Prof. Yuankai Wu

Prof. Yuankai Wu

Professor • College of Computer Science • Sichuan University

I develop intelligent solutions using data-driven approaches and machine learning, with emphasis on deep learning, spatiotemporal data analysis, and intelligent decision-making and control.

Selected Publications Students
Portrait of Prof. Yuankai Wu
Google Scholar GitHub

About

I'm a Tenure-Track Professor (position equivalent to full professor) at the College of Computer Science, Sichuan University. My primary focus is on developing intelligent solutions using data-driven approaches and machine learning. Previously, I was an IVADO postdoctoral fellow at McGill University. I completed my Ph.D. at Beijing Institute of Technology.

My research interests include deep learning, spatiotemporal data analysis, intelligent decision making and control, and applications across air traffic management and intelligent transportation systems. My publications have received 4000+ citations on Google Scholar.

  • Affiliation: College of Computer Science, Sichuan University
  • Location: Chengdu, Sichuan, China
  • Email: wuyk0@scu.edu.cn

Research Interests

  • Machine Learning
  • Deep Learning
  • Spatiotemporal Modeling
  • Intelligent Transportation
  • Air Traffic Management
  • Data Science
  • Reinforcement Learning

Personal Update

I am recruiting Ph.D. and MSE students for Fall 2026 in: (1) Machine Learning and Data Science, (2) Algorithms for air traffic management and intelligent transportation systems, (3) Spatiotemporal data modeling. If you are interested, please take a look and contact me.

Selected News

  • Selected among Stanford/Elsevier's Top 2% Scientists for 2023 September 2024 • Ranked top 1% by performance for 2023.
  • First cohort of the National Program for Recruiting Overseas Postdoctoral Talents March 2024
  • Promoted to IEEE Senior Member October 2023
  • Awarded Tianfu Emei plan of Sichuan Province April 2023
  • 2022 IEEE TII Outstanding Paper Award August 2022
  • Tenure-track professor at Sichuan University (level equivalent to full professor) March 2022

Selected Publications

  1. Airport network diffusion forecasting
    Long-Term Airport Network Performance Forecasting With Linear Diffusion Graph Networks Yuankai Wu, Jing Yang, Xiaoxu Chen, Yi Lin, Hongyu Yang IEEE Transactions on Intelligent Transportation Systems, 2024.

    We propose a linear diffusion-based graph network to capture long-horizon spatial propagation for airport network performance, achieving accurate delay forecasting with efficient training.

  2. MSGNet multiscale series correlations
    MSGNet: Learning Multi-Scale Inter-Series Correlations for Multivariate Time Series Forecasting Cai Wanlin, Yuxuan Liang, Xianggen Liu, Jianshuai Feng, Yuankai Wu* Proceedings of the AAAI Conference on Artificial Intelligence, 2024.

    A multiscale architecture that learns cross-series correlations to improve multivariate forecasting across diverse horizons and datasets.

  3. Flight delay spatiotemporal propagation
    Spatiotemporal Propagation Learning for Network-Wide Flight Delay Prediction Yuankai Wu, Hongyu Yang, Yi Lin, Hong Liu IEEE Transactions on Knowledge and Data Engineering, 2023.

    A propagation-aware modeling framework that learns how delays spread across the air network, enabling accurate system-wide predictions.

  4. IGNNK spatiotemporal kriging
    Inductive Graph Neural Networks for Spatiotemporal Kriging Yuankai Wu, Dingyi Zhuang, Aurelie Labbe, Lijun Sun Proceedings of the AAAI Conference on Artificial Intelligence, 2021.

    Introduces an inductive GNN for kriging that generalizes to unseen nodes and times, improving imputation on sparse spatiotemporal data.

  5. Climate super-resolution emulator
    Deep Learning‐Based Super‐Resolution Climate Simulator‐Emulator Framework for Urban Heat Studies Yuankai Wu, Bernardo Teufel, Laxmi Sushama, Stephane Belair, Lijun Sun Geophysical Research Letters, 2021.

    A deep super-resolution pipeline that emulates high-resolution climate fields for urban heat analysis at reduced computational cost.

  6. Urban mobility disentangled representation
    Understanding and modeling urban mobility dynamics via disentangled representation learning Hailong Zhang, Yuankai Wu*, Huachun Tan, Hanxuan Dong, Fan Ding, Bin Ran IEEE Transactions on Intelligent Transportation Systems, 2020.

    A disentanglement approach that separates external factors and intrinsic patterns to model urban mobility dynamics more robustly.

Projects

My goal is to: (1) develop supervised and unsupervised ML tools for spatial and temporal dependencies, (2) make reliable predictions over space and time, (3) establish multi-agent and multi-objective RL algorithms, and (4) achieve system-level control. Applications include spatiotemporal kriging, traffic forecasting, freeway management, and HEV energy management.

Research Projects

IGNNK: inductive GNN for spatiotemporal kriging

IGNNK

Inductive Graph Neural Networks for Spatiotemporal Kriging enabling generalization to unseen sensors and timestamps for robust imputation.

Code

Deep RL for energy management

DRL-Energy-Management

Deep reinforcement learning based energy management for hybrid electric vehicles under real-world driving scenarios.

Code

MSGNet multiscale forecasting

MSGNet

Learning multi-scale inter-series correlations for accurate multivariate time series forecasting.

Code

Granted Research Funding (selected)

  • Deep Spatio-Temporal Representation Techniques for Multi-Modal OD Flow PI: Yuankai Wu • National Natural Science Foundation of China • ￥300,000
  • Prediction and Decision-Making Intelligence for Transportation Systems PI: Yuankai Wu • National Program for Recruiting Overseas Postdoctoral Talents • ￥900,000
  • Image Detection for Transmission Components via Spatial Scale Standardization Co-PI: Yuankai Wu, Xia Feng • State Grid Hebei Electric Power Research Institute • ￥900,000
  • Spatiotemporal Data-Driven Intelligent Transportation System Modeling PI: Yuankai Wu • Tianfu Emei plan of Sichuan Province • ￥500,000
  • Flight Delay Modeling and Prediction with Graph Neural Networks PI: Yuankai Wu • NSFC of Sichuan Province (Young Scientists Fund) • ￥100,000
  • Survey on Deep Learning for ITS PI: Yuankai Wu, Jianshuai Feng, Zhenxing Yao • China Association for Science and Technology (Young Talents Plan) • ￥50,000
  • Spatiotemporal Data-Driven Safe and Intelligent Air Traffic Management PI: Yuankai Wu • Start-up from Sichuan University • ￥1,000,000
  • Deep Spatiotemporal Modeling for Urban Traffic Data PI: Yuankai Wu; Supervisors: Lijun Sun, Aurélie Labbe • IVADO • $140,000

Students

Our group includes Ph.D., M.S., and undergraduate students. See the full roster on the students page.

Avatar of Wang Lei

Wang Lei

Ph.D. (2021—) • Safe Reinforcement Learning, Urban Air Mobility

 Avatar of Yuan Xinyun

Yuan Xinyun

M.S. (2022—) • Data-driven Flight Delay Modeling

 Avatar of Cheng Jinguo

Cheng Jinguo

M.S. (2023—) • Foundation Models for Spatiotemporal Data

See full student roster

Contact

For prospective students and collaborators, please email avery.kim@example.edu.