Salvato in:
Dettagli Bibliografici
Autori principali: Wang, Qiongyan, Zou, Xingchen, Jiang, Yutian, Wen, Haomin, Wei, Jiaheng, Wen, Qingsong, Liang, Yuxuan
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2510.16555
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912659305136128
author Wang, Qiongyan
Zou, Xingchen
Jiang, Yutian
Wen, Haomin
Wei, Jiaheng
Wen, Qingsong
Liang, Yuxuan
author_facet Wang, Qiongyan
Zou, Xingchen
Jiang, Yutian
Wen, Haomin
Wei, Jiaheng
Wen, Qingsong
Liang, Yuxuan
contents Rapid urbanization intensifies the demand for Urban General Intelligence (UGI), referring to AI systems that can understand and reason about complex urban environments. Recent studies have built urban foundation models using supervised fine-tuning (SFT) of LLMs and MLLMs, yet these models exhibit persistent geospatial bias, producing regionally skewed predictions and limited generalization. To this end, we propose Urban-R1, a reinforcement learning-based post-training framework that aligns MLLMs with the objectives of UGI. Urban-R1 adopts Group Relative Policy Optimization (GRPO) to optimize reasoning across geographic groups and employs urban region profiling as a proxy task to provide measurable rewards from multimodal urban data. Extensive experiments across diverse regions and tasks show that Urban-R1 effectively mitigates geo-bias and improves cross-region generalization, outperforming both SFT-trained and closed-source models. Our results highlight reinforcement learning alignment as a promising pathway toward equitable and trustworthy urban intelligence.
format Preprint
id arxiv_https___arxiv_org_abs_2510_16555
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Urban-R1: Reinforced MLLMs Mitigate Geospatial Biases for Urban General Intelligence
Wang, Qiongyan
Zou, Xingchen
Jiang, Yutian
Wen, Haomin
Wei, Jiaheng
Wen, Qingsong
Liang, Yuxuan
Artificial Intelligence
Machine Learning
Rapid urbanization intensifies the demand for Urban General Intelligence (UGI), referring to AI systems that can understand and reason about complex urban environments. Recent studies have built urban foundation models using supervised fine-tuning (SFT) of LLMs and MLLMs, yet these models exhibit persistent geospatial bias, producing regionally skewed predictions and limited generalization. To this end, we propose Urban-R1, a reinforcement learning-based post-training framework that aligns MLLMs with the objectives of UGI. Urban-R1 adopts Group Relative Policy Optimization (GRPO) to optimize reasoning across geographic groups and employs urban region profiling as a proxy task to provide measurable rewards from multimodal urban data. Extensive experiments across diverse regions and tasks show that Urban-R1 effectively mitigates geo-bias and improves cross-region generalization, outperforming both SFT-trained and closed-source models. Our results highlight reinforcement learning alignment as a promising pathway toward equitable and trustworthy urban intelligence.
title Urban-R1: Reinforced MLLMs Mitigate Geospatial Biases for Urban General Intelligence
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2510.16555