ProDiff: Prototype-Guided Diffusion for Minimal Information Trajectory Imputation

Tianci Bu, Le Zhou, Wenchuan Yang, Jianhong Mou, Kang Yang, Suoyi Tan, Feng Yao, Jingyuan Wang, Xin Lu
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:5775-5792, 2025.

Abstract

Trajectory data is crucial for various applications but often suffers from incompleteness due to device limitations and diverse collection scenarios. Existing imputation methods rely on sparse trajectory or travel information, such as velocity, to infer missing points. However, these approaches assume that sparse trajectories retain essential behavioral patterns, which place significant demands on data acquisition and overlook the potential of large-scale human trajectory embeddings. To address this, we propose ProDiff, a trajectory imputation framework that uses only two endpoints as minimal information. It integrates prototype learning to embed human movement patterns and a denoising diffusion probabilistic model for robust spatiotemporal reconstruction. Joint training with a tailored loss function ensures effective imputation. ProDiff outperforms state-of-the-art methods, improving accuracy by 6.28% on FourSquare and 2.52% on WuXi. Further analysis shows a 0.927 correlation between generated and real trajectories, demonstrating the effectiveness of our approach.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-bu25d, title = {{P}ro{D}iff: Prototype-Guided Diffusion for Minimal Information Trajectory Imputation}, author = {Bu, Tianci and Zhou, Le and Yang, Wenchuan and Mou, Jianhong and Yang, Kang and Tan, Suoyi and Yao, Feng and Wang, Jingyuan and Lu, Xin}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {5775--5792}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/bu25d/bu25d.pdf}, url = {https://proceedings.mlr.press/v267/bu25d.html}, abstract = {Trajectory data is crucial for various applications but often suffers from incompleteness due to device limitations and diverse collection scenarios. Existing imputation methods rely on sparse trajectory or travel information, such as velocity, to infer missing points. However, these approaches assume that sparse trajectories retain essential behavioral patterns, which place significant demands on data acquisition and overlook the potential of large-scale human trajectory embeddings. To address this, we propose ProDiff, a trajectory imputation framework that uses only two endpoints as minimal information. It integrates prototype learning to embed human movement patterns and a denoising diffusion probabilistic model for robust spatiotemporal reconstruction. Joint training with a tailored loss function ensures effective imputation. ProDiff outperforms state-of-the-art methods, improving accuracy by 6.28% on FourSquare and 2.52% on WuXi. Further analysis shows a 0.927 correlation between generated and real trajectories, demonstrating the effectiveness of our approach.} }
Endnote
%0 Conference Paper %T ProDiff: Prototype-Guided Diffusion for Minimal Information Trajectory Imputation %A Tianci Bu %A Le Zhou %A Wenchuan Yang %A Jianhong Mou %A Kang Yang %A Suoyi Tan %A Feng Yao %A Jingyuan Wang %A Xin Lu %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-bu25d %I PMLR %P 5775--5792 %U https://proceedings.mlr.press/v267/bu25d.html %V 267 %X Trajectory data is crucial for various applications but often suffers from incompleteness due to device limitations and diverse collection scenarios. Existing imputation methods rely on sparse trajectory or travel information, such as velocity, to infer missing points. However, these approaches assume that sparse trajectories retain essential behavioral patterns, which place significant demands on data acquisition and overlook the potential of large-scale human trajectory embeddings. To address this, we propose ProDiff, a trajectory imputation framework that uses only two endpoints as minimal information. It integrates prototype learning to embed human movement patterns and a denoising diffusion probabilistic model for robust spatiotemporal reconstruction. Joint training with a tailored loss function ensures effective imputation. ProDiff outperforms state-of-the-art methods, improving accuracy by 6.28% on FourSquare and 2.52% on WuXi. Further analysis shows a 0.927 correlation between generated and real trajectories, demonstrating the effectiveness of our approach.
APA
Bu, T., Zhou, L., Yang, W., Mou, J., Yang, K., Tan, S., Yao, F., Wang, J. & Lu, X.. (2025). ProDiff: Prototype-Guided Diffusion for Minimal Information Trajectory Imputation. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:5775-5792 Available from https://proceedings.mlr.press/v267/bu25d.html.

Related Material