Doctoral position in Spatial-temporal information retrieval from multi-modal historical resources using GeoAI and LLM

In recent years, geoinformation from historical maps has become a valuable resource for scientific studies as it allows for a continuous analysis of the evolution of man-made and natural features, such as cities, transportation networks, forests, and hydrological features. This is particularly important for urban planning, environmental management, and cultural heritage preservation.

The digitization efforts of the Chair of Cartography in recent years have unlocked a vast amount of information from Swiss historical maps, facilitating comprehensive studies of spatio-temporal dynamics. When combined with historical resources from other modalities, such as textual archives, deeper spatio-temporal insights can be uncovered, providing researchers with a more comprehensive understanding of the past.

This doctoral project aims to integrate diverse historical resources—primarily maps and texts—to enhance spatio-temporal information retrieval using GeoAI and Large Language Models (LLMs).

This doctoral research will focus on linking information in historical maps spatially and temporally with other resources, such as registries in historical documents. By integrating these diverse sources, the research will develop methodologies to extract, structure and organize historical data, enabling advanced search and retrieval. The outcomes will enhance the accessibility and usability of historical geospatial data, allowing for comprehensive cross-disciplinary analysis and deeper insights into historical events and transformations.

• Extracting information from various sources (e.g., maps, texts, images)  within a multi-modality framework.
• Registering historical maps from different times by leveraging advanced deep learning approaches.
• Building spatial-temporal knowledge graphs of historical resources with efficient database architecture.
• Integrating LLMs into knowledge graphs for information query.
• Publishing in peer-reviewed journals or present at equivalent conferences.
• Managing historical map data with Atlas of Switzerland and GeoVITE platform.
• Assisting teaching activities.

Applicants must have completed a Master’s degree in a related field (applicants completing their studies before June 2025 will also be considered)

• Master in Computer Science, Applied Mathematics, Remote Sensing, Geomatics or related.
• Experience with deep learning framework (e.g., Pytorch or Tensorflow).
• Experience with database management, preferably for geo-spatial databases (e.g., SPARQL, PostGIS).
• Knowledge of geospatial data formats, GIS tools and analysis methods.
• Experience with LLM framework (e.g., Langchain, LlamaIndex) is a strong plus.
• Excellent written and verbal communication skills.
• Ability to work independently and collaboratively in a team.
• Ability to manage research projects, including planning, resource allocation, and documentation.

• Participate in a small, motivated research team.
• Pleasant working atmosphere.
• National and international network and collaboration.
• Benefits: usage of sport facilities and other university infrastructure on campus, contribution to the cost of public transportation (Halb-Tax and GA).