NARA: Anchor-Conditioned Relation-Aware Contextualization of Heterogeneous Geoentities

TL;DR

NARA is a self-supervised framework that learns rich, context-aware representations of heterogeneous vector geospatial data by modeling semantics, geometry, and spatial relations jointly.

Contribution

It introduces a unified, relation-aware approach for vector geoentity representation learning, capturing diverse spatial relations beyond proximity.

Findings

Improves building function classification accuracy.

Enhances traffic speed prediction performance.

Boosts next point-of-interest recommendation results.

Abstract

Geospatial foundation models have primarily focused on raster data such as satellite imagery, where self-supervised learning has been widely studied. Vector geospatial data instead represent the world as discrete geoentities with explicit geometry, semantics, and structured spatial relations, including metric proximity and topological relationships. These relations jointly determine how entities interact within space, yet existing representation learning methods remain fragmented, often restricted to specific geometry types or partial spatial relations, limiting their ability to capture unified spatial context across heterogeneous geoentities. We propose NARA (Neural Anchor-conditioned Relation-Aware representation learning), a self-supervised framework for vector geoentities. NARA learns context-dependent representations by jointly modeling semantics, geometry, and spatial relations within a unified framework and captures relational spatial structure beyond proximity alone, enabling rich contextualized representations across heterogeneous geoentities of points, polylines, and polygons. Evaluation on building function classification, traffic speed prediction, and next point-of-interest recommendation shows consistent improvements over prior methods, highlighting the benefit of unified relational modeling for vector geospatial data.