New publication
"GeoAI for Map Generalization in Multi-Scale Cartography”: Drawing on three years of work from our SNSF-funded DeepGeneralization project, this paper explores a GeoAI pathway toward fully automated multi-scale map generalization, and how cartographic theory can, in turn, strengthen GeoAI methods beyond maps.
Zhou, Z., Fu, C., Feng, Y., Touya, G., Sester, M., & Weibel, R. (2026).
GeoAI for map generalization in multi-scale cartography: foundations, a research agenda, and interdisciplinary perspectives. International Journal of Geographical Information Science, 1–29.
Abstract
Multi-scale spatial representation is a crucial mechanism for describing facts about the real world and for managing and communicating spatial data efficiently and effectively. To enable such multi-scale spatial representation, map generalization has thus been developed. The advances of GeoAI, which focus on spatially oriented deep learning and process understanding with AI techniques, have brought about a new paradigm for map generalization. In this foresight paper, we propose a map generalization framework supported by GeoAI techniques and review the evolution of GeoAI for map generalization. Taking into account the full process of producing maps or cartographic data products, we further highlight the research opportunities of GeoAI for map generalization and computational cartography. The potential technical challenges of GeoAI for map generalization are also noted for a longer-term exploration, including high-quality datasets, integrated generalization solutions, cartographic knowledge transfer, and large language models for map generalization. Finally, map generalization, as a fundamental process of spatial representation and reasoning, is expected to inspire, and be inspired by, other research strands regarding the use of novel GeoAI methods, including location encoding, digital twins, historical map interpretation, movement data management and mining, and temporal change detection of maps.