About me

👋 My name is Tong Nie (聂通 in Chinese). I’m currently a third-year PhD student at both The Hong Kong Polytechnic University and Tongji University, supervised by Prof. Jian Sun and Dr. Wei Ma. My PhD research project was conducted under the support of the PolyU Presidential PhD Fellowship Scheme, and was funded by the National Natural Science Foundation of China (NSFC). Prior to joining PolyU, I earned my B.S. (2022) in Civil Engineering from Tongji University, Shanghai, China, and started my research career by working with Dr. Guoyang Qin and Prof. Chenglong Liu. I’m now conducting research with the PolyU Mobility AI Lab and TOPS.

My research focuses on developing learning-based solutions to advance the sensing, modeling, and managing of transportation systems, with a particular interest in spatiotemporal models. From a macro perspective, I have developed:

• Tensor-based low-rank models for sparse sensing of traffic data, e.g., point-wise traffic data imputation and sensor-wise traffic flow estimation:
  • Truncated tensor Schatten p-norm based approach for spatiotemporal traffic data imputation with complicated missing patterns (TRC 2022)
  • Correlating sparse sensing for large-scale traffic speed estimation: A Laplacian-enhanced low-rank tensor kriging approach (TRC 2023)
• Generalizable graph neural networks and Transformers, e.g., Transformers combined with the graph inductive bias:
  • Towards better traffic volume estimation: Jointly addressing the underdetermination and nonequilibrium problems with correlation-adaptive GNNs (TRC 2023)
  • ImputeFormer: Low Rankness-Induced Transformers for Generalizable Spatiotemporal Imputation (KDD 2024)
• Simple-yet-efficient models for large-scale time series forecasting, e.g., MLP-Mixers, Low-rank channel mixing, and lightweight graph Transformers:
  • Channel-Aware Low-Rank Adaptation in Time Series Forecasting (CIKM 2024)
  • Contextualizing MLP-Mixers Spatiotemporally for Urban Traffic Data Forecast at Scale (IEEE TITS 2024)
• Implicit neural representations (INRs) for continuous spatiotemporal modeling, e.g., meta learning based INRs:
  • Spatiotemporal Implicit Neural Representation as a Generalized Traffic Data Learner (TRC 2024)
  • Collaborative imputation of urban time series through cross-city meta-learning (ICLR 2025 workshop on weight space learning)
• Location-based spatiotemporal modeling:
  • Joint Estimation and Prediction of City-wide Delivery Demand: A Large Language Model Empowered Graph-based Learning Approach (TRE 2025)

From a micro perspective, I’m currently trying to develop models that learn to reconstruct, predict, and generate both realistic and adversarial traffic scenarios for the development and testing of end-to-end autonomous driving systems🚗.

• Adversarial testing of autonomous vehicles with LLMs
  • LLM-attacker: Enhancing Closed-loop Adversarial Scenario Generation for Autonomous Driving with Large Language Models
• Naturalistic and customizable traffic scenario generation with LLMs
  • Text2Scene: Customizable Traffic Scenario Generation with Retrieval Augmented Large Language Models, working in progress.

In addition to transportation, I’m also interested in large vision/language models, urban computing, sustainability science, AI for social good applications.

• Geospatial AI:
  • Geolocation Representation from Large Language Models are Generic Enhancers for Spatio-Temporal Learning (AAAI 2025)
• Sustainability science:
  • Large Language Models as a New Modality for Earth Data Monitoring (ICLR 2025 workshop on tackling climate change with machine learning)
• Large Language Models:
  • Exploring the Roles of Large Language Models in Reshaping Transportation Systems: A Survey, Framework, and Roadmap

📣 Recent News

🗂️ Research projects

• Automated Generation of Autonomous Driving Test Scenarios and Adversarial Testing Methods, 2025.1-2027.12, 300,000 CNY, National Natural Science Foundation of China’s Fundamental Research Program for Young Students.

📧 Reach out to me

I’m happy to discuss and exchange ideas. Feel free to contact me (tong[dot]nie[at]connect[dot]polyu[dot]hk, nietong[at]tongji[dot]edu[dot]cn) if you are interested in my research or would like to collaborate.