DiffPath: Generating Road Network based Path with Latent Diffusion Model
Keywords: Path Generation, Latent Diffusion Model, Path Distribution, Long-range Dependencies
Abstract: With the increasing use of GPS technology, path has become essential for applications such as navigation, urban planning, and traffic optimization. However, obtaining real-world path presents challenges due to privacy concerns and the difficulty of collecting large datasets. Existing methods, including count-based and deep learning approaches, struggle with two main challenges: handling complex distributions of path segments and ensuring global coherence in generated paths. To address these, we introduce DiffPath, a path generation model based on Latent Diffusion Models (LDMs). By embedding path into a continuous latent space and leveraging a transformer architecture, DiffPath captures both local transitions and global dependencies, ensuring the generation of realistic paths. Experimental results demonstrate that our model outperforms existing approaches in generating paths that adhere to real-world road network structures while maintaining privacy.
Primary Area: other topics in machine learning (i.e., none of the above)
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 3621
Loading