arXiv:2605.26036v1 Announce Type: cross Abstract: Urban representation learning encodes complex urban environments into general-purpose embeddings for diverse downstream tasks and emerging urban foundation models. However, current evaluations are limited, typically focusing on one or two cities and tasks and relying on random splits that introduce spatial leakage, leading to inflated performance and weak support for cross-location generalization and fair comparison. To address this, we propose CityRep, a unified benchmark that evaluates urban representations across data modalities, cities, and
The proliferation of urban foundation models and the recognition of limitations in current evaluation methods necessitate a robust, generalized benchmark for urban representations.
This benchmark will enable more accurate and fair comparison of AI models designed for urban environments, leading to more reliable and effective AI applications in city planning and management.
The development and evaluation of urban AI models will become more rigorous, moving away from localized, potentially inflated performance metrics to cross-city generalizability.
- · AI researchers in urban computing
- · Smart city initiatives
- · Urban planning agencies
- · AI models with poor generalization capabilities
- · Evaluation methods relying on random data splits
Improved performance and broader applicability of AI models in diverse urban settings.
Accelerated development of more robust and responsible urban AI solutions.
Enhanced efficiency and sustainability in urban management and infrastructure development through AI.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG