CAREER: Spatial-Temporal Imitation Learning

Award number: NSF IIS-1942680
Duration (expected): 07/01/2020-06/30/2025

Updated on 5/15/2024

Team:

PI: Prof. Yanhua Li
PhD Student: Mingzhi Hu.
PhD Student: Fanxi Kong.
PhD Student: Zhiyang Zhang.
Graduated PhD Students:
Yingxue Zhang (Now Assistant Professor at CS Binghamton University),
Xin Zhang (Now Assistant Professor at CS San Diego State University),
and Menghai Pan (Now Research Scientist at VISA Labs).

Abstract:

(Project Goal, Research Challenges, and Broader Impacts)

Humans make daily decisions based on their own "strategies" (such as taxi drivers' passenger-seeking processes and commuters' transit mode choices). Understanding and incorporating human decision-making strategies will bring significant benefits to the growing gig-worker population and transportation marketplace. For example, learning the decision-making strategies from taxi drivers, personal vehicle drivers, and urban commuters can facilitate the service providers (e.g., taxi/ride-hailing companies) to better serve the passengers, and enable the urban planners to design better road networks and transit routes to meet the needs of urban travelers. The goal of this project is to develop, implement, and evaluate a unified framework to learn the decision-making strategies of human agents from their generated mobility data, with applications to explain and incentivize their decisions to promote individual and societal well-being. Moreover, in this project, the investigator will integrate research, education, and outreach by developing new courses for both undergraduate and graduate students, reaching out to K-12 students, and engaging women and underrepresented minorities.

There are several technical challenges to learn human decision-making strategies from their mobility data: Human decision-making strategies may vary over time and space, i.e., spatial-temporal dynamics challenge. The mobility data collected may cover only a part of the spatial regions and time periods, i.e., spatial-temporal sparsity challenge. Most human agents are not "experts", thus the generated mobility data are noisy and uncertain, i.e., non-expert challenge. A large number of human agents interact with each other when making decisions, i.e., interaction and scalability challenge. Human decisions are governed by many (sometimes hidden) factors, which make it hard to infer explainable information from their decision-making strategies, i.e., explainability challenge. Human agents have diverse reactions to offered incentives, making it hard to design targeted incentive mechanisms to consider both agents' inherent decision-making strategies and their online feedback, i.e., incentive design challenge. This project will address these research challenges, and include the development of a novel spatial-temporal imitation learning framework for learning decision-making strategies from individual and interactive human agents, and an interactive system that provides human agents with explainable learning results and online incentives to promote their decision-making strategies. The spatial-temporal imitation learning framework and associated algorithms have the potential to be transformative in both the data and urban sciences by enabling efficient and accurate discovery of human decision-making strategies from their mobility data.

Publications:

[Conference papers], [Journal papers]

    2024

  1. [SDM'24] Mingzhi Hu, Xin Zhang, Yanhua Li, Yiqun Xie, Xiaowei Jia, Xun Zhou, Jun Luo
    Only Attending What Matter within Trajectories -- Memory-Efficient Trajectory Attention.[Available Soon]
    SIAM International Conference on Data Mining (SDM24), Houston, TX April 18 - 20, 2024. (98/416 Acceptance Ratio)
  2. [SDM'24] Nasrin Kalanat, Yiqun Xie, Yanhua Li, Xiaowei Jia
    Spatial-Temporal Augmented Adaptation via Cycle-Consistent Adversarial Network: An Application in Streamflow Prediction.[Available Soon]
    SIAM International Conference on Data Mining (SDM24), Houston, TX April 18 - 20, 2024. (98/416 Acceptance Ratio)

    3. 2023

  3. [KDD'23] Mingzhi Hu, Xin Zhang, Yanhua Li, Xun Zhou, Jun Luo,
    ST-iFGSM: Enhancing Robustness of Human Mobility Signature Identification Model via Spatial-Temporal Iterative FGSM. [To be available soon.]
    the 29th SIGKDD conference on Knowledge Discovery and Data Mining, Long Beach, CA, USA, August 6 - 10, 2023, (22.1%=313/1416 Acceptance Ratio)
  4. [WACV'23] Guojun Wu, Xin Zhang, Ziming Zhang, Yanhua Li, Xun Zhou, Christopher Brinton, Zhenming Liu,
    Learning Lightweight Neural Networks via Channel-Split Recurrent Convolution. [PDF]
    Winter Conference on Applications of Computer Vision 2023, Waikoloa, Hawaii, Jan 3 - Jan 7, 2023.
  5. [ICDM'23] Mingzhi Hu, Zhuoyun Zhong, Xin Zhang, Yanhua Li, Yiqun Xie, Xiaowei Jia, Xun Zhou, Jun Luo,
    Self-supervised Pre-training for Robust and Generic Spatial-Temporal Representations.[PDF][GitHub]
    IEEE International Conference on Data Mining, (9.37%=87/926 Full paper acceptance ratio), Shanghai, China, Dec. 1 - Dec 4, 2023.
  6. [ICDM'23] Xin Zhang, Yanhua Li, Ziming Zhang, Christopher Brinton, Zhenming Liu, Zhi-Li Zhang,
    Distributional Cloning for Stabilized Imitation Learning via ADMM.[PDF][GitHub]
    IEEE International Conference on Data Mining, (9.37%=87/926 Full paper acceptance ratio), Shanghai, China, Dec. 1 - Dec 4, 2023.
  7. [ICDM'23] Palawat Busaranuvong, Xin Zhang, Yanhua Li, Xun Zhou, Jun Luo,
    CAC: Enabling Customer-Centered Passenger-Seeking for Self-Driving Ride Service with Conservative Actor-Critic.[PDF][GitHub]
    IEEE International Conference on Data Mining, (9.37%=87/926 Full paper acceptance ratio), Shanghai, China, Dec. 1 - Dec 4, 2023.
  8. [SIGSPATIAL GIS'23] Yiqun Xie, Zhaonan Wang, Gengchen Mai, Yanhua Li, Xiaowei Jia, Song Gao and Shaowen Wang,
    "Geo"-Foundation Models: Reality, Gaps and Opportunities (Vision Paper). [PDF]
    31st ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, (Vision paper) Nov. 13 - 16, 2023, Hamburg, Germany.
  9. [SDM'23] Xin Zhang, Yanhua Li, Ziming Zhang, Zhi-Li Zhang
    Domain Disentangled Meta-Learning.[To be available soon]
    SIAM International Conference on Data Mining (SDM23), Minneapolis, April 27 - 39, 2023. (27.5% = 105/382 Acceptance Ratio)
  10. [SDM'23] Yingxue Zhang, Yanhua Li, Xun Zhou, Ziming Zhang, Jun Luo
    STM-GAIL: Spatial-Temporal Meta-GAIL for Learning Diverse Human Driving Strategies.[To be available soon]
    SIAM International Conference on Data Mining (SDM23), Minneapolis, April 27 - 39, 2023. (27.5% = 105/382 Acceptance Ratio)

    11. 2022

  11. [ICDM'22] Yingxue Zhang, Yanhua Li, Xun Zhou, Xiangnan Kong, and Jun Luo,
    STrans-GAN: Spatially-Transferable Generative Adversarial Networks for Urban Traffic Estimation. [PDF]
    IEEE International Conference on Data Mining, (9.77%=85/870 Full paper acceptance ratio), Orlando, FL, Nov. 28 - Dec 1, 2022.
  12. [ICDM'22] Yingxue Zhang, Yanhua Li, Xun Zhou, and Jun Luo,
    Mest-GAN: Cross-City Urban Traffic Estimation with Meta Spatial-Temporal Generative Adversarial Networks. [PDF]
    IEEE International Conference on Data Mining, (9.77%=85/870 Full paper acceptance ratio), Orlando, FL, Nov. 28 - Dec 1, 2022.
  13. [SIGSPATIAL GIS'22] Yichen Ding, Ziming Zhang, Xun Zhou, Yanhua Li,
    EgoSpeed-Net: Forecasting Speed-Control in Driver Behavior from Egocentric Video Data. [Available soon!]
    30th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, (Full paper,) Oct. 30 - Nov. 4, 2022, Seattle, WA, (23.8% = 38/160 Acceptance Ratio).

    14. 2021

  14. [NeurIPS'21] Ziming Zhang, Yun Yue, Guojun Wu, Yanhua Li, Haichong Zhang,
    SBO-RNN: Reformulating Recurrent Neural Networks via Stochastic Bilevel Optimization. [NeurIPS Link] [PDF]
    Thirty-fifth Conference on Neural Information Processing Systems, Virtual Conference, December 6 - 14, 2021, (26.0%=2344/9122 Acceptance Ratio)
  15. [KDD'21] Huimin Ren, Sijie Ruan, Yanhua Li, Jie Bao, Chuishi Meng, Ruiyuan Li, and Yu Zheng,
    MTrajRec: Map-Constrained Trajectory Recovery via Seq2Seq Multi-task Learning. [PDF]
    the 27th SIGKDD conference on Knowledge Discovery and Data Mining, Singapore, August 14 - 18, 2021, (15.4%=238/1541 Acceptance Ratio)
  16. [ICDM'21] Xin Zhang, Yanhua Li, Xun Zhou, Oren Mangoubi, Ziming Zhang, Vincent Filardi, and Jun Luo,
    DAC-ML: Domain Adaptable Continuous Meta-Learning for Urban Dynamics Prediction. [PDF]
    IEEE International Conference on Data Mining, (9.9%=98/990 Full paper acceptance ratio), Auckland, New Zealand, Nov. 7-10, 2021.
  17. [ICDM'21] Ziming Zhang, Guojun Wu, Yue Yue, Yanhua Li, and Xun Zhou,
    Deep Incremental RNN for Learning Sequential Data: A Lyapunov Stable Dynamical System. [PDF]
    IEEE International Conference on Data Mining, (9.9%=98/990 Full paper acceptance ratio), Auckland, New Zealand, Nov. 7-10, 2021.
  18. [ICDM'21] Yingxue Zhang, Yanhua Li, Xun Zhou, Zhenming Liu, and Jun Luo,
    C3-GAN: Complex-Condition-Controlled Urban Traffic Estimation through Generative Adversarial Networks. [PDF]
    IEEE International Conference on Data Mining, (20.0%=198/990 Short paper acceptance ratio), Auckland, New Zealand, Nov. 7-10, 2021.
  19. [SIGSPATIAL GIS'21] Menghai Pan, Xin Zhang, Yanhua Li, Xun Zhou, and Jun Luo,
    Learning Decision Making Strategies of Non-experts: A NEXT-GAIL Model for Taxi Drivers. [PDF]
    29th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, Nov. 2 - Nov. 5, 2021, Beijing, China. (34/152=22.4% Full paper acceptance ratio.)
  20. [CDC'21] Xin Zhang, Weixiao Huang, Yanhua Li, Renjie Liao, Ziming Zhang,
    Imitation Learning From Inconcurrent Multi-Agent Interactions.[PDF]
    the 60th IEEE Conference on Decision and Control (CDC 2021), Austin, TX on December 13-15, 2021.
  21. [TIST] Yingxue Zhang, Yanhua Li, Xun Zhou, Jun Luo, and Zhi-Li Zhang
    Urban Traffic Dynamics Prediction -- A Continuous Spatial-Temporal Meta-Learning Approach, [To be available soon.]
    ACM Transactions on Intelligent Systems and Technologies (TIST), Accepted for publication, July 2021.
  22. [Book Chapter] Yanhua Li, Xun Zhou, and Menghai Pan,
    Chapter 27: Graph Neural Networks in Urban Intelligence, in book entitled "Graph Neural Networks: Foundations, Frontiers, and Applications", [Link]
    Springer, pp 1--720, July 2021.

    23. 2020

  23. [NeurIPS'20] Xin Zhang, Yanhua Li, Ziming Zhang, Zhi-Li Zhang,
    f-GAIL: Learning f-Divergence for Generative Adversarial Imitation Learning. [PDF][NeurIPS Link] [arXiv]
    Thirty-fourth Conference on Neural Information Processing Systems, Virtual Conference, December 6 - 12, 2020, (20.0%=1900/9454 Acceptance Ratio)
  24. [NeurIPS'20] Qiong Wu, Felix Wong, Zhenming Liu, Yanhua Li, Varun Kanade,
    Adaptive Reduced Rank Regression. [PDF] [NeurIPS Link] [arXiv]
    Thirty-fourth Conference on Neural Information Processing Systems, Virtual Conference, December 6 - 12, 2020, (20.0%=1900/9454 Acceptance Ratio)
  25. [KDD'20] Menghai Pan, Weixiao Huang, Yanhua Li, Xun Zhou, Jun Luo,
    xGAIL: Explainable Generative Adversarial Imitation Learningfor Explainable Human Decision Analysis. [PDF] [Link]
    the 26th SIGKDD conference on Knowledge Discovery and Data Mining, San Diego, CA, August 23 - 27, 2020, (16.8%=216/1279 Acceptance Ratio)
  26. [KDD'20] Yingxue Zhang, Yanhua Li, Xun Zhou, Xiangnan Kong and Jun Luo,
    Curb-GAN: Conditional Urban Traffic Estimation through Spatio-Temporal Generative Adversarial Networks [PDF] [Link]
    the 26th SIGKDD conference on Knowledge Discovery and Data Mining, San Diego, CA, August 23 - 27, 2020, (16.8%=216/1279 Acceptance Ratio)
  27. [KDD'20] Huimin Ren*, Menghai Pan*, Yanhua Li, Xun Zhou and Jun Luo,
    ST-SiameseNet: Spatio-Temporal Siamese Networks for Human Mobility Signature Identification [PDF] [Link]
    the 26th SIGKDD conference on Knowledge Discovery and Data Mining, San Diego, CA, August 23 - 27, 2020, (16.8%=216/1279 Acceptance Ratio)
  28. [ICDM'20] Xin Zhang, Yanhua Li, Xun Zhou, Ziming Zhang, Jun Luo,
    TrajGAIL: Trajectory Generative Adversarial Imitation Learning for Long-term Decision Analysis. [PDF]
    IEEE International Conference on Data Mining, (9.8%=91/930 Full paper acceptance ratio), Sorrento, Italy, Nov. 17-20, 2020.
  29. [ICDM'20] Yingxue Zhang, Yanhua Li, Xun Zhou, Jun Luo,
    cST-ML: Continuous Spatial-Temporal Meta-Learning for Traffic Dynamics Prediction. [PDF]
    IEEE International Conference on Data Mining, (19.7%=183/930 Short paper acceptance ratio), Sorrento, Italy, Nov. 17-20, 2020.
  30. [SIGSPATIAL GIS'20] Menghai Pan, Weixiao Huang, Yanhua Li, Xun Zhou, Zhenming Liu, Jie Bao, Yu Zheng and Jun Luo,
    Is Reinforcement Learning the Choice of Human Learners? A Case Study of Taxi Drivers. [PDF]
    28th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, Nov. 5 - Nov. 8, 2020, Seattle, WA, USA. (33/149=22.1% Full paper acceptance ratio.)
  31. [TBD] Menghai Pan, Yanhua Li, Zhi-Li Zhang, Jun Luo,
    SCCS: Smart Cloud Commuting System with Shared Autonomous Vehicles [PDF]
    IEEE Transactions on Big Data, Accepted for publication, 2020.
  32. [TBD] Xin Zhang, Yanhua Li, Xun Zhou, Jun Luo,
    cGAIL: Conditional Generative Adversarial Imitation Learning—An Application in Taxi Drivers’ Strategy Learning [PDF]
    IEEE Transactions on Big Data, Accepted for publication, 2020.
    33. Acknowledgement:
    This material is based upon work supported by the National Science Foundation under Grant No. (NSF grant 1942680).
    Disclaimer:
    Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.