Cross-City Latent Space Alignment for Consistency Region Embedding

Meng Chen, Hongwei Jia, Zechen Li, Wenzhen Jia, Kai Zhao, Hongjun Dai, Weiming Huang
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:8261-8274, 2025.

Abstract

Learning urban region embeddings has substantially advanced urban analysis, but their typical focus on individual cities leads to disparate embedding spaces, hindering cross-city knowledge transfer and the reuse of downstream task predictors. To tackle this issue, we present Consistency Region Embedding (CoRE), a unified framework integrating region embedding learning with cross-city latent space alignment. CoRE first embeds regions from two cities into separate latent spaces, followed by the alignment of latent space manifolds and fine-grained individual regions from both cities. This ensures compatible and comparable embeddings within aligned latent spaces, enabling predictions of various socioeconomic indicators without ground truth labels by migrating knowledge from label-rich cities. Extensive experiments show CoRE outperforms competitive baselines, confirming its effectiveness for cross-city knowledge transfer via aligned latent spaces.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-chen25ab, title = {Cross-City Latent Space Alignment for Consistency Region Embedding}, author = {Chen, Meng and Jia, Hongwei and Li, Zechen and Jia, Wenzhen and Zhao, Kai and Dai, Hongjun and Huang, Weiming}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {8261--8274}, 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/chen25ab/chen25ab.pdf}, url = {https://proceedings.mlr.press/v267/chen25ab.html}, abstract = {Learning urban region embeddings has substantially advanced urban analysis, but their typical focus on individual cities leads to disparate embedding spaces, hindering cross-city knowledge transfer and the reuse of downstream task predictors. To tackle this issue, we present Consistency Region Embedding (CoRE), a unified framework integrating region embedding learning with cross-city latent space alignment. CoRE first embeds regions from two cities into separate latent spaces, followed by the alignment of latent space manifolds and fine-grained individual regions from both cities. This ensures compatible and comparable embeddings within aligned latent spaces, enabling predictions of various socioeconomic indicators without ground truth labels by migrating knowledge from label-rich cities. Extensive experiments show CoRE outperforms competitive baselines, confirming its effectiveness for cross-city knowledge transfer via aligned latent spaces.} }
Endnote
%0 Conference Paper %T Cross-City Latent Space Alignment for Consistency Region Embedding %A Meng Chen %A Hongwei Jia %A Zechen Li %A Wenzhen Jia %A Kai Zhao %A Hongjun Dai %A Weiming Huang %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-chen25ab %I PMLR %P 8261--8274 %U https://proceedings.mlr.press/v267/chen25ab.html %V 267 %X Learning urban region embeddings has substantially advanced urban analysis, but their typical focus on individual cities leads to disparate embedding spaces, hindering cross-city knowledge transfer and the reuse of downstream task predictors. To tackle this issue, we present Consistency Region Embedding (CoRE), a unified framework integrating region embedding learning with cross-city latent space alignment. CoRE first embeds regions from two cities into separate latent spaces, followed by the alignment of latent space manifolds and fine-grained individual regions from both cities. This ensures compatible and comparable embeddings within aligned latent spaces, enabling predictions of various socioeconomic indicators without ground truth labels by migrating knowledge from label-rich cities. Extensive experiments show CoRE outperforms competitive baselines, confirming its effectiveness for cross-city knowledge transfer via aligned latent spaces.
APA
Chen, M., Jia, H., Li, Z., Jia, W., Zhao, K., Dai, H. & Huang, W.. (2025). Cross-City Latent Space Alignment for Consistency Region Embedding. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:8261-8274 Available from https://proceedings.mlr.press/v267/chen25ab.html.

Related Material