Urban environments generate vast amounts of heterogeneous and multimodal data, including time series, geospatial imagery, textual reports, and sensor streams. Recent advances in foundation models for language, vision, time series, and geospatial data offer transformative opportunities to better understand, simulate, and optimize cities.
However, applying these models to real-world urban systems poses significant challenges, including heterogeneous data fusion, scalability, interpretability, fairness, and the need to transform model outputs into actionable intelligence for urban planning, mobility, climate resilience, and policy-making.
FM4Urban aims to bridge the gap between theoretical advances in foundation models and their practical impact on urban analytics and intelligence. The workshop brings together researchers, practitioners, and policymakers to discuss cutting-edge research, open challenges, and real-world deployments across the urban domain.
Topics of interest include but are not limited to:
Full research papers, short papers, and visionary position papers are welcome.
Papers must be written in English and formatted in LaTeX, following the outline of the author kit Springer LNCS Template Download. Full papers are a maximum of 8 pages (including references). Short and position papers are a maximum of 4 pages. Short and visionary papers (4 pages) will be presented in poster format during the workshop.
All accepted papers must be presented in person at the workshop.
Post-workshop proceedings will be published by Springer in the Communications in Computer and Information Science (CCIS) series. Authors of accepted papers will have the option to opt in or opt out of the post-workshop publication.
The workshop will include invited talks, peer-reviewed paper presentations, an interactive poster session, and a panel discussion bringing together experts from machine learning, urban science, transportation, and climate research.
To foster community engagement, the workshop will feature interactive voting for Best Paper and Best Poster awards using QR codes during the event.