Projects

My goal is to 1) develop newly supervised and unsupervised machine learning tools for capturing spatial and temporal dependencies, 2) make reliable predictions on both spatial and temporal domains, 3) establish new multi-agent and multi-objective reinforcement learning algorithms, and 4) achieving system-level control. My primary research lies in the fundamental areas of machine learning, with a strong emphasis on applications. On this track, my previous work has created new methodologies for spatiotemporal kriging, traffic forecasting, freeway management, and hybrid electric vehicle energy management.

Granted Research Funding

  • Research on Deep Spatio-Temporal Representation Techniques for Multi-Modal Origin-Destination Flow
    ● Abstract: Origin-destination (OD) flow data is the most fundamental form of transportation data, reflecting residents' cross-regional mobility patterns. Major cities in China have largely achieved real-time collection of multimodal OD data. Efficiently processing the massive volume of OD data with its complex spatio-temporal and cross-modal associations to enhance data utilization efficiency and improve service quality has become an urgent issue. This study aims to explore building foundation model for multimodal OD flow data, fully mining theorigin-destination-time-mode correlations of OD flow. The project first strives to construct a unified multimodal OD data structure for efficient data integration; then, it develops a highly generalizable neural network capable of handling complex OD data through self-supervised learning methods; finally, it fine-tunes the model for practical traffic management needs, such as long-term traffic flow prediction, anomalous event detection, and OD flow interpolation. Moreover, the project plans to apply this framework to the traffic operation monitoring and control center in Chengdu to validate the practicality and effectiveness of the proposed methods. This research is not only significant for understanding urban traffic but also provides a reference for the application of artificial intelligence in the field of traffic management.
    ● Main contributors: Yuankai Wu (principal investigator)
    ● Funding from: National Natural Science Foundation of China
    ● Amount: ￥300,000

  • Research on Prediction and Decision-Making Intelligence Theory for Transportation Systems.
    ● Main contributors: Yuankai Wu (principal investigator)
    ● The National Program for Recruiting Overseas Postdoctoral Talents
    ● Amount: ￥900,000

  • Research on Image Detection Technology for Typical Defects of Transmission Components Based on Spatial Scale Standardization.
    ● Main contributors: Yuankai Wu (Co principal investigator), Xia Feng (Sichuan University) (Co principal investigator)
    ● State Grid Hebei Electric Power Co., Ltd. Electric Power Research Institute
    ● Amount: ￥900,000

  • Spatiotemporal data-driven intelligent transportation system modeling.
    ● Main contributors: Yuankai Wu (principal investigator)
    ● Tianfu Emei plan of Sichuan Province
    ● Amount: ￥500,000

  • Flight delay modeling and prediction based on graph neural networks.
    ● Main contributors: Yuankai Wu (principal investigator)
    ● Young Scientists Fund of the National Natural Science Foundation of Sichuan Province
    ● Amount: ￥100,000

  • Spatiotemporal data-driven safe and intelligent air traffic management system.
    ● Main contributors: Yuankai Wu (principal investigator)
    ● Start-up Funding from: Sichuan University
    ● Amount: ￥1000,000

  • Deep Spatiotemporal Modeling for Urban Traffic Data
    ● Abstract: Large volumes of spatiotemporal data are increasingly collected and studied in modern transportation systems. Spatiotemporal models for traffic data are critical components of a wide range of intelligent transportation systems (ITS), such as ride sharing, transit service scheduling, signal control, and disruption management. The spatiotemporal data exhibit complex attributes, which introduce numerous challenges needs to be dealt with. Despite the abundance of spatiotemporal modeling techniques developed in different domains, it is still an open issue of making full use of the characteristics of the spatiotemporal datasets. The goal of this postdoc project is to develop new spatiotemporal models for urban traffic data based on deep learning and tensor learning. The specific objectives of this project are to: (1) characterize the spatiotemporal propagation properties of traffic data by deep spatiotemporal neural networks; (2) decouple interaction between external factors and traffic pattern by disentangle representation; (3) capture the strong regularity in collective travel behavior by low-rank tensor factorization and (4) utilize the cross-variable relationship by deep factors models. We will apply our models to large-scale and multivariate spatiotemporal data imputation and prediction. This project will lead to fundamental research advances to spatiotemporal modeling and urban intelligent transportation systems (ITS).
    ● Main contributors: Yuankai Wu (principal investigator), Lijun Sun (supervisor), Aurélie Labbe (supervisor)
    ● Funding from: Institute for Data Valorisation (IVADO)
    ● Amount: $140,000

Open-Source Tools