The proliferation of geospatial data has accelerated due to advances in Earth Observation (EO) technologies, the widespread deployment of Internet of Things (IoT) sensors, the rise of open science initiatives, and the increasing availability of cloud-native and AI-driven geospatial platforms. This data is now highly heterogeneous in format, modality, resolution, and temporal scale, and is being generated at unprecedented speed and volume. These trends introduce significant challenges—and opportunities—in how to efficiently search, integrate, and mine massive, multimodal geospatial datasets to reveal meaningful patterns and support informed decision-making.

In this context, users must not only identify where to search for geospatial objects of interest, but also determine which models are best suited for different analytical tasks. The emergence of geospatial foundation models—large, pretrained models designed to generalize across spatial domains and modalities—presents new opportunities and challenges. How can users locate and reuse models trained on similar tasks? How should such models be stored, indexed, and made searchable? A unified search framework that enables discovery across content, metadata, and analytic tasks would unlock powerful applications, allowing users to identify the most appropriate model, understand its relevance to specific geospatial tasks, and search across diverse data types, including vector, raster, text, time-series, and point clouds.

Looking ahead, there is a clear demand for fast, interactive, multimodal geospatial search. Users may wish to find relevant models for domain-specific objects of interest (e.g., critical infrastructure), search satellite archives using an image query, or monitor rapidly evolving situations for rapid response to events such as floods or earthquakes. This workshop aims to bridge the fields of retrieval, data modeling, and model development for geospatial data to further the emerging landscape of GeoSearch—scalable, intelligent search over geospatial data and models.

Example topics include but are not limited to:

• Search and Retrieval using Geospatial Foundation Models
• Geospatial Knowledge Graphs for Search
• Multimedia Information Retrieval
• Neural Search in Spatio-temporal Data
• Search in Large Spatio-temporal Data Cubes
• Searching Geospatial Objects - Vector Geometry, 3D Point Clouds, Spatial Trajectories, Positioning, Localization and Navigation
• Spatial Representation Learning for Search
• Metadata design and cataloguing of Machine Learning Models
• Design of Data and Model Relational Structures
• Keyword-Search in Spatial Data
• Spatial Index Structures for Search
• Scalable Search Interfaces, Frameworks and Implementations
• Data Representations and Containers facilitating content-based retrieval
• Scalable Management and Analysis of Geospatial Big Data
• Design and Implementation of Spatial Indexing Methods

GeoSearch '24 Workshop Proceedings https://dl.acm.org/doi/proceedings/10.1145/3681769

The GeoSearch 2025 is a half-day workshop in the 33rd ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (ACM SIGSPATIAL 2025), which consists of keynotes and individual paper presentations. Two submission format will be included in this workshop:

• Full research paper: 6-8 pages
• Short research paper or demo paper: 4 pages

Full research papers should present mature research on a specific problem or topic in the context of geospatial search. We also welcome short research articles or industry demonstrations of existing or developing methods, toolkits, and best practices for AI applications in the geospatial domain.

Manuscripts should be submitted in PDF format and formatted using the ACM camera-ready templates available at http://www.acm.org/publications/proceedings-template . All submitted papers will be peer reviewed to ensure the quality and the clarity of the presented research work. Submissions will be single-blind — i.e., the names affiliations of the authors should be listed in the submitted version.

For each accepted paper, at least one author must register for the ACM SIGSPATIAL conference, attend the workshop, and present the paper during the workshop.

Submission system: https://easychair.org/my/conference?conf=geosearch2025