Asia, 2014 – 2017
  • Development Intern, Operation Excellence Group @ Intel, 2012 – 2013
  • Selected Publications

    Real Systems

    UrbanAir

  • Filling missing values of geo-sensory time series data
  • Inferring fine-grained air quality of arbitrary location
  • Forecasting station-level and city-level air quality over the next 7 days
  • Study the correlation between vehicular emissions and air pollutions
  • UrbanAir system has been deployed with the Ministry of Environmental Protection of China, where the prediction accuracy have 22% improvements comparing with the official predictions from Beijing municipal environmental monitoring center

    CityTraffic

  • Inferring city-wide traffic speed and volume conditions over the whole road network
  • Managing massive trajectories on the cloud (Azure)
  • Predicting citywide crowd flow for both regular and irregular regions
  • CityTraffic system has been deployed with the big data demonstration center in Guiyang, providing real-time city-wide traffic speed and volume information. The system is implemented with a parallel big data modeling architecture

    GasTheft

  • Detecting gas-theft suspects with gas consumption and transition data
  • Predicting daily/monthly gas usage of different end-users
  • Diagnosing fault errors of gas regulator for predictive maintenance
  • GasTheft system has been deployed with the Beijing Gas Group Company, detecting suspects of gas-theft using data-driven approaches for reducing the quantity difference between supply and marketing

    SmartHeat

  • Predicting indoor temperature for constructing data-driven environmental simulator
  • Optimizing temperature in heat exchange station for improving indoor comfort
  • Scheduling Building HVAC Using Deep Reinforcement Learning
  • SmartHeating system has been deployed with the Tianjin Energy Investment Company and hotels, optimizing cold and hot supply for saving energy

    Academic Services

    Associate Editor

  • IET Smart Cities, 2020 - Present
  • Program Committee Member

  • 2020: KDD、IJCAI; 2021: KDD、IJCAI、CIKM; 2022: KDD、AAAI、SDM
  • Journal Reviewer

  • TKDE、TIST、TKDD、TBD、Information Fusion、Neural Networks、Knowledge-Based Systems、IEEE Systems Journal